12 What Explains the Link Between Spanking and Assaulting a Partner?
A substantial body of research, including the chapters in Parts II and III, has found that spanking or corporal punishment experienced as a child is associated with a broad range of serious behavior problems of children and adolescents. The chapters in this part of the book extend the inquiry to crime as an adult. This chapter and the next chapter are about the links between spanking and physically assaulting a romantic partner or spouse. Chapters 14 and 15 are about the relation of spanking to a broad range of criminal acts. Chapter 16 is about the relation of spanking to forcing sex on a partner. In the concluding part of the book, Chapter 19 summarizes the results of many studies of the relation of spanking to crime, with an emphasis on cross-cultural and longitudinal research. The questions to be addressed in this chapter include:
* What has previous research found about the link between being spanked as a child and later in life assaulting a marital or dating partner?
* What is the relationship of spanking to depression as an adult, to approving of violence, and to marital conflict?
* Are depression, approving violence, and marital conflict processes or mechanisms that explain why spanking is associated with an increased probability of assaulting a partner?
* Does the relation of spanking to assaulting a partner apply when the study controls for things that could be the underlying causes?
* Does the relationship between having been spanked as a child and assaulting a partner apply to assaults by women as well as men?
* Do the results apply to both minor assaults such as slapping a partner and severe assaults such as punching or choking a partner?
* What are the policy and practice implications of the results of the study reported in this chapter?
Spanking and Assaults on Partners
Studies over the last 35 years have found that spanking is related to physically assaulting a partner. Gelles (1974) studied 80 families and found that adults who had been spanked frequently as a child (at least monthly) had a higher rate of assaulting a partner than those who had not been hit. Carroll (1977) studied 96 couples and found that "36.6% of those who had experienced a high degree of parental punishment reported assaulting a spouse compared to 14.5% of those who had not." Other studies found similar results. Johnston (1984) studied 61 abusive men and 44 nonabusive men and found that spanking was related to both minor and severe spous~ abuse. Kalmuss's (1984) analysis of a nationally representative sample of 2,143 American couples found that being slapped or spanked as a teenager more than doubled the probability of husband-to-wife and wife-to-husband assaults.
Straus and Kaufman Kantor (1994) studied a second nationally representative sample and found that spanking was a significant risk factor for assaults on wives, even when other potentially influential variables, such as socioeconomic status, gender, age, witnessing violence between parents, and alcohol use, were controlled. Finally, a longitudinal study of high school boys (Simons et al., 1998) found that the more spanking these boys had experienced, the more likely they were to hit a dating partner. The Simons et al. (1998) study is important because it controlled for the level of misbehavior that presumably led' to the parents spanking when the boys were younger. This is critical because, as noted in previous chapters, parents tend to spank in response to misbehavior. Therefore, if a study does not control for the level of misbehavior in adolescents, it may reflect preexisting aggressive and antisocial tendencies of the boys who were spanked, rather than the effect of spanking.
The research showing that spanking is associated with an increased probability of physical assaults against a partner is consistent with many studies that have found that spanking is related to physical aggression against other children and to other behavior problems. This includes the studies on spanking and child behavior problems in Part II, the other chapters on the relation of spanking to adult violence and crime in this part of the book, and the studies reviewed by (Gershoff, 2002).
The Linking Processes
Previous studies have found that spanking is associated with physically assaulting a partner. But those studies do not show why spanking is associated with an increased probability of physically assaulting a partner. This chapter presents the results of investigating three of the many possible processes that could explain what produces the link between spanking and violence against a partner. The three mediating processes are:
* Corporal punishment teaches that it is morally correct to hit to correct misbehavior and that carries over to relationships between adults.
* Corporal punishment limits development of nonviolent conflict-resolution skills that results in a high level of conflict with partners and, therefore, a higher probability of violence.
* Corporal punishment increases the probability of depression, which in turn increases the probability of aggression.
The Morality of Violence
Although physically assaulting a partner is a criminal, act, American culture actually tolerates and legitimizes such acts in various ways. National surveys show that at least one quarter of the population approves slapping a spouse under some circumstances (Gelles & Straus, 1988; Greenblat, 1983; Moore & Straus, 1995; Simonet al., 2001; Straus et al., 2006; Straus, Kaufinan Kantor, & Moore, 1997). When asked for an example of such a circumstance, by far the most frequently mentioned circumstance was sexual infidelity (Greenblat, 1983, pp. 243-246). Among the university students studied for Chapter 13 on spanking and partner violence in 32 nations, the rates for approving of a husband slapping his wife under some circumstances ranged from 26% to 45%, and the rates for approving of a wife slapping her husband ranged from 65% to 82%.
These attitudes are partly a carryover from a previous historical era when husbands did have the legal right to physically chastise an errant wife (Calvert, 197 4 ). American courts began nullifYing this common law principle in the 1870s, but it has survived in American culture and in the informal culture of the criminal justice system. To take just one of thousands of examples, a New Hampshire judge, when sentencing a man who stabbed his wife, admonished him by saying-"if you had just slapped her, you wouldn't be here today" (Darts & Laurals, 1993). The multitude of ways in which the actions and inactions of the criminal justice system continued to legitimize partner assault has been documented for at least a generation (Straus, 1976). There has been remarkable progress since then, largely due to the efforts of the women's movement. Instead of advising police officers to avoid interfering in domestic disturbances, (International Association of Chiefs of Police, 1967), most police departments now require or recommend arrest (Buzawa & Buzawa, 2003; Sherman, Schmidt, & Rogan, 1992).
The reasons for the persistence of norms permitting marital violence are multiple and complex. This chapter examines the hypothesis that one of the explanations is spanking by parents. Because parents who spank or slap a misbehaving child are doing so with community approval and act in the belief that spanking is sometimes necessary for the child's own good, spanking teaches children the unintended lesson that hitting is a morally correct way of dealing with misbehavior. That lesson can then be carried over to also apply to the misbehavior of an intimate partner.
Social learning theory suggests that one of the ways children learn to use and value violence is by observing and modeling the behavior of their parents (Bandura, 1973). We think this is especially likely to happen if the violence they observe is in the form of spanking a misbehaving child because, as just noted, doing so is socially approved behavior. A 1998 survey of a nationally representative sample of U.S. adults found that 73% believe spanking is a necessary part of child rearing (Ellison & Bradshaw, 2009). Thus, when parents spank to correct and teach, it is accompanied by an unintended hidden curriculum. One of the hidden lessons in that curriculum is that violence can be and should be used to secure good ends-that is, it teaches that violence is morally justified, not just in the extreme of self-defense, but when dealing with persistent misbehavior in ordinary human interaction (Wolfe, Katell, & Drabman, 1982). Another lesson stems from the fact that most parents hit a child only after trying other methods of correction and control. From this, children learn that violence is permissible "when other things don't work" (Straus et al., 2006, pp. 103-104).
Parents assume that these lessons about the morality ofhitting someone who misbehaves and "won't listen to reason" will be applied when their child is an adult, only to hitting a child who misbehaves. Studies of children show, however, that children who are spanked tend to apply these principles to interaction with other children who misbehave toward them (see Chapter 6 and Simons & Wurtele, 2010). This chapter builds on that research by investigating the possibility that the lessons learned persist into adulthood and dating and marital relationships. This possibility arises because it is almost inevitable that, sooner or later, a partner will misbehave and not listen to reason as the partner sees it. As one woman put it, "I punch guys for the same reasons people 'discipline' their children. I've got expectations in love, and I want him to improve" (Connell, 2002).
These research results and theories led to the following hypotheses:
Hypothesis 1: The more spanking experienced, the greater the probability of believing that there are circumstances when one would approve of hitting a partner.
Hypothesis 2: Individuals who believe that it is sometimes permissible to hit a partner are more likely to actually hit their partner than those who do not.
Truncated Development of Conflict-Resolution Skills
Another process that we investigated to try to understand what might explain why spanking is associated with an increase in physically assaulting an intimate partner starts from the assumption that the more parents rely on spanking to deal with misbehavior, the lower the child's skills will be in nonviolent problem solving. This is partly because, as was shown in Chapter 10 on spanking and cognitive ability, spanking slows cognitive development. An even more direct relationship may occur because each time parents spank, it denies the child the opportunity to observe, participate in, and learn nonviolent modes of influencing the behavior of another person. These modes include explaining, negotiating, compromising, and modifying their own behavior to adapt to the situation. As one parent we spoke to·put it, in explaining why she spanked, "I don't have time for all that." Based on this line of reasoning, we hypothesized that:
Hypothesis 3: The more spanking experienced, the greater the probability of inadequate skill in managing conflict and, therefore, a higher probability of unresolved conflicts with partners.
Hypothesis 4: A high level of conflict, in turn, is associated with an increased risk of violence (as shown in Straus et al., 2006, Chart 13).
The data available for this chapter let us test the fourth hypothesis because it includes a measure of the presumed consequence of a lack of such problemsolving skills: a measure of unresolved marital conflict.
Depression
Still another process that might explain why spanking is linked to assaulting a partner identifies depression as a mediating variable (also known as an intervening variable). A mediating variable refers to a characteristic or a process that, if supported by the statistical analysis, provides at least part of the explanation for the link between the hypothesized cause variable and the effect variable.
Depression was included in the theory we tested because of the results from two related lines of research. The first line of research shows that spanking is associated with being depressed as an adult. Straus (1995a) and Straus and Kaufman Kantor (1994) found that, after statistically controlling for six risk factors (e.g., witnessing parents assault each other and a low socioeconomic status), individuals who were slapped or spanked frequently during adolescence were twice as likely to experience severe depressive symptoms and suicidal ideation when they were adults. There are at least 13 other studies that have found that spanking is associated with an increased probability of depression, including: Afifi et al. (2006), Bordin et al. (2009), De Vet (1997), DuRant et al. (1995), Fergusson, Boden, & Horwood, (2008), Harper, Brown, Arias, & Brody (2006), Holmes & Robins (1988), Leary et al. (2008), Spencer (1999), Turner & Finkelhor (1996), Turner & Muller (2004).
A second relevant line of research has found that depression is associated with an increased probability of hostile and aggressive behavior toward others. Although depressed individuals are typically thought of as passive and motivationally deficient, a growing body of research suggests that depression is often associated with aggression, especially in the form of uncontrolled violent outbursts against others (Berkowitz, 1993). The co~occurrence of depression and aggression among children (Garber, Quiggle, Panak, & Dodge, 1991) as well as adults, led Berkowitz (1983, 1993) to speculate that depressive symptoms may be linked to hostility or violence against a partner. This speculation is confirmed by the results of studies focusing specifically on domestic violence. For example, one study found that domestically violent males are more than twice as likely (45% versus 20%) to report symptoms of clinical depression than nonviolent males (Julian & McKenry, 1993). The differences remained even when race, quality of the marital relationship, life stress, and alcohol use were controlled for statistically. Further, Maiuro, Cahn, Vitaliano, Wagner, and Zegree (1988) found that 67% of men who assaulted their wives were clinically depressed compared with 34% of young men who assaulted nonfamily members, and 4% ofnonassaultive men (see also Tolman & Bennett, 1990).
The link between depression and partner assault has not yet been adequately explained and most likely represents a complex, reciprocal relationship with· a number of other characteristics of each of the partners and of their relationship. Some researchers (Maiuro et al., 1988; Tolman & Bennett, 1990), however, have suggested that individuals who are depressed may resort to physical violence to help deal with feelings of helplessness that frequently accompany depression. In the case of marital relationships, an individual may act aggressively toward his or her partner in an effort to reestablish control over a discordant marital relationship that is in jeopardy of dissolving. Additionally, enduring patterns of low selfesteem and personal insecurity, or fears of abandonment, may predispose some individuals to respond aggressively to perceived threats of loss of the relationship.
Postulating that depression serves as a precursor to spousal aggression does not contradict the fact that depression can also be a consequence of being assaulted by a partner, as was shown by Stets and Straus (1990). We believe there is a bidirectional relationship and that depression is both a cause and a consequence of partner violence.
These theories and research results led to the following hypotheses:
Hypothesis 5: The more spanking experienced, the greater the probability of depression.
Hypothesis 6: Depression, in turn, increases the probability of violence against a partner.
Sample and Measures
We used path analysis to test the theory that spanking is related to assaulting a partner because spanking is associated with an increased probability that the child will grow up to approve of violence, have a marriage with a high level of conflict, and be depressed. Each of these three problems, in turn, is associated with an increased probability of hitting a partner. We tested this theory separately for men and women to allow for the possibility that the effects of spanking might be different for men and women and because it is widely believed that assaults on partners by women have a different etiology than assaults on partners by men (Cascardi & Vivian, 1995; Medeiros & Straus, 2006; Navaro, 1995; Straus, 2009c; White & Smith, 2009).
Sample
The sample used to test the theory consisted of 4,401 participants in the 1985 National Family Violence survey (2,557 women and 1,844 men). This sample is briefly described in Chapter 1 0 on number of children and spanking and in more detail in Straus & Yodanis (1996) and the Appendix. This is a nationally representative sample, not a sample selected because of involvement in some type of Violence. The study participants were interviewed as adults about whether they were slapped or spanked when they were adolescents. For the generation who were adults at this time of this survey in 1985, being hit when they were adolescents was far from a rare event. Just over one half of American parents at that time hit early adolescent children, and they did so an average of 8 times a year (Straus & Donnelly, 1994). The high prevalence and chronicity of corporal punishment of adolescents in this nationally representative sample is important because it indicates that, despite the fact that the data are on early adolescence, the results are broadly applicable to that generation and are not restricted to a small number of families in which there was an abnormally high level of violence.
Measures
More detailed statistical information about the following measures are in Straus and Yodanis (1996).
Corporal punishment. Corporal punishment was measured by asking each study participant, "Thinking about when you, yourself, were a teenager, about how often would you say your mother or stepmother used physical punishment like slapping or hitting you? Think about the year in which this happened the most." The response categories were never, once, twice, 3 to 5 times, 6 to 10 times, 11 to 20 times, and more than 20 times. We asked a parallel question about corporal punishment by the participant's father. Empirical research and socialization theories of parent-child relationships have noted that mothers and fathers spend unequal time, perform unique parenting roles, have different interactions, and form dissimilar relationships with their children (Demo, 1992; Peterson & Rollins, 1987). The effects of parental use of corporal punishment may also be different depending on whether it is done by the father or the mother. To find out, we examined the effects of fathers' and mothers' use of corporal punishment separately.
A limitation of this measure is that it depends on the ability and the willingness of the study participants to recall these events. Fortunately, as we noted in the previous chapter, there is empirical evidence indicating that adults' recall of events in childhood can provide a valid measure (Coolidge et al., 2011; Fisher et al., 2011; Morris & Slocum, 201 0). Further, the adult recall questions focused on adolescence because asking adults about corporal punishment at earlier ages would be less accurate. Finally, readers should note that never experiencing corporal punishment meant that it was never experienced during the teenage years; respondents could have experienced corporal punishment before they were teenagers, but that was not the focus of this question.
Assaults between partners. Assaults between partners was measured using the original Conflict Tactics Scales (described in Straus, 1979; Straus et al., 1996).
Violence approval. Violence approval was measured using two questions, "Are there situations that you can imagine in which you would approve of a husband slapping a wife's face?'' and "Are there situations that you can imagine in which you would approve of a wife slapping a husband's face?'' The response categories were 1 = strongly disagree, 2 = disagree, 3 = agree, and 4 = strongly agree.
Unresolved conflict. Unresolved conflict in the respondent's marriage or other relationship was measured by questions about how often the respondent and the spouse or partner disagreed on five issues: managing the money; cooking, cleaning, or repairing the house; social activities and entertaining; affection and sexual relations; and issues about the children. The response categories were 0 =never, 1 = sometimes, 2 =usually, 3 = almost always, and 4 = always.
Depression. Depression was measured using four questions from the Psychiatric Epidemiological Research Instrument (Dohrenwend, Askenasy, Krasnoff, & Dohrenwend, 1978): "Have you been bothered by feelings of sadness or worth?," "Have you felt very bad and worthless?," "Have you had times when you couldn't help wondering if anything was worthwhile anymore?," and "Have you had times when you felt completely hopeless about everything?" Respondents were asked to report how often this happened within the past year: never= 0, almost never= 1, sometimes= 2, fairly often= 3, and very often= 4. More information about this variable can be found in Straus (1995a).
Controls
The analysis included controlling for the following factors that might influence the results: age of the participant (to control for generational differences), socioeconomic status of the ethnic group, and the witnessing of violence between parents. To make the results more readily understandable, the charts in this chapter do not include the paths from the control variables. The results for those paths and more information about the measures and statistical methods used for this chapter are in Straus and Yodanis (1996).
Interrelation of Depression, Conflict, and Approval of Violence
Up to this point, the discussion treated each of the three processes that might explain the relation of spanking to physically assaulting a partner separately. However, attitudes about the legitimacy of hitting a partner, conflict resolution skills, and depression are likely to work together to increase the likelihood of partner assault. This perspective is based on research showing that marital conflict and depression are linked (Beach, Sandeen, & O'Leary, 1990; Julian & McKenry, 1993; Maiuro et al., 1988). As for links with approval of violence, although we have not found previous research showing a connection between approval of violence and marital conflict and depression, such a connection is plausible as the cognitive aspect of the link between marital conflict and depression. The results of examining the relation of these mediating variables to each other revealed the following:
* Depression and marital conflict. Women with a high level of depression were about 3.4 times more likely to have a high level of marital conflict, and men with a high depression score were about 2.5 times more likely to be in a relationship with a high level of conflict. It is important to keep in mind that, because this is a cross-sectional study, it is just as plausible to interpret this result as showing that a high level of conflict is associated with an increased probability of depression, or that there is a bidirectional relationship and perhaps also an escalating cycle.
* Depression and approval of marital violence. Women with a high depression score were 1.7 times more likely to approve of marital violence. For men, there was no relationship between depression and approval of hitting a partner.
* Marital conflict and approval of marital violence. Men who approved of slapping a partner under some circumstances were 2.5 times more likely to be in a relationship characterized by high levels of conflict. For women, there was no relationship between conflict and approval of hitting a partner.
The difference between the findings for men and women might reflect gender differences in socialization and conflict management. To be more specific, these fmdings may reflect the tendency of men to externalize problems in the form of aggression and of women to internalize problems in the form of depression (Kramer, Krueger, & Hicks, 2008; Maschi, Morgen, Bradley, & Hatcher, 2008).
Relationship Between Spanking and Assaulting a Partner
The arrows in Chart 12.1 show the connections between the variables that were found to be statistically dependable. The numbers on each of these paths indicate the percent by which an increase of one unit of the variable at the start of the path is associated with an increase or decrease in the probability of the variable at the end of the path. The detailed tests of significance are in Straus and Yodanis (1996).
Chart} 2.1 Three of the Processes Explaining the Link between Spanking and Assaulting a Partner
Direct Links between Spanking and Assault
Spanking by mothers. The upper section of Chart 12.1 (Part A) summarizes the results for assaults by male partners, and the lower section (Part B) summarizes the results for assaults by female partners. In the chart for assaults by men, the 9% on downward sloping arrow in the middle that goes from Spanking by Mother on the left to Assault by Husband on the far right indicates that each increase of one category in the seven-category measure of having been spanked or slapped by a mother is also associated with a 9% increase in the odds of a man assaulting his partner. The comparable arrow in Part B (the lower half of Chart 12.1) shows the same link between spanking or slapping by mothers and women assaulting their male partners.
Spanking by fathers. The results for Spanking by Father in Part A does not show a direct path from Spanking by Father to Assault by Husband because a statistically dependable relationship was not found. On the other hand, the chart for women (Part B) does show a direct path from Spanking by Father to Assault by Wife. The arrow from Spanking by Father to Assault by Wife shows that for each increase of one category in the spanking or slapping scale, there is an 8% increase in the probability of a woman physically assaulting her partner. In short, spanking or slapping by a mother is associated with an increased probability of later hitting a partner for both men and women, but spanking or slapping by a father is only related to women hitting their partner. However, as will be shown in the n'ext section, spanking has similar indirect effects leading to physical assaults against a partner by both men and women.
The Mediating Mechanisms
The main issue addressed by this study is the theory that part of the explanation for the link between spanking and assaulting a partner is that spanking increases the probability of depression, approval of violence, and marital conflict. The paths from having been spanked or slapped to these three variables in the middle of both Part A and B of Chart 12.1 show the results of investigating three processes that might explain why being spanked or slapped by parents is associated with assaulting a partner.
Depression. One hypothesized process that might help explain the link between having been spanked or slapped by parents and physically assaulting a partner is that spanking or slapping increases the probability of depression. The paths from Spanking to Depression in both parts A and B of Chart 12.1 show that, for men and for women, the more spanking by mothers, the more likely both men and women are to have symptoms of depression. For women, but not for men, this also applies to spanking by fathers. The 193% on the path from Depression to Assault in Part A of Chart 12.1 shows that men in the highdepression group were almost 3 times more likely to have assaulted a partner in the previous 12 months than other men. The same path in Part B of Chart 12.1 shows that women in the high-depression group were just over twice as likely (107% increase) to have physically assaulted a partner than were other women.
Approval of violence. Another process that might explain why spanking is associated with an increased probability of assaulting a partner is that spanking increases the probability of approving violence. For this study, that was measured by whether the study participant agreed that there are circumstances when they might approve of a husband slapping his wife and a wife slapping her husband. Because a central theme of this book is that spanking is a fundamental cause of violence in the family and in society, we graphed that relationship in detail. Chart 12.2 shows that spanking or slapping is related to approval of violence by men, even when there was no violence between their parents as well as when there was. This does not mean that witnessing assaults between parents is unimportant. The fact that the line for men who witnessed this type of violence is higher on the graph shows that witnessing violence is also associated with even more approval of violence. Men who both witnessed physical assaults between their parents and were hit by their parents as an adolescent have the highest level of approving a husband slapping his wife. Almost identical results were found for women.
Returning to Chart 12.1 in Part A, there are paths on this issue for Spanking by Father and for Spanking by Mother. For men, the 11% on the path going from Spanking by Mother to Approval of Violence indicates that each increase of one point on the seven-point spanking or slapping scale is associated with an 11% increase in the probability of approving a husband slapping his wife. The 8% on the path from Spanking by Father to Approval of Violence shows that each increase of one unit in the measure of having been spanked or slapped as a teenager is associated with an 8% increase in the probability of approving the slapping of a partner. Part B of Chart 12.1 shows that for women, spanking by mothers, but not spanking by fathers is associated with a greater probability of approving a wife slapping a husband. For women, the 13% on the path in Chart 12.1 from Spanking by Mother to Approval of Violence indicates a somewhat stronger tendency for spanking or slapping by a woman's mother to be linked to approving slapping a husband. However, for women, we did not find a statistically dependable relationship between Spanking by Father and Approval of Violence. For women, spanking by fathers was related to an increased probability of depression.
Chart 12.2 The More Spanking Men Experienced, the Higher the Probability They Will Approve of a Husband Slapping His Wife
* "Approve" was measured as the percent who did not strongly disagree that "I can think of a situation when I would approve of a husband slapping a wife's face."
The paths in Part A going from approval of violence to assault shows that men who agreed there are circumstances when they would approve of a husband slapping his wife are about twice as likely to actually hit their partner than are other men. Similarly, Part B shows that women who agreed that there are circumstances when they would approve of a wife slapping her husband were about twice as likely ( 109% increase) to have physically attacked their partner in the previous 12 months. These findings are consistent with the theory that spanking teaches the moral legitimacy of hitting someone who misbehaves, and this in turn increases the probability of actually physically assaulting a partner who misbehaves as that partner sees it.
Marital conflict. The paths from Spanking by Mother to Marital Conflict in Parts A and B of Chart 12.1 show that the more spanking by mothers, the more likely both men and women are to report high marital conflict. However, there is no path from Spanking by Father to Marital Conflict because we did not find a statistically dependable relationship for this path. The paths from Marital Conflict to Assault in both Parts A and B of Chart 12.1 show that high marital conflict was associated with men being 164% more likely to assault a partner and women being 243% more likely to assault a partner. These results are consistent with the theory that spanking and assaulting a partner are related because spanking restricts a child's opportunity to learn nonviolent modes of conflict resolution and, therefore, increases the probability of a high level of marital conflict, which in tum is associated with an increase in the probability of physically attacking a partner.
Do the Results Apply to Severe Assaults?
We also investigated the possibility that the results just presented might be different if the outcome variable was severe assaults; that is assaults involving attacks with objects, punching, and choking. These are acts, such as kicking and punching, that are associated with a greater risk of causing injury than slapping, shoving, and throwing things (see Straus & Yodanis, 1996), for information on the severe assault scale.) The results for severe assaults were similar to those just reported. The analysis for men found that depressive symptoms, violence approval, and marital conflict were associated with an increased probability of severely assaulting a female partner. Spanking or slapping by mothers, however, was not associated with severe assaults by men. The analysis for women found that depressive symptoms, violence approval, marital conflict, and spanking by mothers were associated with an increased probability of severely assaulting a male partner. Spanking or slapping by the fathers of these women, however, was not associated with an increased probability of severe assaults by women. Other Variables Related to Assaulting a Partner
The five variables included in the analysis as controls are also of interest in their own right. The tables in Straus and Yodanis (1996) show that:
* Having grown up in a family where there was violence between their parents was associated with an increased probability of depression, approval of slapping a partner, marital conflict, and actually assaulting a partner.
* The probability of approving violence and actually assaulting a partner decreases with age. This is consistent with other studies of crime, including assaulting a partner (Suitor, Pillemer, & Straus, 1990).
* Higher socioeconomic status was associated with a lower probability of both the men and women in this study, being high in depression, approving the slapping of a partner, experiencing marital conflict, and actually . assaulting a partner.
* Minority ethnic group respondents in this study had a higher probability of marital conflict and assaulting a partner.
Summary and Conclusions
This chapter tested a theory about some of the processes that bring about a link between spanking and physically assaulting a partner in a marital or cohabiting relationship. The processes investigated are three of the adverse side effects of spanking: depression, attitudes approving violence, and marital conflict (which we used as proxies for deficits in conflict-resolution skills resulting from the use of corporal punishment). Our theory is that spanking is associated with an increased probability of each of these, and each in turn is associated with an increased probability of assaulting a partner. The results of the analyses from a nationally representative sample of married or cohabiting partners were largely consistent with this model. We found that spanking in adolescence was associated with an increased probability of:
* Experiencing depression as an adult
* Approving violence against a spouse
* A high level of marital conflict
In turn, each of these effects of spanking was associated with an increased probability of physically assaulting a partner. Thus, at least part of the association between spanking and adult spousal assaults is explained by these three variables. Moreover, these links were found for both men and women regardless of age, socioeconomic status, ethnic group, and whether or not the participants in the study grew up in a family in which there were assaults between their parents. Thus, spanking had a unique effect that was in addition to the effect of the control variables.
The Dose-Effect of Spanking
Defenders of spanking believe it is not harmful if it is used only rarely. To address this, we examined differences between participants in this study who were never hit, hit only once, hit only twice, hit 3 times, etc. Those analyses showed that each increase in spanking of an adolescent, starting with just one instance, was associated with an increase in approval of violence and actual violence toward a partner. Harmful effects for rare spanking are shown in the chapters on spanking and children's antisocial behavior, the outcomes associated with impulsive spanking (Chapters 6 and 7), and the other four chapters in this part of the book and in Straus (1994, Chart 7-2, 2001a) and in a study by Turner and Finkelhor (1996). Turner and Finkelhor used data from interviews with children age 10 to 16. They found that even one or two instance of spanking at those ages was associated with an increase in stress in children.
Some Limitations
Although the analysis controlled for a number of possible sources of spurious findings, some limitations of the research need to be considered to properly evaluate the findings. First, the effect of spanking might have occurred because some of the participants who experienced spanking might also have experienced more serious violence in the form of physical abuse. If so, that probably accounts for part of the effects on the participants who experienced frequent spanking. It is unlikely to explain the effects of low and moderate amounts of spanking because those participants were unlikely to have been victims of more severe physical attacks. In addition, s.tudies by MacMillan et al. (1999); Straus and Kaufman Kantor (1994); Straus and Donnelly (2001a); Vissing et al. (1991); and Yodanis (1992) were able to exclude abused childfen and, after excluding them, each of those studies found that spanking continued to have significant harmful side effects such as those reported in this chapter.
Another limitation is that the spanking data were obtained by asking the study participants about being hit by their parents when they were adolescents. This raises several problems. Those who were hit at that age might be an unrepresentative sample. However, as reported earlier, more than half of the cohort in this study recalled being hit at the age. Thus, corporal punishment of early adolescents was typical of the U.S. population of that time. Since then, the percent of early adolescent children who are slapped or spanked has decreased. The data are, therefore, dated in respect to the prevalence of spanking, but as we pointed out earlier, this does not necessarily affect the relationship between spanking and later assaulting a partner. Another problem is that adults' recall data can be biased. However, it is not necessarily biased, as shown by studies such as (Coolidge et al., 2011; Fisher et al., 2011; Morris & Slocum, 2010). Nevertheless, participants in the study who hit their partners might perceive and report their parents as more violent than those who did not assault their partner. Although that is a concern for the study in this chapter, there are studies that do not depend on recall. These include the chapters on spanking and child behavior (Chapter 6), impulsive spanking (Chapter 7), the child-to-mother bond (Chapter 8), and spanking and adult crime (Chapter 15). Strassberg et al. (1994), Gershoff et al. (2010), and Taylor et al. (2010) all found that spanking is associated with subsequent violence. It is unlikely that, 20 years later, the study participants can accurately recall how many times they were hit. As a consequence, we view this as an ordinal measure that indicates who was hit more than others, not a measure of the actual number of times. Finally, most of those who were not spanked in their early teenage years were probably spanked at earlier ages. Thus, the results do not refer to those who were spanked and those who were not. They mostly refer to those for whom spanking continued into the early teenage years. One implication of this limitation is that the results shown may be minimum estimates of the effects of spanking because the most : of the group who were not spanked in their early teenage years experienced the spanking earlier.
Finally, this was not a longitudinal study that controlled for misbehavior that led to the spanking. As a consequence, unlike the three longitudinal studies in this book and the longitudinal studies listed in Chapter 19 on spanking and its implications for societal-level crime and violence, the relationships found may reflect the effects of the maladaptive characteristics of the study participants who hit their partner, rather than the effects of spanking.
Policy Implications
To the extent that these cross-sectional findings can be interpreted as reflecting a causal relationship between spanking and assaulting a partner later in life, eliminating or reducing spanking can contribute to reducing marital violence. This is because, although spanking has declined in the United States, at least one third of American parents continue to hit early adolescent children (see Chapter2).
Reducing or ending spanking may also make an indirect contribution to ending marital violence through its effect on the way the criminal justice system deals with partner violence. Police and prosecutorial policies intended to reduce partner violenc~ have often been weakly implemented (Garner & Maxwell, 2009). It is possible that the link between spanking and approval of slapping a partner may be one of the factors underlying this weak implementation. The attitudes of police, prosecutors, and judges, like most other Americans, may reflect violence-justifying effects of spanking. That may be part of the reason the criminal justice system so often fails to act against all but the most egregious cases of marital assault. To the extent that criminal prosecution in cases of partner violence is an effective policy, ending spanking in child rearing could contribute to ending the de facto institutional practices that tolerate violence between marital, cohabiting, and dating partners.
The results in this chapter also suggest that a reduction in spanking could have a beneficial impact on one of the most pervasive forms of psychological distress-depression. Close to 15% of the population will experience an episode of major depression at some point in their lives (Substance Abuse Mental Health Services Administration, 2005). Moreover, mood disorders account for more usage of mental health services than any other psychiatric disorder and are responsible for the majority of all attempted and completed suicides (Boyer & Guthrie, 1985; Charney & Weissman, 1988). Thus, in addition to the association of spanking with physical assaults, the severity of the consequences associated with depression also underscores the need for professionals working with parents (such as nurses, parent educators, and pediatricians), to inform parents about the harmful side effects of spanking and help them avoid using spanking.
13 Assault and Injury of Dating Partners by University Students in 32 Nations
The previous chapter showed that spanking is associated with an increased probability of physically assaulting a partner. It also provided information on some of the processes or mechanisms that could explain why spanking is related to assaulting a partner among couples in a large and nationally representative of U.S. households. The research described in this chapter extends the examination of the link between spanking and assaulting a partner by examining these questions:
* What percent of university students in 32 nations physically assaulted a dating partner in the previous year?
* Does the link between spanking and physically assaulting a partner found for U.S. couples in the previous chapter apply to dating relationships of university students in 32 nations?
* What is the effect of being in these 32 different national social contexts on the prevalence of attitudes favoring slapping a partner and actually assaulting a partner, and the relationship of spanking to those attitudes and behaviors? That is, when the percent of the students in a national context who were spanked a lot is large, is there a correspondingly large perpent of students who:
• Approve of slapping a partner? • Actually assaulted a partner? • Injured a partner?
Information on the extent to which physical violence in university student dating relationships result in physical injury is important because we think that many people believe that acts of violence between dating couples is rarely serious or dangerous.
Emily M. Douglas is the first author of this chapter.
National Differences in Dating Partner Violence
There is an important difference between this chapter and the previous chapter in the method of research. This chapter presents the results of a cross-national comparative study in which the issue is the potential effect of the national context on violence in couple relationships. To do this, we used data on the percent in each of the 32 nations of university students who experienced corporal punishment and the percent in each nation of students who had attitudes favorable to hitting a partner under certain circumstances, the percent in each nation who actually assaulted a partner, and the percent who assaulted seriously enough to physically injure their partner. The underlying theoretical reason for the focus on nations is to find out if social settings where there is a high rate of violent socialization of children in the form of spanking also tend to be social settings where there is a high rate of physically assaulting a dating partner.
Assaults on Dating Partners
Student Rates versus General Population Rates
As high as the percent of married couples who engage in physical assaults is (see previous chapter), numerous studies have found even higher high rates of physical assault on dating partners by university students. The typical results show that from 20% to 40% of students physically assaulted a dating partner in the previous 12 months (Archer, 2000, 2002; Katz, Washington Kuffel, & Coblentz, 2002; Sellers, 1999; Sugarman & Hotaling, 1989). The dating couple rates are 2 to 3 times greater than the rates typically found among representative samples of American households, and both the household and the dating couple assault rates are many times the rate of assaults known to the police. Assaults known to the police are reported as rates per 100,000 population, whereas assaults between couples are so prevalent that percentages (the rate per 100 couples) are more appropriate.
The much higher rate of assault by university students than in surveys of households is probably because most students are much younger. The average age of university students is about 20, whereas the average age of community samples of couples is about 40. The long established age-crime curve refers to the fact that most crimes, and especially violent crimes, peak in the late teens or early twenties and then decline rapidly. That is clearly the case for the crime of assaulting a partner. The rates decline from a peak of over 30% for the late teens to less than one half that at about age 40 (Straus & Ramirez, 2007; Suitor et al., 1990). Most of the dating violence studies have been in the United States and Canada. As noted previously, one of our objectives was to determine the extent to which these monumentally high assault rates are found among students in other national settings around the world. If high rates of physical assaults against dating partners are found to be characteristic of university students in most or many countries, it adds urgency to research which can help explain why so many students engage in this type of behavior.
Hypotheses
Like any form of violence, assaulting a partner has multiple causes. For purposes of this book, we investigated whether the prevalence of corporal punishment is one of them. The specific hypotheses we tested are:
The higher the percentage of students in a nation who are spanked as children, the higher the percent of students who:
* Approve of a husband slapping his wife and a wife slapping her husband * Assaulted a dating partner * Injured a dating partner
Sample and Measures
Sample
The hypotheses were tested using data on assaults perpetrated against dating partners by university students who participated in the International Dating Violence Study in 32 nations. The study and the sample are described in Chapter 3 on the worldwide prevalence of spanking and in (Rebellon et al., 2008; Straus, 2008b, 2009b). We analyzed data on the 14,252 ofthe 17,404 students in the study who were in a romantic relationship. Of the 32 nations, there were two in Africa, seven in Asia, thirteen in Europe, four in Latin America, two in the Middle East, two in North America, and Australia and New Zealand.
Measures
Spanking. Spanking was measured by asking the students whether they agreed or disagreed that: "I was spanked or hit a lot by my parents before age 12" and "When I was a teenager, I was hit a lot by my mother or father." The response categories were: strongly agree (1), agree (2), disagree (3), and strongly disagree (4). Anyone who did not strongly disagree was considered to have been spanked as a child. This is based on assuming that if their parents had not spanked or hit a lot they would choose strongly disagree. We tested that assumption before proceeding by computing the correlations of these variables with approval of a husband hitting his wife under some circumstances. Each pair of correlations compared counting anyone who chooses any response, except strongly disagree, as having been spanked with correlations using agree or strongly agree as the criterion. All the correlations were higher using not strongly disagree as the criterion for having been spanked.
The question asks about being spanked or hit a lot because of evidence that a lot is the typical experience in the United States, and perhaps in most of the other nations in the study. Studies that measured how often toddlers were hit have found an average to 2 to 3 times a week (Giles-Sims et al., 1995; Holden et al., 1995; Stattin et al., 1995). One study found that even at age 6, 70% of children were hit once a week or more (Vittrup & Holden, 2010).
Approval of partner violence. Two questions were used to measure approval of violence against a partner: "I can think of a situation when I would approve of a husband slapping a wife's face" and "I can think of a situation when I would approve of a wife slapping a husband's face." The response categories for these questions were the same as for the spanking question. The cut points were again the percentage of students at each nation who did not strongly disagree. And again, exploratory analyses found stronger correlations using this cut point than with other possible cutting points.
Measures of partner violence. Physical assault and injury were measured by the revised Conflict Tactics Scales or CTS2 (Straus et al., 1996). In the past 25 years, the Conflict Tactics Scales have been used in hundreds of studies in many nations, but especially in economically developing nations because it is part of the World Health Organization sponsored studies of maternal health. It has demonstrated cross-cultural reliability and validity (Archer, 1999; Straus, 1990a, 2004; Straus & Mickey, 2012). For this chapter, we used the CTS2 scales measuring physical assault and physical injury, and the subscales for severe assault and severe injury. As in previous studies using the CTS with general population samples, most of the assaults and injuries were in the minor category. Because severe violence may be a unique phenomenon with a different etiology, (Straus, 1990c; Straus, 2011; Straus & Gozjolko, in press) all analyses were conducted for the overall rates of partner violence, and then repeated for the rates of severe violence.
Physical assault. The CTS2 uses five behaviors to measure minor assault, for example, slapping a partner. For severe assault, there are seven behaviors, for example, punched or choked a partner. The overall rate of partner assault is the percent in each nation who perpetrated any one or more of the 12 acts.
Injury. There are five CTS2 items to measure injury inflicted on a partner, such as "Having a sprain, bruise, or cut after a fight with a partner" (a minor injury item) and "Needed to see doctor because of a fight with a partner." The scales for assault and injury were coded 1 if any of the acts occurred in the past year and coded 0 if there were none. The data used for this chapter are the percentage of students at each nation with a score of 1, which is the percentage who assaulted or injured a dating partner. The reliability for the overall physical assault scale for the samples in this study was .88. For the injury scale, the alpha coefficient was .89.
Moderator and control variable. Because the etiology and the effects of assaulting a partner may be different for men and women, we used gender as a moderator variable by repeating all analyses for male and female students.
The analysis controlled for the score for each nation on the limited disclosure scale (Chan & Straus, 2008; Sabina & Straus, 2006; Straus et al., 2010; Straus & Mouradian, 1999). In research on self-reported criminal behavior, differences between nations could reflect differences in the willingness of people in different nations to report socially undesirable behaviors and beliefs as much or more than real differences in crime and criminogenic beliefs. For example, we found that the higher the score on the limited disclosure scale, the lower the percent who agreed that they could think of a situation when they would approve of a husband slapping his wife (r = -.36) and the lower the percent who said they had injured a partner (r = -.21). Scores of the students in each nation on the limited disclosure scale permit a statistical control for nation to nation differences in willingness to disclose such information. Without this control, the results could be spurious; that is, the correlations might reflect differences between nations in the willingness of students to self-report socially undesirable behavior and beliefs rather than differences in spanking.
Data Analysis
The analyses used a nation-level data file, in which the cases are the 32 nations, not individual students. The data for each case consists of summary statistics for the nation, such as the mean or the percentage of students with a certain characteristic. Separate data files were created for males and females, based on aggregating the data for the males and females in each site, and the analyses were replicated using those files.
Partial correlation analysis was used to test the hypothesized relationships of corporal punishment to approval of slapping a spouse, perpetration of physical assault against a dating partner, and injuring a dating partner. The analyses controlled for the age of the respondent, length of the relationships,· social desirability, and socioeconomic status-and for analyses of the total sample, gender of the respondent.
Differences Between Men and Women and Between Nations
This section describes the percent of students at each of the 32 nations who were spanked as a child and as a teenager, the percent who approved of violence against a dating partner, and the percent who physically assaulted and injured a dating partner in the previous 12 months. For each of these beliefs and behaviors, we give the average percentage for all32 nations, and also the nation with the lowest and the highest percent and the percent for the United States. The section below on differences between nations explains how to use Charts 13.2 to 13.7 to get a rough estimate of the percentages for each nation, and for men and women students.
Spanking before age 12. The percent of students who were spanked as a child ranged from 15% in the Netherlands to 75% in Taiwan. The average for the 32 nations was 47%. Table 2 in Chapter 3 and Chart 13.2 give the percent for each of the 32 nations and also show that in almost all nations, boys were hit somewhat more than girls.
Chart 13.1 Nations Where Students Have Experienced More Spanking Tend to Also Be Nations Where a Larger Percent of Students Approve of Partner Violence, Assault a Partrler, and Injure a Partner"
• M = Correlation for male student data. F = correlation for female student data.
*p < .05, **p < 01. Coefficients on each line are the partial correlations using nations as the cases (N = 32), controlling for mean limited disclosure scale score of each nation.
Chart 13.2 The Larger the Percent in a Nation Who Experienced Spanking as a Child, the Larger the Percent Who Approved a Husband Slapping His Wife
Chart 13.3 The Larger the Percent in aN ation Who Experienced Spanking as a Child, the Larger the Percent Who Approved a Wife Slapping Her Husband
Chart 13.4 The Larger the Percent in a Nation Who Experienced Spanking as a Child, the Larger the Percent Who Assaulted Their Partner
Chart 13.5 The Larger the Percent in a Nation Who Experienced Spanking as a Child, the Larger the Percent Who Severely Assaulted Their Partner
Chart 13.6 The Larger the Percent in a Nation Who Experienced Spanking as a Child, the Larger the Percent Who Injured Their Partner
See Chart 3.2 for the key to nation abbreviations.
Chart 13. 7 The Larger the Percent in a Nation Who Experienced Spanking as a Child, the Larger the Percent Who Severely Injured Their Partner
See Chart 3.2 for the key to nation abbreviations.
The 47% who experienced spanking as a child, as reported by the students in this sample, is almost certainly an underestimate for the children and families from the nations in this study. Among the U.S. students, for example, only 61% reported spanking before age 12, whereas Chapter 2 on spanking in the United States shows that 94% of the parents interviewed reported spanking toddlers at least once in the preceding year. The difference between 94% and 61% probably reflects at least three things. First, many people cannot remember much of what happened when they were toddlers (the peak ages for spanking). Second, the question asked about being spanked a lot, whereas the percent in Chapter 2 is for any spanking in the previous 12 months. Third, this is a sample of university students and, therefore, has relatively few from the lowest socioeconomic status households where spanking is most prevalent. These reasons for underestimating probably apply to students in all 32 nations. To the extent that the underestimate applies to all the nations, it means that the percentages can be used to rank the percent of students in the 32 nations who were spanked. That rank order is crucial for determining if the extent of spanking in a national context is associated with the percent who assaulted.
Teenagers. We had expected the percent of students who were hit as teenagers would be substantially lower than the percent who were spanked before age 12, in part because that is what our three national surveys of parents in the United States have found (see Chapter 2). However, the percentages for the university students in this sample are lower, but not that much lower. They ranged from 10% hit as a teenager in the Netherlands to 66% in Tanzania, with the average for the 32 nations being 33%. This is lower than the 44% for being hit before age 12, but not nearly as much lower as we expected. As we found for being hit before age 12, the percent of boys in each nation who were hit by parents as a teenager was almost always slightly higher than the percent of girls who were hit as a teenager (see Chapter 3, Table 3.2).
Approval of husband slapping his wife. The students were asked whether they agreed or disagreed with "I can think of a situation when I would approve of a husband slapping a wife's face." The percent who did not strongly disagree that they could approve of a husband slapping his wife in some situations ranged from 28% in the Netherlands to 85% in Russia. The average for the 32 nations was 51%. Among U.S. students, 38% did not strongly disagree that they could approve of a husband slapping his wife in some situations.
Approval of a wife slapping her husband. The students were also asked the extent to which they agreed or disagreed with "I can think of a situation when I would approve of a wife slapping a husband's face." The percent who did not strongly disagree that they could approve of a wife slapping her husband in some situations ranged from 59% in Japan to 96% in Russia. The median for the 32 nations was 77%. Among U.S. students, 77% did not strongly disagree that they could approve of a wife slapping her husband in some situations.
The much higher percent of students who would approve of a women slapping her husband in some circumstance (median of 51%) than approve of a man slapping his wife (median of 77%) is consistent with several other studies (Felson, 2000; Felson & Feld, 2009; Gulas, McKeage, & Weinberger, 2010; Nabors, Dietz, & Jasinski, 2006; Simonet al., 2001; Sorenson & Taylor, 2005; Straus, 1995b; Straus, Kaufman Kantor et al., 1997). It is an indication of cultural toleration of a woman hitting her partner. Part of the explanation is probably public recognition of the greater potential for injury when men assault women (Greenblat, 1983). Regardless of why assaults by women are more accepted, that cultural acceptance is probably also one of the reasons that more than 200 studies have found higher rates of assault by female than male partners (Archer, 2002; Fiebert, 2010; Straus, 2008b, 2009c). Of course, many other variables also contribute to the high rate of assaults by women on male partners, for example, women are protected (up to a certain point) by chivalry norms (Felson, 2002; see also Straus, 1999; Winstok & Straus, 2011a).
Differences between nations. The approximate values for each nation of the variables in this study can be found by reading the values shown on the vertical and horizontal axes in Charts 13.2 to 13.7. For example, in Chart 13.2, the plot point for Israel (ISR) in the lower left of the panel for men shows that 33% of the male students in Israel did not strongly disagree that "I can think of a situation in which I would approve of a husband slapping a wife'sface." On the other hand, when it comes to a wife slapping her husband, the plot point for Israel in Chart 13.3, shows that almost twice as many of the men (63%) did not strongly disagree that there are situations in which they might approve of a wife slapping her husband. As pointed out earlier, greater acceptance of a wife slapping her husband than the husband slapping his wife is consistent with several other studies.
Assaulting a partner. The percent who physically assaulted a partner in the previous 12 months differed greatly from nation to nation. For all assaults regardless of whether it was a minor assault like slapping or a severe assault like punching, the rates ranged from 17% in Portugal to 40% in Iran, and the average for the 32 nations was 30%. A rate in the range of20% to 40% is typical of most studies of dating partner violence (Archer, 2002; Katz et al., 2002; Stets & Straus, 1990; Sugarman & Hotaling, 1989). Even the lowest assault rate in this study is still very high, compared with rates for other nonclinical populations. For example, it is about 3 times higher than the percent of married and cohabiting couples who physically assault (Gelles & Straus, 1988; Straus & Gelles, 1986; Straus & Smith, 1990).
Our finding that, in most of the 32 nations, female students were somewhat more likely than male students to have assaulted a partner is also typical of other studies of university students (Archer, 2002; Desmarais et al2012; Straus, 2008b ). The percent of male and female students in each nation who assaulted a partner in the previous 12 months is given in Straus (2008b) and can also be determined by reading the value for a nation in the vertical axis of Chart 13.6.
For severe assaults, the rates are of course much lower. The percent who severely assaulted a partner in the previous 12 months ranged from 1.8% in Sweden to 22% in Taiwan, and the average for the 32 nations was 11%. Among U.S. students, 11% reported severely assaulting a partner. Considering that these are severe assaults (roughly analogous to an aggravated assault in the U.S. legal system), these are extremely high rates.
Injured a partner. Although in this and more than 200 other studies (Desmarais et al2012; Fiebert, 2010), as many or more female than male students assaulted a partner, the attacks by male students usually resulted in more injury than attacks by female students (Archer, 2000, 2002; Capaldi et al., 2009; Whitaker, Haileyesus, Swahn, & Saltzman, 2007). The typical finding is that women suffer about two thirds of the injuries. However, for this study, as in a few other studies, the difference was much smaller. The 6.9% of female students who were injured is only 15% higher than the 6.0% of male students who suffered an injury from an attack by a female partner. The results for severe injury show a larger difference: 1.6% of the women suffered a severe injury compared with 1.1% of the men, which is a 45% higher injury rate for women. For the U.S. part of the sample, the rates for any injury of students are 7.8% of women and 7.6% of men, which is only a 3% difference. For severe injury however, the rate of injury of U.S. women students was 46% greater than the rate of severe injury of male students (1.9% of women, 1.3% of men). These injury rates are roughly consistent with the rates for a nationally representative sample ofU.S. couples (Stets & Straus, 1990).
Relationship between Spanking and Partner Violence
As astonishing and important as are the high percent of university students who approve of slapping a partner under some circumstances, and who actually assaulted a partner or inflicted injury, the crucial issue for this chapter is why these aspects of violence are so prevalent. Of course, there are many forces at work. The one we investigated for this book is whether nations in which spanking is more prevalent also tend to be nations in which there is a high rate of approving of slapping a partner and actually assaulting and injuring a partner. To find out, we computed the correlation of the level of spanking in each nation with each of the six aspects of partner violence described above, again, by nation, controlling for the score on the limited disclosure scale. The correlations are shown in Chart 13 .1. Most of these correlations are large enough to be, statistically dependable despite being based on only 32 cases. Because the link between having been spanked and assaulting a partner might be different for men and women, we ran the analysis separately for male and female students. In addition, the analysis controlled for nation-to-nation differences in reluctance to disclose socially undesirable beliefs and behaviors. Chart 13.1 summarizes what we found in the form of the correlations between spanking and each of the six aspects of partner violence studied. Charts 13.2 to 13.7 are scatter plots with the regression line showing the relation of the percent spanked in each nation to each of the six aspects of violence. They also indicate (with the aid of a ruler), the percentages for each of the 32 nations for each of these aspects of violence, separately for men and women. See the example presented earlier in the section on differences between nations.
Spanking before Age 12
The correlation coefficient of .57 on the top arrow in Chart 13.1 shows that the larger the percent of male students in a nation who were spanked or hit a lot before they were 12, the larger the percent who did not strongly disagree with the statement "I can think of a situation when I would approve of a husband slapping a wife's face." The correlation of .42 on that line is for women. It shows almost as strong a relationship between being spanked as a child and approval of a husband slapping his wife. Chart 13.2 lets the reader see which nations are low and high on spanking and approval of a husband slapping his wife. The trend line shows that the higher the percent of students in a nation who were spanked or hit a lot before they were 12, the higher the percent who approved of a husband slapping his wife under certain circumstances, and that this applies to both men (the left side of the chart) and women (the right side of the chart).
Spanking during the Teenage Years
The correlations on the underside of the top arrow in Chart 13.2 are for teenag _ ers. For men, the correlation of .53 is almost as strong a relationship to approval of a husband slapping his wife as was found for being spanked as a young child. For the women in this sample, however, the correlation of .28 is not large enough to be statistically dependable when there are only 32 cases.
In Chapter 5 on approval of violence and spanking, we presented similar results, but for that chapter, the focus was on whether people who approve of slapping a wife are more likely to approve of spanking. We believe both are correct-that is, that spanking increases the probability of being inclined to violence (our interpretation of the results in this chapter) and that being inclined to violence increases the probability of spanking (our interpretation of the results in Chapter 5).
Spanking and Approval of a Wife Slapping Her Husband
The correlations on the top of and below the arrow leading to Approve of Slapping by Wife in Chart 13.1 show that the only statistically dependable result we found for the issue of whether spanking is related to approval of a wife slapping her husband was for women who were spanked or hit a lot before they were 12 years old. Even that correlation is smaller than the correlations for approval of husband slapping a wife. This was unexpected because, as shown earlier in this chapter, a larger percent of the students approved of a wife slapping "her husband. Although only one of the four relationships is statistically dependable, the percent of men and women students in each nation, who approved of a wife slapping her husband under certain circumstances, are important descriptive statistics.
Spanking and Assaulting a Partner Any assault. The percentages of men and women who assaulted a partner are given in Chart 13 .4. The arrows leading to Assault in the center of Chart 13.1 give the partial correlations from testing the hypothesis that the higher the prevalence of spanking in a nation, the more likely students would be to physically assault a dating partner. Contrary to what we expected, for men we found no relationship between the percent in a nation who were spanked and physically assaulting a partner. However, for women, the larger the percent in a nation who experienced spanking, the larger the percent who physically assaulted a partner.
Severe assault. Most of the assaults in the Assault measure are minor, such as slapping or throwing something at a partner. Because severe violence may be a unique phenomenon with a different etiology (Straus, 1990c; Straus, 2011; Straus & Gozjolko, in press), we repeated the analysis using severe assaults such as punching and choking a partner. The correlations on the arrows to Severe Assault against Partner are all large and statistically dependable. Thus, for both men and women, the percent in a nation who were hit by parents to correct misbehavior, either as a child or as a teenager, is related to an increased percent in who severely attacked a partner. Again, the link between spanking and severe assault on a dating partner is slightly stronger for the women in this study than for the men. The percentage in each nation who severely assaulted a partner in each nation is indicated in the vertical axis of Chart 13 .5. See the example for Israel of how to find these percentages in the charts given in a previous section on differences between nations.
Spanking and Injuring a Partner
Any injury. Charts 13.6 show the percentages in each nation who injured a partner. The arrow leading to Injury in the lower part of Chart 13.1 shows that the larger the percentage of students in a nation who experienced spanking as a child, the higher the percentage who injured a dating partner. This relationship applies to both male and female students when the outcome variable is any injury.
Severe injury. When the outcome variable is severe injury to women, the results, are similar in the sense that all four of the correlations are in the predicted direCtion, but only the correlation of .63 for the relation of corporal punishment as a teenager to severely injuring a partner is large enough to be statistically dependable. Although the correlation of spanking before age 12 with severely assaulting a partner is not strong enough to be statistically dependable, we included Chart 13.6 (the scatter plot for this relationship) because it shows the approximate percent of students in each of the 32 nations who severely injured a partner, including the fact that there are six nations in which 3% or more of the men severely injured a partner and four nations in which 3% or more of the women students severely injured a partner.
Summary and Conclusions
Although not all the correlations were statistically dependable, the general pattern is that nations where spanking is more prevalent tend to also be nations in which a higher percent of the students in this study approved of slapping a partner, actually physically assaulted a partner, and injured a partner. These societal-level results show that the link between spanking and violence applies not only to the characteristics of individual persons, but also to the national contexts in which the students in this study lived.
Limitations
Before drawing further conclusions from these results, some important limitations need to be mentioned. First the data is on university students and may be unique to that sector of the population of a nation. However, the results are consistent with those of other studies (including the preceding chapter), which have found similar relationships for representative samples. In addition, there is evidence that nationto-nation differences found for these students corresponds to national differences in the same variables found by other studies. Evidence showing the validity of the International Dating Violence Study data is in Straus (2009b ).
Another limitation is that about two thirds of the sample is female. We reduced the potential problems resulting from this limitation by conducting separate analyses for males and females.
The measure of spanking is another limitation because it is recall data. However, as we pointed out in a previous chapter, there is empirical evidence indicating that adult recall of events in childhood can provide a valid measure of childhood experiences (Coolidge et al., 2011; Fisher et al., 2011; Morris & Slocum, 201 0). Another limitation of the measure is that it asks about having been spanked or hit a lot. As explained in the description of this measure, we think the phrase a lot fits the experience of most children. Nevertheless, the number of times the students had in mind for a lot is unknown, and it likely varies between students and sites. In addition, because the question asked about being spanked or hit a lot, it could be interpreted as a measure of physical abuse. Our opinion, however, is that this is not likely because, as pointed out earlier, the research shows that two or more times a week is typical. As a consequence, one would have to conclude that physical abuse is typical. Our opinion is that any spanking of children is abuse, but that is not the cultural or statistical norm in the United States or most other nations.
Finally, although we controlled for the score of each nation on the limited disclosure scale, because of the small sample size (32), we did not control for enough other variables to have more confidence that the results really reflect the effect of spanking, such as whether there was violence between the parents. However, the study in the preceding chapter did control for violence between the parents and found that spanking made a unique and additional contribution to explaining the occurrence of assaulting a partner.
Links between Spanking and Assaulting Dating Partners
Keeping these limitations in mind, the results support the hypothesis that the larger the percentage of persons in a social context who were spanked, the higher the prevalence of three aspects of violence in partner relationships: cultural norms tolerating or supporting hitting a partner, physically assaulting a dating partner, and assaulting severe enough to injure a partner. An unexpected fmding is that the links between spanking and these three aspects of partner violence are stronger for women than for men. More than 200 other studies have found the percent of women who assault a partner is as high or higher than the percent of male partners who assault a partner (Archer, 2002; Desmarais et al., 2012). The results in this chapter on the stronger link between spanking and partner violence for women, when combined with the data showing high rates of spanking experienced by these women, may be part of the explanation for the high percentage of women who physically assaulted a partner.
Why Is Spanking Linked to Violence against a Partner?
Elsewhere in this book, we discussed and presented results on the relationship between spanking and antisocial and aggressive behaviors by individual persons. In this chapter, the issue included whether there is a similar relationship at the societal level. Is a society in which spanking is prevalent also likely to be a society in which other types of violence are prevalent? Levinson's study ( 1989) of the societies in the Human Relations Area Files, for example, found a correlation of .32 between the extent of corporal punishment used and the extent of wife beating. Levinson's findings, like those in this chapter, are consistent with the cultural spillover theory of violence (Baron & Straus, 1989; Baron, Straus, & Jaffee, 1988). The cultural spillover theory asserts that the more a society uses violence for socially legitimate purposes such as bringing up children or punishing criminals, the more individual citizens are likely to use violence for socially illegitimate and often criminal purposes. That is, violence in any sphere oflife will tend to engender violence in other spheres oflife (see Chapter 19 for a discussion of spillover of violence from one sphere oflife to another). To the extent that this theory is correct, prevention of spanking will contribute to prevention of partner violence and vice versa. That is, as we noted earli~r, it is also likely that the relationship between spanking and violence against a partner is bidirectional; that is, societies in which violence is prevalent are also likely to be societies in which children are brought up violently, that is, with the use of spanking.
Prevalence of Corporal Punishment in Adolescence
A previous study found that, for Americans who were adults at the time of this survey in 1985, being hit as an adolescent was not a rare event. Just over one half of American parents at that time continued corporal punishment into adolescence, and they did so an average of 8 times in the previous 12 months (Straus & Donnelly, 1994). The high prevalence and chronicity of corporal punishment of adolescents is important because it indicates that, despite the fact that the data is about early adolescence, the results are applicable to the majority of the participants; that is, they are not restricted to a small number of families in which there was an abnormally high level of violence.
Policy Implications
The research in this chapter revealed that despite considerable progress toward ending partner violence (Durose et al., 2005; Smithey & Straus, 2004; Straus, 1995b; Straus & Gelles, 1986), a large proportion of university students continue to physically attack a dating partner. To the extent that the results of this research are correct in identifying spanking as a risk factor for partner violence, further steps toward primary prevention of partner violence should include increased policy and programmatic efforts to end all use of spanking by parents.
The finding that the social context, as measured by the percentage of students who experienced spanking, is related to the percent of students who assaulted a partner calls for an expanded view of the processes linking spanking to criminal violence. It suggests the process involves more than the criminogenic effect of spanking on specific individuals who were hit by their parents. These results suggest that when a large percentage of the population are brought up violently, it may create a social climate that increases the probability of other kinds of violence by amplifying the effect of spanking, and perhaps by increasing the probability of violence by those who did not themselves experience a high level of spanking. If so, it strengthens the argument in Chapter 19 on spanking and societal-level crime and violence that spanking is part of the process that creates and maintains a violent society, and that one way to curb this violence is through public policy than bans all forms of corporal punishment, including so-called mild spanking.
14 Cultural Context and the Relation of Spanking to Crime
A disproportionate number of delinquent children in the United States are Black. Poverty, racism, and neighborhood violence are major parts of the explanation for the disproportionate number of Black delinquent children. However, the results of the studies in Part II and the other chapters in Part IV suggest that violent child rearing in the form of spanking is also part of the explanation. That idea has been heatedly rejected by most Blacks we have talked to, including Black social scientists such as Polite (1996). Obviously more research is needed. This chapter provides some of it by presenting results from an analysis of a nationally representative of U.S. households on the following questions:
* What is the theoretical basis for believing that the relation of spanking to crime is different for different race or ethnic groups?
* What has previous research found about the belief that spanking does not have harmful side effects in sociocultural contexts where it is the statistical or cognitive norm?
* What percent of Americans who were adults in 1985 were hit by parents when they were teenagers?
* Is there a link between having experienced corporal punishment as a teenager and the following nine aspects of criminal behavior and beliefs among Black and Hispanics as well as Whites?
• Child aggression • Child delinquency • Assaulted a spouse in previous 12 months • Approval of a husband slapping his wife in some situations • Approval of a wife slapping her husband in some situations • Assaulted a nonfamily person in previous 12 months • Severely assaulted a child in previous 12 months • Drug use • Arrested in previous 12 months
* How consistent are the results of this study with those of 15 other studies examining whether the effects of spanking are mitigated by cultural norms supporting spanking?
Sean Lauer is the coauthor of this chapter.
Sociocultural Context Differences in the Effects of Spanking
Many Blacks, including some Black social scientists, believe that spanking not only corrects the immediate misbehavior but also helps make the child generally law abiding. They argue that spanking is part of the Black cultural tradition and that it is necessary to use strong discipline (usually a euphemism for spanking) to protect children in the difficult circumstances of inner-city life (Daniel, 1985; Ealey, 1980; Hampton, 1987; Harris, 1992; Ispa & Halgunseth, 2004; Polite, 1996; Taylor, Hamvas, Rice, Newman, & DeJong, 2011; Thomas & Dettlaff, 2011 ). There is also an important minority of Black social scientists who stress the unintended harmful effects of spanking-for example, Hampton (1987), Comer and Poussaint (1992), and Earls and colleagues (Molnar et al., 2003 ). They oppose spanking in any form. One has developed a parent education program, Effective Black Parenting (Alvy & Marigna, 1987), which explicitly argues that continued reliance on spanking is a dysfunctional carryover of patterns based on slavery.
Catholic schools have traditionally used corporal punishment. But by the spring of2011, of the thousands of Catholic schools in the United States, only one continued to do so-St. Augustine's High School in New Orleans. This is an all-male and predominately Black school. Archbishop Gregory Aymond of New Orleans called on St. Augustine to end the practice and pointed out that no other Catholic school in the United States uses corporal punishment. The request was met with demonstrations by more than 500 students, parents, and other supporters of the school's use of corporal punishment. They contended that its discipline code and practices are the reason their young men are successful in life and their future careers. Rev. John Raphael, president of St. Augustine High School, said the controversy is not as much about the paddle as about the right of Black parents to discipline their children in the matter they see fit (Dequine, 2011).
Why Might Spanking Have Different Effects for Black Children?
Academic defenders of spanking by Black parents have used cultural relativity theory and adaptation to difficult life circumstances theory to argue that spanking may be beneficial for Black children.
Cultural relativity theory draws on the slavery experience and argues that under those circumstances it was a matter of life and death for Blacks and their children to be obedient. Misbehavior could even result in being sold. This brutal reality required unquestioned obedience of children that was achieved by spanking. To the extent that spanking remains an element of Black culture, children of parents who do not spank may perceive their parents as not caring or not loving, with all the negative consequences that flow from such feelings (Alvy, 1987; Kohn, 1969; Peters, 1976; Young, 1970).
Deterrence theory (Gibbs, 1975) is implicit in the perspective ofBaumrind (1991a) and Belsky (1991). They argue that the dangers of ghetto life, although different from the dangers of the plantation, are equally lethal; for example, there is peer pressure to use drugs and participate in crime. Under those circumstances strong deterrents are needed to prevent children from succumbing to alcohol and drug abuse, violence, and other crime that surrounds them in the ghetto. A beating by parents when the child misbehaves is thought to provided that.
Bronfenbrenner (1985) and others present an argument that has much in common with social disorganization theory (Bursik, 1988; Faris, 1955). Because society, and especially the social circumstances of low-income minorities, has become increasingly unstable, there is a need for families to provide more structure and stability in the lives of their children, including a high level of engagement and firm discipline. Bronfenbrenner rejected spanking as part of the firm discipline, but others such as Baumrind (1972, 199lb, 1992a, 1992b, 1996; Baumrind et al., 2002) accept it.
Deater-Deckard and Dodge (1997) suggest that because spanking is considered appropriate and normal in Black families, it is delivered dispassionately and matter-of-factly rather than in the context of a rageful, out-of-control parent. To the extent that this is correct, spanking would not be accompanied by the psychological maltreatment, such as terrifying threats or emotionally abusive verbalizations, that much research has indicated has more pernicious effects on child development than spanking alone (Vissing et al., 1991). We have not located empirical research on the emotional tone or impulsivity of corporal punishment by Black parents, but there is research that can throw some light on the issue of the extent to which spanking and psychological aggression are linked among both Black and White parents. It is a nationally representative sample that included 792 White and 120 Black parents (Straus et al., 1998). We computed the correlation between spanking and psychological aggression for both groups of parents. The results showed a stronger link between spanking and psychological aggression among Black parents than among White (correlation for Black parents .75 and .51 for White parents; p (less than) .001 for both correlations).
Deater-Deckard et al. (1996) and Gunnoe and Mariner (1997) found that spanking was associated with an increase in the child fighting in school for White children but not for Black children. Both emphasize this cultural context difference in discussing their results. However, using a composite scale to measure aggressive and antisocial behavior, both Gunnoe and Deater-Deckard also found that spanking was related to more aggressive and antisocial behavior for Black as well as for White children. In the Gunnoe and Mariner paper, the findings section includes only one brief sentence acknowledging that their study "replicates the Straus et al. findings" on the link between spanking and antisocial behavior regardless of ethnic group. The extensive discussion and conclusion sections omits mentioning the results showing that spanking at Time 1 was associated with more antisocial behavior subsequently for children of all ages and all ethnic groups.
A similar discrepancy between the results and the discussion of the results occurred in a six-nation study of the effects of spanking (Lansford et al., 2005). The results section states that "in all countries higher use of physical discipline was associated with more aggression and anxiety." However, the final sentence of the article says that the findings "suggest potential problems in using physical discipline even in contexts in which it is normative." Thus, the unambiguous results have become something that is suggested by the findings, and only that spanking is a potential problem. The Duke University press release contradicted the actual results even more directly. It declared that, "A particular parenting practice may become a problem only if parents use it in a cultural context that does not support the practice" [emphasis added].
Given the competing theoretical arguments and the inconsistency in research on this important issue, there is a clear need for additional research. The study reported in this chapter was undertaken to provide some of the needed additional research, and specifically to test what can be called the Universal Harm Theory of spanking, that:
Regardless of cultural context, spanking is associated with an increased probability of violence and other crime, and drug abuse.
The universal harm theory is in contrast to the cultural relativity theory that argues that spanking has no adverse effects (and many would argue has beneficial effects) if done in compliance with cultural norms that prescribe spanking to correct misbehavior.
Sample and Measures
The data for this chapter come from the second National Family Violence Survey (Gelles & Straus, 1988; Straus & Gelles, 1986) that included 2,557 mothers and 1,844 fathers of children under age 18 living at home. This is the same sample used for Chapter 12 on the link between spanking and assaulting a partner. Although the study in this chapter was mainly stimulated by the controversy over whether the adverse effects of spanking apply to Black children, Hispanic Americans were also analyzed because the contextual conditions of poverty, segregation, and dangerous high-violence neighborhoods also apply to many Hispanic families. In such neighborhoods, parents of all race and ethnic groups tend to spank more (Winstok & Straus, 2011 b). The study oversampled minorities to provide enough cases to be examined separately.
Measures of Spanking
Two measures of spanking were used. The first measure used adult recall data. It was obtained by asking the adult participants in the survey, "Thinking about when you yourself were a teenager, about how often would you say your mother or stepmother used physical punishment, like spanking or slapping or hitting you? Think about the year in which this happened the most. Never, once, twice, 3 to 5 times, 6 to 10times, 11 to20times, morethan20times."The second measure uses contemporaneous parent report data and will be referred to as contemporaneous data. It was obtained using the Conflict Tactics Scales to ask parents with a child under 18 living at home whether and how often they had spanked during the preceding year. The version of the CTS used for this study is very similar to the revised CTS used for Chapter 3 on spanking in the United States.
The Data on Crime
Using the adult recall data on spanking, we were able to test the relatipnship between having experienced spanking as a child and seven aspects of violence and crime later in life: assaults on individuals who are not part of their family, arrests, physical abuse of children, approval of a wife slapping a husband, approval of a husband slapping a wife, actually assaulting a partner, and drug use.
Using the contemporaneous data, we could test the relationship between spanking experienced by children as reported by their parents and two outcomes during childhood: aggressiveness and delinquency. The above measures are described in the Appendix.
Control Variables
The data for this sample permitted us to control for several characteristics of the participants and their families. These are characteristics that might be the underlying reason for the relation between spanking and crime. For example, if there is a tendency for parents who are physically violent to each other to also hit their children more than other parents, a correlation between spanking and violence and other crime by their children might be due to the children having witnessed parents being violent toward each other rather than due to the child being spanked. As a consequence, the analyses using the adult recall measure of spanking controlled for:
* Whether the participants' parents were violent to each other * The age of the participant * Family socioeconomic status * Gender of participant
The analysis using the contemporaneous data controlled for all of the above, as well as four additional variables:
* Number of children in the household
* Whether the spanking was by the mother or the father
* The amount of reasoning used by the parent when dealing with misbehavior by the child (see Appendix for·details on the questions)
* Physical abuse
As a result of these controls, the statistics in this chapter show the net effect of spanking. The net effect is the degree to which spanking is related to crime over and above the relation to crime of the -characteristics that were controlled.
In addition, it was also possible to control for the confounding of legal spanking with physical abuse by excluding from the sample children who experienced one or more of the items in the severe assault scale of the parentchild Conflict Tactics Scales (Straus et al., 1998). This cou~d only be done, however, for the analysis using the contemporaneous data because no measure of physical abuse was available for the adult recall data sample.
Chart 14.1 The More Spanking, the Greater the Probability of Assaulting a nonfamily Person
Corporal Punishment and Assault and Arrest as an Adult
Assault
Assaults of a nonfamily person. The parts of Chart 14.1 show that for all three ethnic groups, the more spanking experienced as an adolescent, the higher the probability that they had physically assaulted someone outside their family during the 12 months preceding the survey.
Each part of Chart 14.1 has three lines, one for each ofthree different socioeconomic status levels: low socioeconomic status (20th percentile, i.e., the bottom fifth), middle socioeconomic status (50th percentile), and high socioeconomic status (80th percentile, i.e., the highest fifth in socioeconomic status). These lines show the relation of spanking to the probability of assaulting someone other than a family member by study participants who are low, middle, and high in socioeconomic status. The lines in the charts for the lowest socioeconomic status group (the 20th percentile) are higher than the other lines, indicating a higher assault rate for the lowest socioeconomic status group. This is consistent with all other U.S. national data on assaults. More important, these results show that spanking as a child is associated with an increased probability, of having assaulted someone during the year covered by this study, even though the effect of socioeconomic status is controlled because everyone in each group is in the same socioeconomic status group.
The detailed statistics in the Appendix indicate that, among Black men, each increase of one point in the spanking scale was associated with a 13% increase in the probability of assaulting a nonfamily person. An increase of 10% and 13% percent may seem to indicate that, despite being statistically significant, the effect size is small (i.e., that spanking does not strongly influence later behavior). However, those percentages mean an average increase of 10% and 13% associated with each step in the eight-interval spanking scale. Thus the cumulative effect of these step-by-step increases is substantial. For example, in the upper left section of Chart 14.1, the line for Blacks in the high socioeconomic status group starts at about 8% who assaulted among those who reported no spanking as a teenager and increases to about double that for those at the right side of the line who experienced the most spanking as a teenager. The upper line in the Black part of Chart 14.1 shows that among low socioeconomic status Blacks, the probability of assaultive behavior in the preceding 12 months increased from about 15% for those who did not recall any instance of having been hit by their parents as a teenager to almost double (30%) for those who reported 30 or more instances of having been hit by a parent.
The results are similar for Hispanic American and White men (increases of 13% and 10%, respectively). For women, the rate of assaults was much lower than for assaults by men, but the results show that spanking of girls is associated with a similar increase in the probability of women physically assaulting someone when they are adults.
The upper right part of Chart 14.1 shows that Hispanic Americans who experienced the most spanking had an assault rate that is 2.6 times greater than those who were not hit as a teenager. The three plot lines overlap so much that it seems as though there is only one line because, for the Hispanic Americans in this study, socioeconomic status was not related to assault.
Section C ofthe chart shows that for Whites, although on average the assault rates are lower than for the other two ethnic groups, those who experienced the most spanking were almost twice as likely to assault someone outside their family, as those who experienced the least spanking.
Arrest
The results from testing whether spanking is related to having been arrested during the year covered by this study were in the predicted direction (i.e., the more spanking, the greater chances of having been arrested). None of the relationships in these analyses, however, met the scientific standard for statistical dependability (statistical significance), and thus we cannot conclude that spanking is related to arrest based on this study. This may be a reflection of the small proportion who were arrested in the year covered by this study (3% of Blacks, 2% of Whites, and 1% of Hispanic Americans). Another problem was an error in asking about arrest. The phrase "for something serious" was supposed to be included in the question about having been arrested, but it was omitted and the error was not discovered until after the survey was administered and the data had been collected. Thus, an unknown proportion of the arrests may have been for traffic violations.
Corporal Punishment and Family Violence
Physical Abuse of Children
A previous ~alysis of this sample found that the more spanking or slapping experienced, the greater the probability of physically abusing a child in the year covered by this study (see Straus & Yodanis, 2001, for the results and for an analysis of the processes linking corporal punishment and physical abuse). That analysis, however, did not investigate whether the results applied to different racial or ethnic groups. The analysis for this chapter found a strong relationship between having been spanked and later in life physically abusing a child. We also found that this relationship applied to each of the three racial or ethnic groups. In fact, of the possible effects of spanking on the 12 aspects of crime and violence examined for this chapter, spanking had the strongest relation to physical abuse of a child. This is not surprising if one takes the view that physical abuse is part of a continuum that begins with spanking (Straus, 2000; Straus & Yodanis, 2001).
Spouse Assault
This section presents results for an aspect of crime and violence that we think of as criminogenic attitudes (i.e., attitudes that are related to or might increase criminal activity). We asked the participants in the survey a slight variation of the question used in the study of university students (Chapter 13), "Are there any situations that you can imagine in which you would approve of a husband slapping his wife's face?'' The question was repeated for the approval of a wife slapping her husband's face. We found that in all three racial or ethnic groups, . spanking is associated with an increased probability of the participant approving. For approval of a wife slapping her husband, the findings for Hispanic and Whites are similar, but the overall approval rates are higher, and the relationship between spanking and approval of a wife slapping her husband is stronger than for approval of a husband siapping his wife. Among Blacks, however, the relationship between spanking and approving slapping a wife was not statistically significant.
The next question is whether the link between spanking and attitudes approving violence against a partner translates into actually hitting a partner. The analysis in the previous chapter using the same sample and found that the more spanking experienced, the more approval of violence, and the greater the probability of having assaulted a partner in the previous year (Chapter 12). What t the analysis for this chapter adds is whether spanking is a risk factor for partner assault1 in all three racial or ethnic groups.
Corporal Punishment and Drug Abuse
The hyPothesis that, for both men and women, spanking is associated with drug use was supported. However, consistent with the cultural context theory, the association was not significant for Black males or Hispanic females.
Corporal Punishment and Child Delinquency and Aggression
For all three ethnic groups, the more spanking the parents reported using in the previous 12 months, the greater the probability that their child was classified as physically aggressive (see Appendix for the measure). For Hispanic and White children, but not for Black children, spanking was associated with an increased probability of delinquency.
Systematic Review of Cultural Context Effects
Studies of the effect of the cultural context on the relation between spanking and later maladaptive behavior are hard to compare and to arrive at a conclusion about what they show overall. First there are the usual problems making the studies difficult to compare such as differences in the type of research design, sample, the measures used, and the mode of analysis. Second, as documented in the introduction to this chapter, when reviewing the studies, some authors who favor spanking tend to focus on the studies that found no harmful effect. Similarly, some authors who are opposed to spanking tend to emphasize results and cite studies showing that the harmful effects apply to all children. A metaanalysis of the research on cultural context would be the best way to avoid this type of bias. Although we did not do a formal meta-analysis in the sense of a specified search for all available studies and computation of effect sizes under different conditions, we did do a systematic ~omparison. We located and tabulated the results of 17 studies that compared the side effects of spanking in different cultural contexts, one of which is the study in this chapter. Many of the studies reported more than one test of the effects of spanking in different cultural groups, or reported separate tests for different side-effect variables. For example, Lansford, Deater-Deckard, Dodge, Bates, and Pettit (2004) reported separate tests for children and adolescents and repeated these tests for different outcome variables of whether spanking was associated with harmful side effects for the child in different cultural contexts. As a consequence, these 17 studies provided results for 60 comparisons of cultural groups. To provide the basis for systematic analyses of these results, we created a table listing each of the studies. For each study, the table gives information on the sample, measures, and the results of comparisons of cultural groups provided by the study. Because the table is so long (27 pages), we created Table 14.1, which summarizes the results. The full table can be downloaded from the section on corporal punishment (pubpages. unh.edu/-mas2/).
Table 14.1 Percent of Cultural Context Comparisons that Found a Cultural Context Effect (This table is a summary of the detailed table in pubpages.unh.edu/~mas2/.)
The upper panel of Table 14.1 summarizes the results of 17 comparisons using as the outcome various aspects of antisocial behavior by the child. The total row for that section of the table shows that only 4 of the 25 comparisons found support for the theory that spanking is harmless in cultural contexts where spanking is the cognitive or statistical norms.
The center panel of Table 14.1 summarizes the results of 11 comparisons using as the outcome various aspects of child internalizing problems, and one study reporting results for a measure that combined internalizing and externalizing problems. None of the 11 comparisons supported the theory that spanking is harmless in cultural contexts where spanking is the cognitive or statistical norm.
The lower panel of Table 14.1 summarizes the results for 26 comparisons using as the outcome variable externalizing and criminal behavior by adolescents and adults. Ten of the 26 comparisons (38%) supported the cultural context theory that spanking is unrelated to crime in cultural contexts where spanking is the cognitive or statistical norm. However, caution is needed in interpreting these results because 7 of the 1 0 comparisons supporting the idea that cultural norms mitigate the harmful effects of spanking are from one study (Lansford et al., 2004).
Overall, 77% of the 60 tests contradicted the cultural context theory. However, the 23% of the comparisons that were consistent with the cultural context mitigation theory cannot be dismissed. Further research is necessary to try to identifY why a cultural context effect. was found in those studies. For example, because so many Blacks believe in spanking (see Chapter 5 and Ellison & Bradshaw, 2009; Flynn, 1996a; Straus & Mathur, 1996), spanking is virtually synonymous with discipline. In that context, no spanking may mean no discipline. If that is correct, in view of the research on the importance of parental monitoring and control for prevention of delinquency (Buker, 2011; Patterson, DeBaryshe, & Ramsey, 1989; Sampson & Laub, 1993), it is no wonder that spanking has no harmful effect, or even a positive effect, when no spanking tends to mean no discipline. This issue urgently needs further research. Such research could include data to test both the no spanking means no discipline explanation and the cultural legitimacy explanation.
Summary and Conclusions
The controversy over cultural context effects has sometimes become painful because it involves not only science, but also racial pride and the well-being of children, as was illustrated by the controversy about the use of corporal punishment in a Black New Orleans high school described earlier in this chapter. We think the balance is shifting away from corporal punishment by Blacks, even though slowly. For example, in 2011 a Black anti-spanking educational organization was founded (sparethekids.com/). They describe their mission as providing "Black parents, families and communities with a full range of alternatives to corporal punishment." Nevertheless, as illustrated by the principle of St. Augustine High School, many Black leaders continue to believe that efforts to end spanking are an imposition of middle class, White cultural values in the guise of science (Daniel, 1985; Ealey, 1980; Hampton, 1987; Harris, 1992; Polite, 1996). One of us experienced this personally after speaking to a meeting of mostly Black social workers. In the discussion period after the talk, a member of the audience said it was "racism in disguise-trying to ram white middle class values down our throats." This was followed by loud applause.
A frequent theme was that attempts to end the use of spanking will put Black children in even more danger than they already face. Our view is exactly the opposite. We believe that continued use of spanking by Black parents makes it even more likely that that their children will be violent, become victims of violence, and be trapped in the life of poverty and crime. This opinion is likely to outrage many who will point to the array of other criminogenic conditions faced by Black children, such as poverty and racism, and experiences correlated with poverty such as inadequate diet and failing schools. The need to address our conclusion that spanking is a criminogenic aspect of Black experiences in no way denies or subtracts from the importance of the problems just listed. It simply adds to the list.
On the optimistic side, spanking may be more amenable to change than other causes of crime such as poverty, racism, and failing schools. In addition, the evidence in Chapter 17 on trends in use of spanking found that Black support for spanking has declined, even though not as much as among Whites. As the average education of Blacks increases, that decline is likely to accelerate.
What was called the universal harm theory of spanking was examined by testing the hypothesis that spanking is associated with an increased probability of criminal beliefs and behavior regardless of racial or ethnic group. All together, 36 analyses were conducted using 12 measures of criminal beliefs and behavior to determine if spanking was related to crime by Blacks, Hispanic Americans, and Whites. The findings are mostly consistent with the universal harm theory because 35 out of36 resulted in finding that spanking was associated with an increased odds of crime or criminal beliefs, and 29 of the 36 odds· ratios were statistically dependable (see Appendix). The crucial issue for this chapter is whether the link between spanking and crime also applies to all three race or ethnic groups. The 36 analyses included examined the same set of 12 crime variables for each cultural group. Among Whites, 11 of the 12 tests found that spanking was related to crime. Among Blacks, it was 8 out of 12, and among Hispanics, 10 out of 12 tests found that spanking is related to crime.
These findings are noteworthy because, even though they are not based on longitudinal data, they are from a study that overcomes some of the limitations of many previous studies. First, the findings refer to large and nationally representative samples. Second, the analyses controlled for many potentially confounding variables. As a result of these controls, we know that adverse side effects of spanking manifest themselves regardless of socioeconomic status, gender of the parent doing the punishment, gender of the child, age at which spanking was used, number of children irt the household, amount of reasoning used by the participant when dealing with misbehavior by the child, or physical violence between the participant and his or her spouse. In two thirds of the tests, they also apply to Blacks. For the relation of spanking to child aggressiveness and delinquency, it was also possible to exclude children who had experienced physical abuse, thus providing confidence that the relationship was not the result of the overlap between physical;abuse and spanking.
Theoretical Implications
For the most part, the findings support the perspective of universal harm from spanking rather than the idea that the harmful effects of spanking are culturally relative. Universal harm does not mean that every child is damaged by being spanked. It is only a minority, just as only one third of heavy smokers die from a smoking related disease. See the explanation of risk factors in the concluding section of Chapter 1.
The findings can also be taken as lending some support for cultural relativity theory because the hypothesized links between spanking and crime were less consistently supported within the Black sample. On the other hand, such an interpretation would have to ignore the fact that a significant relationship between spanking and crime was found in two thirds of the tests for Black families and five out of six tests for Hispanic American families. Moreover, the weaker fmdings for the Black and Hispanic American samples might occur because the forces producing the high rates of substance abuse and crime in Black and Hispanic neighbors may be so powerful that it overwhelms the importance of whether parents spank.
Cultural relativity theory. There is an important aspect of the cultural relativity theory with which we agree. It is the idea that parental disciplinary methods that are appropriate for socializing a child to adapt to one type of society may not be appropriate in another society that has a different set of values and behavioral expectations. If, for example, a society emphasizes close adherence to rules more than it emphasizes individual initiative and creativity, parents need to emphasize obedience control. If the society is one in which physical fighting is needed for survival, either individual fighting or fighting as part of an army (as in ancient Sparta or certain nonliterate bellicose societies), spanking and harsh discipline are likely to prepare children for survival in such a society. The available evidence, however, suggests that this will be accomplished at the cost of more unwanted aggressive and violent behavior and psychological well-being (Lansford, 2010). Rohner et al. (1991) found that even in St. Kitts-a society in which the cultural norms strongly approve of spanking-the more spanking actually experienced, the higher the incidence of psychological problems such as low self-esteem, emotional instability, and emotional unresponsiveness. In Jamaica, another nation where spanking is the norm, Smith, Springer, and Barrett (20 1 0) found that the overwhelming majority of adolescent respondents experienced spanking and that it was associated with adverse psychological and behavioral consequences.
Cross-cultural comparative studies show that societies that are high in use of spanking also tend to be high in frequency of warfare and violent interpersonal relationships (Lansford, 2010; Otterbein, 1974; Russell, 1972) including violence against spouses (Levinson, 1989). Lansford's review of the cultural context theory concludes that "Although corporal punishment is generally related to more behavior problems regardless of cultural group, this association is weaker in countries in which corporal punishment is the norm. Yet cultures in which corporal punishment is the norm also have higher levels of societal violence" (Lansford, 2010, p. 105).
In the context of a post-industrial society, deductions from cultural relativity theory lead to the opposite hypothesis than has been argued on the basis of the presumed cultural appropriateness of spanking. In a post-industrial society, children need experience in negotiation and cognitive methods of influencing others rather than experience in use of physical force to reach desired outcomes. To the extent that is the case, cultural relativity theory suggests that optimum socialization for life in such a society should emphasize reasoning and explanation and avoid spanking. Moreover, a high level of cognitive ability and higher education are critical for optimum adaptation to a post-industrial society, and the results in Part III of this book show that spanking hinders achievi11g those characteristics.
Social disorganization theory. Social disorganization theory leads to the proposition that, in the context of a high level of social disorganization, parents need to compensate for the lack of structure and social control by a more structUred pattern of family relations and control than would be optimum in other settings in order to increase the probability of the child developing clear conceptions of right and wrong and avoiding demoralization. Neither our theoretical perspective nor the results of this study contradict that proposition. Our results do suggest, however, that a high level of supervision and control should not include control by hitting the child for misbehavior. In fact, as Sears et al. (1957) have shown, children are more likely to develop an internalized conscious and control if parents monitor.
Deterrence theory. This theory leads to the proposition that, because of the intense peer pressure to participate in substance abuse and crime faced by children living in poverty-stricken neighborhoods, to protect the child there need to be strong penalties for transgression and that this is best accomplished by spanking. However, deterrence theory also holds that for punishment to deter, the offender must perceive it as certain and severe. Neither of these applies to children at the age in which delinquency typically occurs because parents probably do not know about most of the child's transgressions. In addition, in the context of the fights and violence that pervade in neighborhoods where poverty is concentrated, being slapped by a parent, or even paddled with a hairbrush, may not be very severe.
Policy Implications
It would be unrealistic not to recognize that non-spanking could be a disaster if parents take it to mean being permissive and ignore a child's misbehavior. Children do need firm but loving control, as exemplified in what Baumrind (1991b) calls an authoritative parenting style. That said, spanking is not a necessary part ofBaumrind's authoritative type. Parents do not need to spank to be authoritative. If social norms change enough to prevent parents from spanking, some parents will give up trying to correct misbehavior, but that is likely to be rare because most parents are deeply committed to producing responsible children. If they stop spanking while continuing to use the same nonviolent methods of correction and control as before, the evidence in this book and many other studies suggest that they will be more effective parents just by leaving out the spanking. This is also the implication offour studies in Sweden (Durrant & Janson, 2005).
Another possible danger is that efforts to correct misbehavior will shift from physical aggression (spanking) to verbal aggression (attempts to cause psychological pain by insulting or depreciating the worth of the child). That would truly be disastrous because psychological attacks have more harmful psychological effects than physical attacks (Vissing et al., 1991). As a consequence, educational programs to end spanking must give equal weight to discouraging or ending verbal attacks on the child, or, as in the Swedish no-spanking law, avoiding all humiliating treatment of children.
There is a tragic irony in the belief, often found in poor urban neighborhoods, that spanking is necessary to bring up children to resist the drugs, crime, and violence. From a societal perspective, it is ironic because spanking is one of the bases of cultural legitimization of violence (see Chapter 5 and Baron & Straus, 1989; Straus, 1991). From a family-level perspective, spanking tends to undermine the very things parents, regardless of culture, most want-a child whose behavior is governed by internalized standards and "stays out of trouble." Instead, spanking teaches that force rather than reason prevail, that even good people (i.e., one's parents) hit others. Moreover, the use of spanking inhibits the development of internalized controls and conscience (Sears et al., 1957). As shown in Chapter 8 on spanking and the child-to-mother bond and Chapter 9 on spanking and risky sex, spanking also tends to undermine the bond between parent and child that is crucial if parents are to exercise influence that will protect children from drugs and crime once they are too big to spank.
15 Spanking High-Risk Children and Adult Crime
Part II, III, and IV of this book, especially the longitudinal studies in Chapters 6 and 10 and the longitudinal studies by others on the effects of spanking, such as the 16 longitudinal studies reviewed in Chapter 19 lead to the ironic conclusion that spanking to correct misbehavior, although it may halt the immediate misbehavior, is associated with an increase in subsequent misbehavior. Moreover, there is a remarkably high level of consistency between studies on the harmful effects of spanking. Among the 112 tests of long-term effects of spanking reviewed by Gersh off (2002), 108 or 96% showed that spanking was associated with later problematic behavior such as less moral internationalization, poor parent-child relationships, delinquency, and mental health problems.
Despite the mass of evidence just cited, when the outcome variable is crime as an adult, additional research is needed. One reason is that most of the previous research has investigated the relation of spanking to aggression and other antisocial and criminal behavior of children. Perhaps even more important, the few studies that investigated the link between spanking and adult crime have been limited by being cross-sectional studies or by an inadequate longitudinal design. Some longitudinal studies could not provide information about spanking per se because the spanking data was analyzed as part of a harsh parenting factor or index that included other undesirable parent practices such as psychological aggression against the child, for example, Eron et al. (1991). The aggression by these children later in life might be the result of the other adverse parent behaviors in the index. Other longitudinal studies, although they measured criminal behavior years after the spanking, did not control for the level of antisocial behavior that led the parent to spank at Time 1. Thus they could not determine if the spanking led to a subsequent change, for better or for worse. They are little better than cross-sectional studies in determining whether spanking causes the higher rate of crime later in life. Aggression and antisocial behavior tend to be persistent characteristics. Thus, the higher rate of crime could be another manifestation of the behavior problems that led the parents to spank. Because antisocial behavior causes parents to spank, it is crucial to control for the level of antisocial or delinquent behavior at Time 1. Doing so allows a
John P. Colby, Jr. and Rose A. Medeiros are coauthors of this chapter.
Because antisocial behavior causes parents to spank, it is crucial to control for the level of antisocial or delinquent behavior at Time 1. Doing so allows a researcher to determine if the link between spanking and antisocial behavior is because (1) when parents spank to correct antisocial behavior it actually increased the probability of antisocial behavior, (2) antisocial behavior on the part of children increases the probability of parents resorting to spanking, or (3) as we believe, both. Longitudinal research can determine this by measuring the level of antisocial behavior that presumably led to the spanking and then measuring the antisocial behavior later to see if, as most people believe, the spanking taught them a lesson and if those who were spanked are less likely to commit crimes or if, as we hypothesize, their behavior, on average, gets worse. Statistically, this is called the causal order problem. If experiments are not possible or practical, the best way of dealing with the causal order problem is probably to statistically control for the amount of antisocial behavior at Time 1. Doing that results in the measure at Time 2 being a measure of the amount of change in antisocial behavior and whether, as we said, the change is for the better or for the worse.
The few studies that have dealt with the causal order problem by controlling for the Time 1 antisocial behavior, which could have led the parents to spank such as the study in Chapter 6, have not investigated 'whether spanking is a risk factor for crime by adults. The longitudinal research reported in this chapter helps fill this gap because it provides data on some· important questions:
* Is spanking to correct antisocial behavior as a child associated with an increase in the probability of adult antisocial behavior in the form of crime?
* Does the link between spanking as a child and crime as an adult apply to property crimes as well as violent crime?
* When parents are emotionally supportive and monitor the child's activities, does that mitigate the effect of spanking?
* Does a relationship between spanking and crime as an adult apply to women as well as men, and does it apply regardless of whether the child grew up in a high-risk situation as indicated by a low-education mother, being the child of a very young mother, not having a father present in the household, having many siblings, and having low emotional support and low parental monitoring of the child's behavior?
The sample for this study was deliberately designed to ensure that there are sufficient children in these high crime risk groups to enable an adequate statistical analysis.
Sample and Measures
Sample
The data for this chapter comes from the National Longitudinal Survey of Youth. That study oversampled minorities and low-income mothers in order to make it possible to have enough young adults from Black, Hispanic American, and low socioeconomic status backgrounds to investigate issues that are particularly important for these groups. Thus, it provides an opportunity to test the hypotheses that spanking is related to later criminal behavior among the very groups who are most committed to the idea that strong discipline (by which they mean spanking) is necessary to make sure that children resist the opportunities and pressures to engage in crime that are so often prevalent in their neighborhoods. In short, the theory underlying the research in this chapter is that spanking as a method of discipline, rather than protecting children from engaging in criminal activity, increases the chances of that happening.
The characteristics that led us to identify the 468 children studied as highrisk for later life crime were not just based on race or ethnicity, which were 51% Black and 21% Hispanic, but a number of risk factors for crime. These risk factors correspond with the characteristics in the Moore, Vandivere, and Redd (2006) Sociodemographic Risk Index. The average level of education of the mothers was less than a high school education-11 years, and the average age of the mothers at the time of the birth of the children in this sample was 17.3 years. Moreover, the father was present in the household at the time of interview for only 36% of the children. The average number of children in the household in 1986, when this data was collected, was high (3.6).
Measures
Spanking. Spanking was measured by interviews with the mothers when the children were between 8 and 13 years old. The mothers were asked, "Sometimes kids mind pretty well and sometimes they don't. About how many times, if any, have you had to spank your child in the past week?"
Crime. The data on crime were obtained by interviewing the children when they were between 18 and 23 years old. They were asked if, in the past 12 months, they had done any of the following 14 criminal acts:
Violent Crime
* Hurt someone badly enough to need bandages or a doctor
* Gotten into a fight at school or work
* Used force to get money or things from someone else
* Hit or seriously threatened to hit someone
* Attacked someone with the idea of seriously hurting or killing them
Property Crimes
* Taken something from a store without paying for it
* Other than from a store, taken something not belonging to you that was worth less than $50
* Other than from a store, taken something not belonging to you that was worth $50 or more
* Intentionally damaged or destroyed property that did not belong to you
* Tried to get something by lying to someone about what you would do for him or her (tried to con someone)
* Taken a vehicle without the owner's permission
* Broken into a building or vehicle to steal something or to just look around
* Helped in a gambling operation, like running numbers or books
* Knowingly sold or held stolen goods
The overall crime scale included all 14 items. The violent crime scale includes five items; the property crime scale includes nine. The scale score could range from zero (committed none of the 14) to 14 (committed all of the 14). There was no one who committed all14. In addition, we created a subscale for violent crime that could range from 0 to 5, and a score for property crime that could range from 0 to 9.
The study participants were asked only if they had engaged in a specific behavior in the past 12 months, not the number of times they committed a criminal act. Thus, the scores represent the total number of different types of crimes they committed, rather than the number of crimes. There is evidence that the number of different types of crime is a better measure of criminality than the frequency of criminal acts (Sweeten, 2009).
Antisocial behavior. Antisocial behavior was the proxy for criminal behavior at Time I. It was measured in this study by a six-item subscale of the National Longitudinal Survey of Youth behavioral problems scale. Parents were instructed, "As you read each sentence, decide which phrase best describes your child's behavior over the last three months." Parents were asked if each of the following were often true, sometimes true, or not true. "He or she cheats or tells lies"; "he or she bullies or is cruel or mean to others"; "he or she does not seem to feel sorry after he or she misbehaves"; "he or she breaks things on purpose or deliberately destroys his or her own or another's things"; "he or she is disobedient at school"; and "he or she has trouble getting along with teachers."
Controls. The data available enabled us to control for ten child and family characteristics that are associated with both spanking and with crime. These controls are very important because, if we fmd a relationship between spanking and crime, they might be the underlying reason for that relationship, not the effect of spanking per se. These variables are mother's education, mother's age when she gave birth to the child, child's race, child's gender, number of other children in the household, whether the child's father was present in the household, and parental emotional support and monitoring of child behavior. Each of these was used in two different ways: first, as control variables to see if the effect of spanking is in addition to those seven variables and to answer the question of whether spanking makes a unique contribution to explaining crime. We also used them as moderator variables to answer the question of whether the relation of spanking to crime applies to different levels or categories of these variables; for example, when parental warmth and support is high, is there a relation between spanking and crime or, among race or ethnic groups, is there a relation between spanking and crime within each of the groups (Black, White, Hispanic)?
Prevalence of Corporal Punishment and Crime
How Much Spanking and Who Does It
The children in this sample ranged from 8 to 13 years old the year when spanking was measured. This is well beyond the peak years for spanking documented in Chapter 2. Despite that, 29% were spanked during the week prior to the Time 1 interview in 1986, including 6% who were spanked 3 or more times that week. This is consistent with the ethnic or racial composition and the socioeconomic status of the sample. The Appendix gives other information on the use of spanking by the mothers in this sample.
Although mothers with less education and mothers with more children spanked more, the differences were not large enough to be statistically significant. This may be due to the limited education of most of the mothers in this sample. Slightly more than one half of the mothers (52.2%) had not completed high school, and only 13.5% of mothers had any education beyond a high school diploma. There was unexpectedly little difference between race or ethnic groups in spanking, probably because of the low average socioeconomic status of all the race and ethnic groups in the sample. There was also little difference in spanking by father's presence in the household, mother's age at the birth of the child, and the level of parental monitoring.
Of the seven characteristics of the children and families in this study, only the child's antisocial behavior was related to spanking strongly enough to be statistically dependable (see Appendix for statistical details). The average antisocial behavior scale score of the children who were not spanked in the seven days before the interview was 1.6. Among those who were spanked once in the past week, the average antisocial score was 2.4, and among those spanked twice or 3 or more times, the average antisocial scale score was 3.0 and 3.0, respectively. Thus, the more spanking, the more antisocial behavior or the more antisocial behavior, the more spanking. The simple cross-sectional analysis cannot tell us which causes which, or whether (as we believe), there is bidirectional causation. To find out, it is crucial to examine what happens subsequently. That is, when parents spank to correct misbehavior, does it teach the child a lesson in the sense of reducing the probability of subsequent misbehavior, or does it boomerang as was shown in Chapter 6, and as we investigated for this chapter, will it be associated with an increase rather than decrease in the chances of the child becoming involved in crime as a young adult?
Crime
When these children were interviewed as young adults, over one half of the males and almost one half of the young women reported having committed at least one crime in the previous 12 months. Specifically:
* 58.2% of the men and 45.2% of the women had committed at least one of the crimes
* 45.9% of the men and 27.1% of the women had committed a violent crime
* 50.5% of the men and 38.9% ofthe women had committed a property crime
These are extremely high crime rates, but they are consistent with the crime rates found by many other studies of self-reported crime, especially among of youth with similar sociodemographic profiles (Junger-Tas et al., 2003; Nettler, 1984; Thornberry & Krohn, 2000).
Corporal Punishment and Crime as a Young Adult
We examined the relationship between spanking and the three measures of crime: violent crime, property crime, and the combination, which we call the overall crime scale. Using the entire sample, although all three relationships were in the hypothesized direction (the more spanking, the more crime), none were large enough to be statistically dependable. When the hypotheses were tested separately for men and women, however, we found that for men, there were statistically dependable relationships between spanking and crime. For the overall crime scale, the regression coefficient was .20, for property crime it was .23, and for violent crime, .24. Chart 15.1 graphs these coefficients. These are substantial relationships, especially because they indicate a relationship of spanking to crime that is over and above 1 0 key risk factors for later criminal behavior, such as early antisocial behavior by the child and parental emotional support and monitoring of child behavior, and shows that for boys, spanking is related to later crime that is in addition to the effect of the 10 control variables. However, Chart 15.1 also shows that for the girls in this study, there was no relationship between spanking and later crime. The statistical details are in the section of the Appendix for this chapter.
In addition to controlling for six variables, we also looked into whether the relation of spanking to crime applies to different levels or categories of these variables; for example, when parental warmth and support is high, is there a relation between spanking and crime, and is there a relation between spanking and crime within all three race or ethnic groups (Black, White, and Hispanic)? We did not find any of them to make a statistically dependable difference. That is, the relation of spanking to crime that is graehed in Chart 15.1 applies regardless of the mother's education, race or ethnic group, amount of emotional support, and cognitive stimulation provided to the child. As we have seen, only the sex of the child made a difference: Spanking was related to later crime for boys but not for girls. The fmdings on the relation of spanking to crime applied to all the ethnic groups. This is particularly important because it contradicts the claim (discussed in Chapter 14) that spanking is harmless when it occurs in a cultural context in which spanking is the norm, such as in many poor Black neighborhoods.
Chart 15.1 For Men, the More Times Spanked, the More Criminal Acts. For Women, There Is No Relationship*
*Controlling for mother's education, child's race, child's gender, number of other children in the household; whether the child's father was present in the household, the mother's age when she gave birth to the child, and parental monitoring of child behavior.
Other Variables Related to Crime
Although we did not find that the other variables in the study affected the relation of spanking to crime, that does not mean those variables are not risk factors for crime. In this study, as in many other studies, the more monitoring by parents, the less total crime and the less violent crime. When we examined the relationship of the risk factors to crime separately for males and females, emotional supportiveness of the mother and having an older mother were associated with less violent crime for girls, but not for boys. Fpr boys, Hispanic ethnicity was associated with the overall crime scale and monitoring the child's activities was associated with a lower probability of violent crime as an adult. (See Appendix for the statistical details.)
Summary and Conclusions
The characteristics of this sample of 8- to 13-year-old children provided an opportunity to examine whether the adverse effects of spanking apply to children from low socioeconomic status families, many of whom were ethnic minority children and lived in disadvantaged high-violence neighborhoods. These life circumstances are part of the explanation for the high rate of spanking (Winstok & Straus, 20llb), which we and others have found. The high level of spanking, despite the euphemism spanking, describes violent child rearing, and these data suggest it is another part of the explanation for the high crime rate we found for the young adults in this study. Moreover, these results are particularly noteworthy because they are based on a longitudinal study that controlled for 10 other variables that are associated with an increased probability of later criminal behavior. Thus, spanking, rather than helping to protect children from crime, is associated with an increased probability that they will later engage in criminal behavior.
Parents in high-violence neighborhoods tend to believe that immediate compliance is necessary because noncompliance may have dire consequences (Garbarino, Kostelny, & Barry, 1997; McLeod, Kruttschnitt, & Dornfeld, 1994; Staples & Johnson, 1993). Spanking does work in securing compliance in the immediate situation, but it is not more effective than nonviolent methods of correction, (Larzelere, Schneider, Larson, & Pike, 1996). However, parents do not know this, and cultural myths described in Chapter 18 and in Beating the Devil Out of Them (Straus, 2001a, Chapter 10) lead most U.S. parents to erroneously believe that spanking is more effective than other methods of discipline and is therefore sometimes necessary. That belief, in combination with the high priority parents in such neighborhoods give to immediate compliance (Kohn, 1977) compared, for example; with the ability to be self-directed, helps explain their high use of spanking. Instead, for the boys in this study, even after controlling for misbehavior that might have led parents to spank, spanking was found to be associated with an increased probability of crime as a young adult. For the girls, spanking was not associated with either an increase or a decrease in the probability of crime. Therefore, although spanking did not increase the probability of crime later in life by girls, neither did it help them avoid crime. In short, for boys (who are most at risk for criminal behavior), spanking was associated with a worsening of the antisocial behavior for which they were spanked as young children, and it did not protect girls in the sense that spanked girls engaged in less crime as an adult.
Our not finding a criminogenic effect of spanking for girls, was consistent with the results of the study by Boutwell, Franklin, Barnes, and Beaver (2011) described in Chapter 20. That study examined the combined effect of genetic heritage and spanking and found that the effect of spanking on child antisocial behavior applied only to boys. Boutwell et al. suggest that this is because of a genetic difference between males and females. One of several other possible explanations is routine activities theory (Cohen & Felson, 1979; Marcum, 2010; Osgood, Wilson, O'Malley, Bachman, & Johnston, 1996). Women are less likely to engage in activities that facilitate crime because they spend less time with men than men spend with men, and many more men than women commit crimes. Thus women are less likely to interact with and be influenced by a criminal person. In addition, women spend less time in locations such as bars and less time at certain types of sports such as boxing that are high in crime and are frequented by more persons prone to crime than other locations. A flip side of this gender difference is that the increasing equality of women and men in the good things in life also means women gain more equality in exposure to criminogenic situations and hence an increase in crime by women. Dawson and Straus (20 11) tested this theory cross-nationally and found that the more gender-equal the nation, the closer the crime rate of women was to the crime rate· of men. They also found that part ofthe lower male-predominance in crime in more gender-equal societies occurs because male crime rates are lower in gender-equal societies.
Limitations
It can be argued that using antisocial behavior as the Time 1 control for criminal behavior is not appropriate because it is qualitatively different than adult criminal behavior (Kandel, 1991). On the other hand, there is a growing body of research showing that early antisocial behavior is associated with later criminal behavior and violent behavior (Loeber & Farrington, 1995; Stouthamer-Loeber, Loeber, Huizinga, & Porter, 1997; Tremblay, 2006).
The measure of spanking used for this study was whether the child was spanked in the previous week. This gets more accurate information about spanking than a longer recall period (see Chapter 2), but it is also a limitation. It is not possible to differentiate between children who were never spanked and those who were not spanked during the seven days before the Time 1 interview. The latter group might include a substantial number of children whose parents do spank, but not on a weekly basis. Future research can deal with this limitation by using the new response categories by using the new response categories for the Conflict Tactics Scales (Moore, Straus, & Winstok,2013) which have also been used in the Dimensions Of Discipline Inventory( Straus & Fauchier, 2011). These make it is possible to have both the greater accuracy of asking about spanking in the past week and simultaneously determine how often spanking occurred in the other weeks of the year. They would also permit determining whether the never spanked fare better than children who were rarely spanked. It is possible that there is a threshold in the use of spanking, up to which point spanking does not result in maladaptive social and psychological outcomes for children. This is the belief of those who defend moderate use of spanking (Baumrind et al., 2002; Friedman & Schonberg, 1996a) but, except for the fact that the harmful effects of spanking are in the form of a dose-response (see, for example, Chapters 6, 7, and 1 0) that belief is based on cultural tradition rather than empirical evidence.
Another limitation is that there is no infoniiation on spanking by the fathers ofthese children. It was not practical to get that information for a number of reasons, but especially because two thirds of the children were in households where the fathers did not reside. Nevertheless, for the one third who ,:were living with their father, this measure erroneously includes in the not-spanked group children who were not spanked by their mother, but were spanked by their father.
Implications of Study Findings
As pointed out in several chapters, the findings of most of the research showing a link between spanking and adult crime could be spurious because those studies did not control for the misbehavior that led parents to spank. Thus, the adult criminal behavior of the spanked children, rather than being an effect of the spanking could be just another manifestation of the tendency to antisocial behavior that led the parents to spank. The research reported in this ch~pter and in Chapter 6 (on the boomerang effect of spanking) and the studies reviewed in Chapter 19 makes that a less likely explanation because they controlled for antisocial behavior that might have caused parents to spank at Time 1.
The controls used for this study also included the amount of parental emotional support provided, cognitive simulation, and parental monitoring of the child's behavior. These are important protective factors. It is sometimes argued that what seems to be an adverse effect of spanking is really an effect of bad parenting. This study shows that, important as those protective factors are, after taking them into account, spanking still has a net effect; that is, spanking explains child-to-child differences in crime as an adult that are in addition to what is explained by the presence or absence ofthese protective factors.
Our results showing that the spanking is related to adult crime for children from low socioeconomic status families, many of whom were living in highviolence neighborhoods, does not contradict the fact that an important part of the explanation for crime by youth lies in the social setting-for example, living in areas of concentrated poverty, family and neighborhood disorganization, and the perception or reality of being locked out oflegitimate employment (Sampson, Raudenbush, & Earls, 1997; Wilson, 1996). Nor does it contradict cultural explanations such as the code of the streets (Anderson, 1999). It adds one more risk factor to life circumstances that are already overflowing with risk factors for crime. Although respect for the cultural autonomy of racial and socioeconomic groups is an important principle (Polite, 1996; Rohner et al., 1996), when an element of that culture, such as spanking, even though undertaken with the intent of enhancing the well-being of children and protecting neighborhoods from crime, has the opposite effect, we believe the effort must be made to change that aspect of the culture.
16 Sexual Coercion and Sexual Assault
The chapters up to now have shown that corporal punishment by parents is associated with an increased risk of antisocial behavior and several types of crime. But why should something that parents do to produce a well-behaved child be associated with crime later in life? The chapter on the relation of spanking to physically assaulting a partner presented evidence on that issue. It showed that being spanked is a risk factor for physically assaultive behavior because spanking is associated with an increased probability of approving violence, depression, and conflicts)with a partner.
This chapter also examines another possible linking process: Whether the relationship between spanking and crime occurs because spanking increases the probability of antisocial personality traits. This follows up what was reported earlier: That spanking has the long-run side effect of increasing the probability of the child developing antisocial personality traits and behavior (Chapter 6). Manuel Gamez-Guadix and Scott Hershberger are coauthors of this chapter.
The chapter also takes into account two other victimization experiencessexual abuse and neglect. The issue investigated is whether spanking, neglect, and sexual abuse are each uniquely associated with an increased probability of sexually coercing and sexually assaulting a partner. Although the study is about the relation of three kinds of victimization to sexual coercion, somewhat more attention is presented about spanking because spanking is the focus of the book. The questions to be addressed are:
* What percent of male and female university students engage in verbally coercing sex and physically forcing sex on a dating partner?
* Are being spanked and experiencing relatively minor forms of neglect, such as not being comforted when distressed, associated with an increased probability of antisocial traits and sexual violence, and does this apply to women as well as men?
* Are the effects of spanking and minor neglect unique in the sense of being in addition to the effect of having been sexually abused?
* Is part of the reason why spanking, sexual abuse, and neglect are related to sexual coercion because these three victimization experiences increase the probability of antisocial traits and behavior? Prevalence of Sexual Coercion and Gender Differences
Sexual coercion is a widespread problem with deleterious short- and long-term consequences (Demaris, 2005; Kernsmith & Kernsmith, 2009; Temple, Weston, Rodriguez, & Marshall, 2007; Tewksbury, 2007). A number of terms in addition to sexual coercion are used to refer to the range of physical or nonphysical behaviors to coerce a partner into unwanted sexual activity. These include sexual assault, sexual aggression, sexual violence, sexual pressure, and date or acquaintance rape. Sexual coercion can be verbal coercion such as insisting on sex when the partner does not want to or threatening an unwilling partner to gain sexual contact against the partner's will. Or it can be physical coercion, which we will call sexual assault.
Research on sexual coercion has consistently found high rates in college and community samples (Basile, Chen, Black, & Saltzman, 2007; Dekeseredy & Schwartz, 1998; Dekeseredy, Schwartz, & Tait, 1993; Koss, Gidycz, & Wisniewski, 1987; Tjaden & Thoennes, 2000). Most of the studies have asked only about men sexually coercing women (e.g., Koss et al., 1987), although a growing body of literature has showed that women also sexually coerce (Anderson & Struckman-Johnson, 1998; Davies, 2002; Hettrich & O'Leary, 2007; Kar & O'Leary, 2010; Muehlenhard & Cook, 1988; Panuzio & DiLillo, 2010). There has also been considerable research on the etiology of sexual coercion, but again, primarily studies of male perpetrators.
The prevalence of perpetration of sexual coercion and assault by men has been widely researched and has been extensively reviewed (Hines, 2007; Spitzberg, 1999). Therefore, this review focuses on studies that present rates for both men and women in dating relationships. All these studies used student samples. Examples include a study by Struckman-Johnson (1988) that found that 10% of men and 2% of women in a sample of university students admitted to forcing a date into sexual activity at least once in their lifetime, including psychological pressure and physical tactics to engage in sexual intercourse, but the study did not provide separate rates for verbal coercion and sexual assault. O'Keefe (1997) found that 12.3% of male students and 3% offemales reported forcing sex in their dating relationships, but this study did not specify what was meant by force. A study that did distinguish between verbally coercing and physically forcing sex is Poitras and Lavoie (1995). They found that
* 12% of males and 6.3% of females used verbal sexual coercion
* 0.3% of males and none of the females misused authority
* 2.3% of males and none of the females used alcohol or drugs
* 3.9% of males and 0.3% offemales threatened to or used physical force
Hines and Saudino (2003) found that 29% of male and 13% of female university students reported one or more acts of sexual corrcion including verbal and physical tactics,in the previous 12 months. In short, these studies show that in relationships with acquaintances both men and women use a broad range of sexually coercive behaviors, but many more men than women do so.
Prior Victimization and Perpetration of Sexual Coercion
We decided to investigate the relation of spanking, minor neglect, and sexual abuse as a child or adolescent to sexually coercing or assaulting a partner because, on theoretical grounds and on the basis of the studies reviewed below, each of these three childhood maltreatment experiences could contribute a distinctive piece to the explanation of sexual coercion.
* Minor neglect can contribute to less moral internalization and less selfcontrol, including control of aggression (see review below). * Spanking can contribute examples of using physical force and violence to control the behavior of another person. * Sexual abuse history can provide an example of coercion to obtain sex.
Together, we believe they are almost a perfect storm that increases the probability of sexual coercion. Moreover, in addition to these direct relationships, the three types of victimization increase the probability of antisocial traits and behavior, which then further increases the probability of sexual coercion.
Previous studies ofthe relation of childhood victimization to sexual coercion and assault have a number oflimitations. For example, some studies combined several types of prior victimization into a single variable (e.g., Wolfe, Scott, Wekerle, & Pittman, 2001), thus providing no information on the separate effect of each type of childhood victimization. Most studies of sexual coercion have considered exclusively male samples. As mentioned previously, there are some studies of sexual coercion by women. However, because samples and measures differ from those used in studies of sexual coercion by men, they cannot accurately compare male and female prevalence rates and etiology. That requires studies that include both men and women in the same study, as in this chapter. Finally, many of the studies of the relation of prior victimization to sexual coercion examined the effect of only one type of victimization in isolation from other types (Loh & Gidycz, 2006). Because poly-victimization is common (Finkelhor, Ormrod, & Turner, 2007), research on a single type of victimization tends to overestimate the effect of that one type of victimization. There is a need to determine the relation of different forms of victimization to sexual coercion while controlling for each of the other types of victimization in order to estimate the unique effect of each. This chapter does that by considering the combined effect and the net effect of three childhood and adolescent types of victimization: neglect, spanking, and sexual abuse in a sample of men and women. In addition, because a number of studies have shown that antisocial traits and behavior are important in the etiology of sexual coercion by men against women (Knight & Sims-Knight, 2004; Malamuth, Linz, Heavey, Barnes, & Acker, 1995), we investigated the extent to which antisocial traits and behavior explains the link between earlier victimization and sexual coercionthat is, whether the relation of the three types of victimization (minor neglect, sexual abuse, and spanking) to sexual coercion occurs because they increase the likelihood of developing antisocial traits and behavior. The following sections review the research on the relation of the three types of victimization to sexual coercion.
Why Focus on Spanking and Neglect?
There has been little research on whether spanking a child and minor neglect (such as not helping a child experiencing a difficulty) are risk factors for sexual coercion. Perhaps this is because it has been presumed that spanking and minor neglect constitute a much less severe level of victimization than sexual abuse. However, there are important reasons to think that spanking and minor forms of neglect increase the risk oflater sexual coercion.
One reason to focus on these presumably less serious types of victimization is the well-established public health principle that mitigating a frequently occurring risk factor with a low effect size (such as spanking a child) can result in a greater reduction in the prevalence of a disease than mitigation of a relatively rare risk factor with a large effect size, such as physical abuse (Rose, 1985). Thus, if spanking and minor neglect are found to be related to sexual coercion, reducing these two risk factors could be a major step in preventing sexual coercion because they are such prevalent forms of child victimization. This is because, as shown in Chapter 2 on the use of spanking in the United States, over 90% of toddlers are spanked. Similarly, studies of minor forms of neglect by parents such as failing to console a child who is sad or in physical pain, have found very high rates (Straus et al., 1998; Straus & Savage, 2005). A cross-national study of minor neglect in a large sample of university students used a scale of eight neglectful behaviors, such as "Did not comfort me when I was upset." It found that about one half of the students experienced one or more of eight neglectful behaviors, including 12% who reported two of the eight, and another 12% experienced three or more (Straus & Savage, 2005).
Neglect History and Sexual Coercion
Neglect is "behavior by a caregiver that constitutes a failure to act in ways that are presumed by the culture of a society to be necessary to meet the developmental needs of a child and which are the responsibility of a caregiver to provide" (Straus & Kaufman Kantor, 2005). We located only one study investigating the link between neglect and sexual coercion. DeGue and DiLillo (2004) found that a history of physical or psychological abuse was related to sexual coercion, but that a history of neglect was not. Nevertheless, there is both a theoretical and an empirical basis for hypothesizing a link between neglect and sexual coercion and assault.
Tremblay (2003) argues that a responsive parent and consistent discipline enable children to learn nonviolent strategies for achieving their goals and expressing anger. A child who does not have a responsive caregiver and consistent guidance may not adequately learn nonviolent strategies and, consequently, is more likely to engage in maladaptive coping strategies and an aggressive pattern of interaction (see also Chapple, Tyler, & Bersani, 2005; Gottfredson & Hirschi, 1994; Spitz, 1959). Consistent with this theory, a number of empirical studies have found that neglect constitutes a risk factor for aggressive and antisocial behavior (Chapple et al., 2005; Hildyard & Wolfe, 2002; Horwitz, Spatz Widom, McLaughlin, & Raskin White, 2001). There is also a study of the link between minor neglect and physical violence against a dating partner using the same measure of neglect as used for this chapter and found that even this relatively minor level of neglect experienced as a child was significantly related to assaulting and injuring a dating partner (Straus & Savage, 2005}.
One can conclude from these studies that there is a large body of evidence showiJ/g that neglect is a risk factor for later aggressive and criminal behavior, including physical aggression against dating partners. Therefore, even though one study that investigated sexual coercion did not find a relationship between neglect history and sexually coercing a partner, we hypothesized that:
The more minor neglect experienced, the greater the probability of engaging in sexual coercion of a dating partner.
Spanking and Sexual Coercion
The chapters in Part II and the previous chapters in this Part and Chapter 19 show that spanking is associated with an increased probability of antisocial behavior as a child and criminal behavior as an adult. Several mechanisms that might explain the association between spanking and aggressive antisocial and criminal behavior were identified in a review by Simons, Burt, an.d Simons (2008). According to social learning theory (Bandura, 1977), children spanked as a means of discipline learn through the example of this legal and morally correct behavior by their parents that aggression is a socially acceptable and effective means for getting others to do what you want. Consistent with this, in the United States and around the world, spanking is associated with an increased probability of physical violence against a marital or dating partner (Chapters 12 and 13).
Although spanking produces compliance in the short term, the studies in this book have found that the long-term effect is the opposite: Spanking increases the probability of deviance, including antisocial behavior. The association between spanking and antisocial behavior was found by the studies in the chapters on spanking and child antisocial behavior (Chapter 6), impulsive spanking (Chapter 7), spanking and the child-to-mother bond (Chapter 8) and a large number of studies reviewed in the chapter on spanking and societal-level rates of crime and violence (Chapter 19). Gershoff's meta-analysis (2002) included 40 tests of the hypothesis that spanking is associated with an increased probability of aggressive and delinquent behavior. Of nine tests of the relationship between spanking and subsequent abuse of the victim's own children or partner, all nine found the hypothesized relation.
Given the consistency of the research showing that spanking is related to aggression and crime, it is not surprising that it has also been. found to be related to sexual coercion of women (DeGue & DiLillo, 2004; Malamuth, Linz, Heavey, Barnes, & Acker, 1995; Simons et al., 2008). On the basis of both theory and those empirical results, we hypothesized that spanking as a child is associated with an increased probability of perpetration of sexual coercion. Moreover, because Chapters 12 and 13 show that the relationship between spanking and assaulting a partner applies to both men and women, we hypothesized that:
Spanking experienced as a child is associated with an increased probability of sexual coercion by men and women in this study.
Sexual Abuse History and Sexual Coercion
Previous reviews have shown that experiencing sexual abuse is associated with a wide variety of problematic behaviors. Because previous reviews are available (Berliner & Elliott, 1996; Kendall-Tackett, Williams, & Finkelhor, 2001) and because of space limitations, we mention only studies that examined the relation of sexual abuse to sexual coercion. Loh and Gidycz (2006) found that men with a history of childhood sexual victimization were over 6 times more likely to perpetrate sexual assault as adolescents and adults. Lyndon, White, and Kadlec (2007) found that males who used force to gain sexual contact reported significantly more childhood history of sexual abuse compared with both men who used manipulation and men who reported engaging in only consensual sex. Sexual abuse history has also been found to be a predictor of perpetration of sexual coercion by women. Krahe, Waizenhofer, and Moller (2003), for example, found that the probability of sexual coercion of a male was 2.62 times higher for females who reported childhood sexual victimization than for females without a history of sexual abuse. Based on these results, we hypothesized that:
A history of sexual abuse is related to sexual coercion by both the male and female students in this study.
Antisocial Traits and Behavior and Sexual Coercion
The link between an antisocial orientation and sexual coercion against women has been extensively studied in both criminal and noncriminal samples of men (Abbey & McAuslan, 2004; DeGue & DiLillo, 2004; Knight & Sims-Knight, 2004; Malamuth et al., 1995; Simons et al., 2008). The related concept of psychopathy has also been found to be associated with sexual coercion (Hare, Clark, Grann, & Thornton, 2000; Knight, 2006). These studies leave little doubt that antisocial traits and behavior are associated with sexual coercion by men but do not answer the question of whether this relationship also applies to women. We therefore tested the hypothesis that:
Antisocial personality and behavior is associated with an increased probability of sexual coercion by both the male and female students in this study.
Sample and Measures
Sample
The sample for this study is from the International Dating Violence Study that also provided the data for the chapters on spanking in world perspective (Chapter 3) and spanking and partner violence among university students across the globe (Chapter 13), namely 14,252 university students in 32 nations. The study and the sample are described in Chapter 3 on the worldwide prevalence of spanking (in Rebellon et al., 2008; Straus, 2008b, 2009b). Most of the data were obtained by administering a questionnaire during regularly scheduled classes. Because this study is on issues in which gender differences are crucial, the analyses either controlled for gender or were replicated for male and female students.
Measure of Sexual Coercion
Sexual coercion and assault of a dating partner was assessed using the Sexual Coercion Scale of the revised Conflict Tactics Scales or CTS2 (Straus et al., 1996). Example items are: "Used threats to make my partner have sex" and "Used force (like hitting, holding down, or using a weapon) to make my partner have oral or anal sex." Participants who reported perpetrating one or more of the verbal coercion items in the past year were coded as 1, and all others as 0. The same procedure was used to identify participants who had committed a sexual assault by physically forced sex. Because research suggests that individuals who use nonphysical coercion to gain sexual contact differ in several ways from individuals who use physical force (DeGue & DiLillo, 2004; Lyndon et al., 2007), these two scores were used to create a Sexual Coercion Severity typology that enables separately analyzing verbal sexual coercion and physically forced sex. The typology classified each participant into one of three mutually exclusive categories: 0 = no sexual coercion, 1 = verbal sexual coercion without physical force, and 2 = physically forced sex.
Measures of Neglect, Sexual Abuse, and Spanking
The measures of minor neglect, sexual abuse history, and spanking are from the Personal and Relationships Profile (Straus et al., 2010; Straus & Mouradian, 1999). The response categories for the questions making up all the scales are (1) strongly disagree, (2) disagree, (3) agree, and (4) strongly agree.
Minor neglect. The Personal and Relationships Profile includes an eight item short form of the Multidimensional Neglectful Behavior Scale (Straus, Kinard & Williams, 1995). It has two items to measure each ofthe four dimensions of neglectful behavior: cognitive ("My parents did not help me to do my best"), supervisory ("My parents made sure I went to school"), emotional ("My parents did not comfort me when I was upset"), and physical ("My parents did not keep me clean"). The scale was scored by adding all the items to which the participant agreed or strongly agreed. Thus, the scores indicate the number of different types of minor neglect experiences that each respondent experienced as a child. The Neglectful Behavior Scale has demonstrated good cross-cultural construct validity and reliability, with an overall alpha of . 72 (Straus, 2006). For this study, the alpha coefficient of internal consistency reliability was .70 for males and . 71 for females.
Spanking. The Personal and Relationships Profile includes the question "I was spanked or hit a lot by my parents before age 12." Participants responded using the 1 (strongly disagree) to 4 (strongly agree) categories described earlier.
Sexual abuse history. Sexual abuse was assessed using the eight-item sexual abuse history scale of the Personal and Relationships Profile. This scale includes questions on whether participants had experienced contact and/or noncontact sexual abuse by family members, peers, and/or nonfamily adults. The scale asks about two behaviors: "Made me look at or touch their private parts (sex organs), or looked at or touched mine" and "Had sex with me (vaginal, anal, or oral)." Each of them was asked for the following four situations: perpetration before age 18 by an adult family member, by another child in the family, by a nonfamily adult, and by nonfamily children. The total score was computed by adding the items to which the respondent marked agree or strongly agree. Thus, the score on the sexual abuse history scale corresponds to the number of different experiences of sexual abuse that each participant experienced as a child and/or adolescent, with a maximum of eight experiences. Strong internal consistency reliability was found for previous samples (Straus & Mouradian, 1999). For this study, alpha coefficient was .80 for males and .79 for females.
Antisocial Traits and Behavior
This scale is derived from the DSM-IV (American Psychiatric Association, 1994). Although the questions were derived from the DSM-IV definition of antisocial personality disorder, this scale was not designed as a diagnostic tool. Because the DSM-IV definition of antisocial personality disorder includes criminal behavior, both before and after age 15, the scale combines the nine antisocial personality traits scale and the criminal history scale of the Personal and Relationships Profile (Straus et al., 2010). Examples of the nine antisocial personality trait items are "I often lie to get what I want" and "I don't think about how what I do will affect other people"). Each of the four criminal behaviors are asked for "before age 15" and for "since age 15." Examples of the eight criminal history scale items are "Before age 15, I stole money from anyone, including family" and "Since age 15, I have physically attacked someone with the idea of seriously hurting them." The 17 items comprising both subscales were summed and divided by the number of items to obtain a mean antisocial traits and behavior score. The internal consistency reliability (Chronbach's1 a) was .82 for males and .81 for females. Information on the construct and concurrent validity of the antisocial traits and behavior subscale is in Hines and Straus (2007). To examine the extent to which medium and high levels of antisocial traits and behavior are associated with an increased probability of sexual coercion, we coded the antisocial traits and behavior scores into three categories: 1 = low antisocial traits and behavior (scores below the 25th percentile), 2 =medium antisocial traits and behavior (scores between the 25th and 75th percentile), and 3 =high antisocial traits and behavior (the top 25% of the distribution).
Control Variables
A relationship between spanking and sexual coercion might just reflect some underlying third V!ifiable. To take that into account, we statistically controlled for five other variables that could be related to both spanking and sexual coercion. Each ofthese is described in more detail in the Appendix. Social desirability. The tendency of some participants in a study to minimize disclosure of socially undesirable behavior could lead them to underreport both spanking and sexual coercion and, therefore, create a spurious correlation. This was controlled using the limited disclosure scale of the Personal and Relationships Profile (Chan & Straus, 2008; Sabina & Straus, 2006; Straus et al., 2010; Straus & Mouradian, 1999).
Family socioeconomic status. Family socioeconomic status was controlled using a three-item scale based on the education of the student's father and mother and family income.
Age. Age was controlled because it is well-established that younger ages are associated with higher rates of violent crime, including partner violence (Stets & Straus, 1989).
Relationship length in months. It was important to control for the length of time the couple had been together because the longer the relationship, the greater the opportunity for sexual coercion to have occurred.
Prevalence and Gender Differences in Sexual Coercion
As has been found in previous studies, a high percent of students sexually coerced a partner. More male than female students sexually coerced a partner, but a &ubstantial percent of women also engaged in sexual coercion. Statistically dependable gender differences were found for the verbal coercion scale and the physical coercion scale and for most of the items making up these scales (see Appendix for the tests of significance).
Verbally coercing sex was reported by 27% of the male students and 20% of the female students. Statistically dependable differences between men and women were also found in most of the verbal sexual coercion items: insisting on sex when the partner did not want to (males= 15.9%, females= 8.2%); threatening the partner to have sex (males= 1.4%, females= 0.8%); insisting on sex without a condom (males= 14.9%, females= 12.9%); insisting on oral or anal sex (males= 11%, females= 3.8%); and threatening the partner to have oral or anal sex (males= 1.2%, females= 1.1). Physically forcing sex was reported by 2.4% of the male students and 1.8% of the female students. Rates for the two specific items were: 1.3% of males and 1.0% of females used force to have sex, and 1.6% of the males and 1.0% of the females used force to have oral or anal sex. Thus, as in other studies that compared sexual coercion by men and women in the same study, both men and women engaged in sexually coercive behavior, but men predominate.
Tests of the Theory
Chart 16.1 summarizes the results of testing the theory that the causes of sexual coercion include having been spanked, and that the effect of spanking is in addition to the effects of having been sexually abused or neglected. Moreover, the theory also specifies that one of the reasons spanking is related to sexual coercion is because, as was shown in Chapter 6, spanking increases the probability of antisocial traits The lines in the chart (called paths) show the relation of each of the three risk factors to antisocial traits and the relationship of antisocial traits to sexual coercion.
Chart 16.1 gives only the most directly relevant results because, ifthe results for all the possible paths were included, there would be so many lines that it would be very difficult to read and understand. Therefore, only paths that are statistically dependable are in the chart, and the paths for the control variables are also not shown. However, the coefficients and tests of significance for all the variables are in the section of the Appendix for this chapter.
The numbers on each path are the amount by which each increase of one point in the variable at the left side ofthe path is associated with an increase in the variable at the right side of the path. These numbers are odds ratios. There are two odds ratios on each path: The one on the right is the odds ratios for women and the one on the left is the odds ratio for men.
Chart 16.1 Spanking Is One of the Risk Factors for Sexually Coercing a Partner by Both Men and Women
Spanking
Verbal sexual coercion. The flrst of the two odds ratios on the path from Spanking (upper left) to Verbal Sexual Coercion (upper right) is marked as ns (not statistically significant) because, for male students, there was not a statistically dependable relationship between Spanking and Verbal Sexual Coercion. But for women, the second odds ratio of 1.10 indicates that each increase of one point on the four-point scale of spanking before age 12 is associated with increasing the odds of Verbal Sexual Coercion by the women in this study 1.10 times (i.e., a 10% increase in Verbal Sexual Coercion).
The paths from Spanking to Antisocial Personality Criminal History, and from there to Verbal Sexual Coercion show that Spanking is also associated with Verbal Sexual Coercion indirectly by increasing the odds of Antisocial Personality Criminal History (center box) by just over 3 times (odds ratio of 3.07) for males and just over 2 times (2.25) for females. The path from Antisocial Personality Criminal History to Verbal Sexual Coercion, in turn, shows that Antisocial Personality Criminal History is associated with increasing the odds of Verbal Sexual Coercion 1.33 times for males and 1.44 times for females. Thus, part of the reason spanking is associated with verbal sexual coercion is because spanking increases the probability of antisocial traits and behavior in general, of which sexual coercion is one example.
Physically forced sex. The diagonal path from Spanking to Physically Forced Sex shows that spanking is directly associated with an increased probability of Physically Forced Sex. The first of the two odds ratio on this path is for the male students. The 1.32 indicates that each increase of one point on the spanking measure is associated with a 33% increase in the probability of men physically forcing sex. The second odds ratio of 1.24 for female students indicates that each increase of one point on the spanking measure is associated with a 27% increase in the probability of the women in this study physically forcing sex.
The path from Spanking to Antisocial Personality Criminal History in the center of the chart indicates that spanking is also related to physically forcing sex, by increasing the probability of Antisocial Personality Criminal History, which in tum, is associated with increasing the probability of physically forcing sex by 1.69 times for males and by 2.08 times for females.
Charts 16.2 and 16.3 graph the relation of spanking to sexual coercion in more detail. The lines for men are higher in both graphs than for women because the men in this study engaged in more sexual coercion than the women. However, Chart 16.2 shows that for both men and women, the more spanking, the higher the percent who used verbal sexual coercion. Chart 16.;3 shows that spanking is associated with an increased probability of physically forced sex by both men and women, but the increased probability is greater for the men.
Chart 16.2 The More Spanking, the Greater the Percent Who Verbally Coerced Sex in the Previous 12 Months
Chart 16.3 The More Spanking, the Greater the Percent Who Physically Coerced Sex in the Previous 12 Months
Sexual Abuse History
Returning to Chart 16.1, the paths from Sexual Abuse History to Verbal Sexual Coercion show that each increase of one point on the four-point Sexual Abuse History scale is associated with a 20% increase in the probability of verbal sexual coercion by men and a 17% increase in the probability of verbal sexual coercion by women.
Having been sexually abused is also associated with an increase in the probability of physically forcing sex. For men, each increase of one point of the Sexual Abuse History Scale is associated with a 1.23 times increase in men physically forcing sex, and a 1.14 times increase in women physically forcing sex.
Prior sexual abuse is also associated with sexual coercion indirectly through an increased probability of antisocial traits and behavior. Thus, as hypothesized, the more sexual abuse experienced, the greater the probability of both verbal and physical sexual coercion.
Neglect History
The ns/ns on the path from Neglect History to Verbal Sexual Coercion indicates that we did not find a direct link between neglect and verbally coercing sex. For physically forcing sex, the path from Neglect History to Physically Forced Sex shows that for men, each increase of one point in the Neglect History scale is associated a 1.2-time increase in the probability of physically forcing sex, but for women, the ns means that we did not find a relationship.
Chart 16.4 The More Antisocial Traits and Behavior, the Greater the Percent Who Verbally Coerced Sex in the Previous 12 Months
Chart 16.5 The More Antisocial Traits and Behavior, the Greater the Percent Who Physically Forced Sex in the Previous 12 Months
For both the men and the women in this study, neglect is indirectly related to physically forcing sex because, the paths from Neglect History to Antisocial Personality Criminal History show that neglect is associated with a 1.29-time increase in antisocial traits by men and a 1.3 7 -time increase for women, and the path from antisocial traits to physically forcing sex shows that being high in antisocial traits and behavior is associate with a 1.62-time increase in the probability of men and a 2.08-time increase in the probability of women physically forcing sex.
Antisocial traits and behavior. Because antisocial traits and behavior appears to play such a central role in explaining sexual coercion by both men and women, we examined this relationship in more detail in Charts 16.4 and 16.5. These plot lines were computed with the value of all other variables controlled. Both charts show that although the probability of both kinds of sexual coercion is higher for male than for . female students, antisocial traits and behavior is associated with a sexual coercion to about the same extent for both the men and women in this study.
Summary and Conclusions
We studied a large multi-nation sample of university students and, consistent with other studies, found high rates of verbally coercing sex and of physically forcing sex. To be more specific, 27% of the male students and 20% of-the female students verbally coerced sex in the previous 12 months, and 2.4% of the male students and 1.8% of the female students physically forced sex during that period. Thus, as in other studies that compared sexual coercion by men and women in the same study (e.g., Banyard et al., 2007; Hines & Saudino, 2003; Kar & O'Leary, 2010; Panuzio & DiLillo, 2010), both men and women engaged in sexually coercive behavior, but the rates for men are higher. The data on physically forcing sex by women are difficult to interpret because the questions did not specify the specific acts of force used. Most of the instances may be manually or orally stimulating the male partner despite physical resistance, whereas forced sex perpetrated by males probably includes penetration attempts and actual penetration.
The results show that, with two exceptions, the relation to sexual coercion of the three risk factors we investigated (spanking, sexual abuse history, and neglect history) are parallel for men and women. The two exceptions are that spanking is related to verbal sexual coercion by women but not by men, and having been neglected as a child is related to physically forcing sex by men but not by women.
Tests of the Theoretical Model
The main objective of the study was to test the theory that the three types of child and adolescent victimization (spanking, neglect, and sexual abuse) are each independently associated with an increased probability of engaging in sexually coercive behavior later in life, both directly and because those victimization experiences increase the probability of the child developing antisocial traits and behavior, which, in tum, increases the probability sexual coercion.
Spanking. Frequent spanking of a young child was found to be an important risk factor for sexual coercion, both directly and because it increases the probability of antisocial traits and behavior. Experiencing frequent spanking was associated with a tripling of the probability of being high in antisocial traits and behavior for men and doubled the probability for the women in this study. In turn, antisocial traits and behavior significantly increased the likelihood of either verbal sexual coercion or physically forced sex for both males and females.
It is important to keep in mind that these results are based on a question that asks about experiencing a lot of spanking. Research shows that the harmful side effects of spanking take the form of a dose-response pattern (Gunnoe & Mariner, 1997; Straus, Sugarman et al., 1997). Therefore, the effect of only occasional spanking is likely to be much lower. Nevertheless, these results are consistent with the results of a meta-analysis of research on spanking (Gershoff, 2002), which found a large and unusually consistent body of research showing that spanking tends to increase the probability of deviant behavior and psychological problems. This includes the longitudinal studies in this book and those summarized in Chapter 19.
Sexual abuse history. For both men and women, a history of sexual abuse was associated with an increase in the probability of sexual coercion and of antisocial traits and behavior. Antisocial traits and behavior, in turn, was associated with a significantly greater probability of either physical or verbal sexual coercion. These findings are in keeping with past research that found that jndividuals with a history of sexual abuse are associated with an increased probability of antisocial and criminal activities (Herrera & McClosky, 2003; Swanston et al., 2003; Widom & Ames, 1994) and sexual coercion against men (Krahe et al., 2003) or women (Loh & Gidycz, 2006; Lyndon et al., 2007).
Neglect history. The hypotheses concerning experiencing mild forms of neglect as a child were only partly supported. Minor neglect was not significantly associated with a higher probability ofVerbal Sexual Coercion. For male students in this study, but not for female students, even this level of neglect was associated with an increased probability of physically forcing sex. The results show that minor neglect is associated with an increased probability of being in the high antisocial traits and behavior category, which in turn is associated with either verbal sexual coercion or physically forced sex for both men and women. This is consistent with the studies reviewed earlier that found an association of neglect with partner violence and antisocial behaviors (see also Chapple et al., 2005).
Antisocial traits and behavior. The results underscore the importance of antisocial traits and behavior in understanding the association between victimization as a child or adolescent and sexual coercion by both men and women. Although the role of an antisocial orientation on sexual coercion has been demonstrated among men in previous studies (Abbey & McAuslan, 2004; Knight, 2006; Knight & Sims-Knight, 2004; Malamuth et al., 1995), the current results indicate that this applies to women as well. However, as in other studies, the rate of antisocial behavior was higher for the men in this study than for the women. In addition, effect size for the association of spanking and sexual abuse with antisocial traits and behavior is higher for the male students than for the female students. This is congruent with research that found that men abused as children are more likely than women who were abused to have antisocial behavior symptoms (e.g., Horwitz et al., 2001).
Limitations
Although this study avoided some of the limitations of previous research, such as restricting the study to male perpetration, failing to control for socially desirable response bias, examining a particular type of prior victimization in isolation from other types of victimization (i.e., not allowing for polyvictimization), and combining several types of victimization into a single variable, there are nonetheless a number of limitations to keep in mind when considering the conclusions and implications.
First, like much previous research on the origins of sexual coercion, the study participants were university students and the data on prior victimization depends on their willingness and ability to recall and report prior victimization. Second, because the rate of most crime, including sexual crime, peaks at about the age of university students, these studies may overestimate the prevalence of sexual coercion in the general population. On the other hand, the opposite bias may be present because university students tend to be from higher socioeconomic status families, and there is abundant evidence that parent-child violence and partner violence become less prevalent with increasing socioeconomic status (see Chapter 2 and for partner violence, Straus et al., 2006; Chapter 12; Straus et al., 1996). Moreover, as pointed out in the Methods section, the study used convenience samples. However, the risk factors for partner violence found in studies using convenience samples of students are almost identical to those found in community sample studies.
An important limitation is that the cross-sectional design does not permit concluding that the relationships found reflect a cause-effect sequence. This is particularly important in the case of spanking because that is something parents typically do to correct misbehavior. Thus, spanking, rather than being a cause of later antisocial traits and behavior, may be a consequence of early child behavior problems, which carries over into adulthood. However, there are at least seven longitudinal studies that show that although misbehavior does caus-e parents to spank and that spanking does result in cessation of the misbehaviot at the time, in the longer run, the U:se of spanking boomerangs in the sense of increasing the probability of subsequent antisocial behavior (Straus, 2001a). At least two of the measures have important limitations. The Neglect History scale measures relatively mild forms of neglect, and even students with high scores would be unlikely to be classified by child protective services as neglected The item used to measure spanking has almost the opposite problem. As pointed out previously, the question refers to having been spanked or slapped a lot before age 12. Thus, the results may not apply to students who were spanked only occasionally. In addition, the question is defective in not using a more specific indication of frequency of spanking, as do the response categories of the parent-child Conflict Tactics Scales (Straus et al., 1998) and the Dimensions of Discipline Inventory (Straus & Fauchier, 2011).
Implications
Although stranger rapes are almost exclusively perpetrated by men, this study, like other studies (Banyard et al., 2007; Kar & O'Leary, 2010; Panuzio & DiLillo, 2010) found that, in partner relationships, women as well as men engage in both verbal and physical coercion of partners. That is an important finding for its practical implications and because it requires an alteration of the theory that sexual coercion and sexual assault are primarily motivated by and instrumental in maintaining male-dominance in r~lationships and in the society. The practice implication is that, even though men predominate in sexual coercion, it should not obscure the fact that women also engage in sexual coercion. Public service announcements and other educational efforts to prevent sexual coercion of partners need to be explicitly addressed to women as well as men.
Our results also contribute to the debate about whether spanking is harmful enough to be banned. They add to the already large and consistent body of evidence showing that spanking children is a risk factor for a wide variety of maladaptive conditions documented in other chapters and by Gershoff (2002). This study adds sexual coercion to the identified aggressive and criminal behaviors that are linked to spanking. Moreover, because having been sexually abused or neglected were controlled in the statistical analysis, the study shows that the effect of spanking is unique; that is, it is in addition to the effect of those two other types of victimization.
Because antisocial traits and behavior seem to be part of the explanation for the link between victimization and s~exual coercion, further development of brief but effective methods of treating this disorder can contribute to preventing sexual coercion and sexual assault in couple relationships.
Programs directly focused on the prevention of rape, such as the date-rape prevention programs at many U.S. universities, are extremely important, even though the evidence on their, effectiveness is mixed (Davis & Liddell, -2002; Gidycz et al., 2001). This suggests that prevention of sexual coercion and sexual assault also needs to address victimization experiences that have no obvious link to sexual coercion, such as spanking and neglect. Helping parents avoid spanking and avoid even seemingly innocuous forms of neglect could contribute to reducing sexual coercion because spanking and minor neglect are so prevalent in the lives of children and youth.
Finally, the results of this study provides another of many examples of the principle that humane treatment of children can have major benefits in creating more humane relationships in general, including sexual relationships between men and women.
PartV
Social Change and Trends in Spanking
17 The Decline in Spanking
Spanking and more severe corporal punishment to correct misbehavior by a child has been an almost universal part of the childhood experience of children t1rroughout history (DeMause, 1984; Levinson, 1989; Montague, 1978; Newell, 1989). The results of our national survey on corporal punishment in the United States (Chapter 2), in 32 nations (Chapter 3), and in this chapter, show that this is still the case in most of the world. However, change has been taking place, and we will examine the available data to get a picture of the pace of the .change. Some of the data are the results of surveys that asked about attitudes concerning spanking and provide information about cultural norms supporting spanking. We also present data on change in the actual use of spanking. Together, they provide a better understanding of trends in use of corporal punishment. To do this, the chapter pieces together several studies to investigate the following questions:
* To what extent has approval of spanking in the United States decreased since the late 1960s?
* To what extent is the decrease of the result of demographic changes such as the increase in the average level of education and to what extent is it a reflection of changes in cultural nqrms about the appropriateness and necessity of spanking?
* Do various sociocultural groups in American society, such as Blacks and Whites, differ in approval of spanking, and are there differences between groups in the degree to which approval of spanking has decreased?
* To what extent has the actual use of spanking decreased during this period?
* To the extent that there has been a decrease in spanking, does it apply to children of all ages, and specifically to toddlers as well as older children?
Anita K. Mathur is the coauthor of this chapter
Cultural Norms ou the Necessity of Spanking
Interrelation of Social Organization and Cultural Norms
W}len a behavior is almost universally prevalent, there are likely to be cultural norms that encourage or at least legitimate that behavior. If changes in the organization of society bring about a change in behavior, such as families having fewer children which, as shown in Chapter 4 on family size and spanking, tends to result in less spanking, new norms are likely to emerge that justifY and regularize the new pattern. For example, in the second half of the 20th century, as middle class married women increasingly engaged in paid employment, the norms that had previously disparaged paid employment by married women have been replaced by norms that favor it. The new norms almost require paid employment by married women, as they do for men.
Regardless of which occurs first, and despite many exceptions, over time cultural norms and actual behavior tend toward being consistent with one other. A recent example of cultural norms changing to reflect and facilitate change in the organization of society is the norms about unmarried cohabitation. As that has become more and more prevalent, the old stigma and pressures to marry have changed to accepting it as just one additional form of relationship. This was illustrated by the election in 2010 of Edward Miliband to be the head of the Labor Party in the United Kingdom and potentially the next prime minister. He is the father of two children living with his partner. A reporter interviewed residents of Pang bourne, a village of 2,900, how they felt about this. Only a minority in the village voted for his party in the last election and are, therefore, unlikely to be biased because of favoring him politically. The interviewees were just about unanimous in saying his cohabiting relationship was unimportant and some asserted that as a prescriptive norm, "None of that stuff bothers me, nor should it bother anyone" (Henley, 2010).
The shift to what is called a post-industrial economic system in many high economically developed nations is a major change in the economic organization of society, and we believe that these economic changes and the increases in education and other social changes that accompany economic development are some of the things that have produced the decline in spanking. The new economic order brought with it a change in the occupations of most people in economically developed societies. Employment in manufacturing has declined drastically, as has the demand for unskilled manual workers. The predominant occupations are now in services, management, the professions, and sciences. In 1950, 30% of U.S. workers were employed in manufacturing compared with less than 15% in 2008. In 2008 there were more choreographers (16,340) than metal-casters (14,880; Caldwell, 2008). These trends mean that an increasing percentage of the population needs to have interpersonal skills to cooperate, explain, and negotiate, and to be self-directed, autonomous, and creative. These are not skills that are fostered by spanking. When parents require unquestioning obedience and hit rather than explain to enforce it, they model an economic and social system that needs more obedient workers than self-directing workers, and that economic system is disappearing.
Kohn and others (Kohn, 1969; Kohn & Schooler, 1983; Pearlin, 1967; Straus, 1971) found that frequent use of spanking is associated with an emphasis on obedience rather than on reasoning, negotiation, and problem solving. Those parents also have less interest in the child going to college (Pearlin, 1967). Because each instance of spanking is a lost opportunity to learn cognitive, interpersonal negotiation, and managerial skills, there is a concordance between spanking and parental expectations for their child. The studies in Part III are consistent with that theory. These studies held many other things constant such as the parents education, supportiveness, and cognitive stimulation. One is a longitudinal study that followed up a large national sample of children and found that the more spanking, the more a child was likely to fall behind other children in development of mental ability (see Chapter 10).
The second study found that the more spanking, the lower the probability of graduating from college (see Chapter 11). Moreover, one of our previous studies found that, even among those who did graduate from college, spanking is associated with a lower probability of being in the top fifth of the U.S. occupation/income distribution (Straus, 200 1a, pp.13 7-142). Of course, this does not mean that spanked children will not graduate from college. If that were the case, because as shown in research on corporal punishment throughout the United States (see Chapter 2), over 90% ofU.S. parents spank toddlers, almost no one would graduate from college. It just means that a lower percent will graduate. (See the section in Chapter 1, Risk Factors: The Real Meaning of Social Science Research Results.)
Consistent with this theory about the links between the occupational structure of society and modes of discipline, Kohn (1969) found that parents who expected their child to go to college spanked less. He interpreted these results as anticipatory socialization by parents who expect their child to be employed in nonmanual work occupations. That is, when parents envision their child in roles where what is needed is cognitive ability, critical thinking, and negotiation skills rather than physical strength and compliance with the rigid routines of the assembly line, they tend to rely on cognitive more than physical modes of teaching and correcting. To the extent that theory is correct, the norms and practices of parents in the ;united States and other post-industrial societies should be moving away from the use of corporal punishment in child rearing. These characteristics of a post-industrial occupational system may be part of the explanation for a moral passage (Gusfield, 1963, 1981) that could eventually change the world from one in which almost all children are socialized by spanking to one in which this occurs for only a small minority of the population. Such a change would be part of the centuries-long "civilizing process" described in the concluding chapter.
Cultural Norms Supporting Spanking
Although cultural norms supporting spanking children may be changing, there is abundant evidence that these norms are still deeply rooted and pervasive in American culture (Greven, 1990; Straus, 2001d). One of the most fundamental ways in which cultural norms supporting spanking are expressed is in the criminallaw on assault. In every U.S. state, hitting a child for purposes of correction or control is exempt from the crime of assault, usually with the provision that it must be limited to reasonable force (Davidson, 1997). In practice, numerous recent state supreme court decisions have ruled that this includes the right to hit with belts and paddles, provided the child is not physically injured (e.g., rulings by the supreme court ofNew Hampshire in 1992 and of Minnesota in 2008). In the Minnesota case (Olson, 2008), the father paddled his unruly son 36 times. In a unanimous decision, the court ruled that spanking a child is not necessarily abuse. The judges said they declined to rule that the infliction of any pain constitutes either physical injury or physical abuse, because to do so would effectively prohibit all corporal punishment of children by their parents and that they did not intend to ban corporal punishment. Nevertheless, the fact that these and other similar cases even got to a state high court is an important indicator of change.
In the 1960s, every state in the United States passed child protection legislation designed to protect children from physical abuse and to provide services for abused children. Ironically, in order to gamer sufficient votes to pass these laws, it was typically necessary to include a provision declaring that parents continued to have the right to use corporal punishment (i.e., to hit children for purposes of correction and control). Thus, legislation intended to protect children from physical abuse contained provisions that further legitimated hitting children in the form of corporal punishment.
* Indiana provides an example of such legislation, where the state's definition of child abuse and neglect reads: "This chapter [on child abuse and neglect] does not limit: The right of the parent to use reasonable corporal punishment to discipline the child." (Indiana Ann. Code§§ 31-34-1-12; 31-34-1-14; 31-34-1-15)
* Similarly, Ohio's child abuse and neglect statute reads: "A child exhibiting evidence, of corporal punishment or other physical disciplinary measures by a parent is not an abused child if the measure is not prohibited under Ohio state statute." (Ohio Rev. Stat.§§ 2151.03(B); 2151.031; 2919.22)
* South Carolina's child abuse and neglect statute also addresses corporal punishment. "The term child abuse or neglect excludes corporal punishment or physical discipline that: Is administered by a parent or persons in loco parentis; is perpetrated for the sole purpose of restraining or correcting the child; is reasonable in manner and moderate in degree; has not brought about permanent or lasting damage to the child; [and] is not reckless or grossly negligent in behavior by the parents. (South Carolina Ann. Code § 63-7-20)
The irony is that about two thirds of the cases of physical abuse that come to the attention of child protective services started out as spanking and then escalated into injury-producing attacks (Gil, 1970; Gonzalez, Durrant, Chabot, Trocme, & Brown, 2008; Kadushin & Martin, 1981; Trocme, Tourigny, MacLaurin, & Fallon, 2Q03; Straus, 2000, 2008a).
Cultural norms supporting spanking are also documented by public opinion surveys asking respondents whether they approve of spanking. A 1968 survey commission by the National Commission on the Causes and Prevention of Violence found that 94% ofthe population approved of spanking a child who misbehaves (Stark & McEvoy, 1970). There may be no other aspect of child rearing except provision of basic physical necessities on which almost everyone agrees. In addition, the normative support for spanking goes well beyond mere permission to hit children. It is a morally correct and expected action. The question asked in the 1968 survey was whether they thought that spanking was sometimes necessary. Parents who do not spank are often thought of as bad parents whose children will grow up wild. That these are true cultural norms with expectations for conformity to the norm is shown by two studies. Carson (1986) found that non-spanking parents come under considerable pressure to spank from relatives, friends, and neighbors who questioned the adequacy of their parenting. Walsh (2002) studied the 998 mothers in our research on impulsive spanking (Chapter 7) and the child-mother bond (Chapter 8) and found that a majority of parents who do did spank were advised to do so. In the six months before being interviewed for the study, 61% of the mothers of children age 2 to 3 were advised to spank their child 1 or more times, and 52% of mothers of children age 13 to 14 were given the same advice.
Change in Approval of Spanking
The surveys of attitudes toward spanking that have been completed since the late 1960s have found both decreases and, despit~ that, continuing high rates of approval. The percentage who approve of spanking also varies considerably from study to study, probably because the questions used and the samples are so different. Therefore, these studies cannot be compared to determine if there has been a change in support for spanking. For example, a 1975 survey of a national sample of parents found that 77% believed that slapping or spanking a 1-yearold who misbehaved is normal and necessary (Straus et al., 2006). A survey of pediatricians published in the 1990s found that 77% supported spanking (White, 1993). These figures although a decade apart cannot be used to estimate change in attitudes because they used different questions and studied different populations. Nevertheless, the changes in the nature ofU.S. society and the antispanking laws passed in many other nations led to the following hypothesis:
Hypothesis 1: The percent ofthe U.S. population who believe that spanking is necessary has declined since 1968.
For this chapter, we also examined whether public approval of spanking decreased as result of demographic changes in U.S. society rather than because of a change in the cultural norms. For example, the percent of college graduates in the United States has been increasing tremendously. College graduates are less favorable to spanking. Therefore, the large increase in the percent of college graduates should result in less approval of and less spanking. A similar effect might result from the increase in the average age of the population. If people become less favorable to spanking as they mature, the overall level of approval could decrease when there are more older people in the population. However, that might not be the case because older persons are not only more mature, they also represent a previous historical setting that was favorable to spanking and that might produce the opposite result. To find out the extent to which such demographic changes explain the decrease in the percent who approve of spanking, we tested a second hypothesis about change in the public approval of spanking:
Hypothesis 2: The decrease in approval of spanking will still be present after controlling for six demographic variables.
If hypothesis 2 is confirmed, it is plausible to infer that the decrease reflects a change in cultural norms rather than a change in the demographic composition of the population.
The Seven Surveys
These two hypotheses could be tested because there are publicly archived data files for seven surveys that used the identical or very similar question to determine the approval of spanking. All seven surveys used nationally representative samples of adults and collected data on the six sociodemographic variables that were included in the analysis for this chapter: the age of the respondent, income, education, gender of the respondent, and race. The samples, which together had 6, 794 participants, are:
* Survey 1 is a 1968 national survey, designed and conducted for the National Commission on the Causes and Prevention ofViolence. It studied 1,160 adults (Stark & McEvoy, 1970).
* Surveys 2-6 are the General Social Surveys for the years that included the question on spanking (1986, 1988, 1989, 1990, and 1991). (Davis & Smith, 1992).
* Survey 7 is the survey that provided the data on corporal punishment in the United States in Chapter 2.
Six of the seven studies obtained data on approval of spanking by asking, "Do you strongly agree, agree, disagree, or strongly disagree that it is sometimes necessary to discipline a child with a good hard spanking?" A somewhat different question was used in the 1968 survey that asked: "Are there any situations that you can imagine in which you would approve of a parent spanking his or her child assuming the child is healthy and over a year old?" Respondents in this survey could answer only yes or no, rather than choose from four response categories. We tried to make the data from studies that used four response categories somewhat more equivalent by combining strongly agree and agree into agree, and combining disagree and strongly disagree into disagree. Thus, the data for all seven studies classifies the respondents as either agreeing or disagreeing. Nevertheless, some of the difference between the 1968 study and the other six studies may be a result of the difference in the wording of the question.
Trends in Support of Spanking
Chart 17.1 shows that the belief that spanking is sometimes necessary decreased from 94% in 1968, to 65% in 1994. Another survey asked the same question in 1999 and found a further decrease in the approval of spanking to 55% (Children's Institute International, 1999). In one generation, approval of spanking seems to have gone from something almost everyone in the United States thought was necessary to a bare majority. Of course, 55% is still a majority. However, other surveys since then suggest that the decrease to 55% may not be accurate. The decrease to 55% is inconsistent with more recent results from the General Social Survey to be .discussed later. Nevertheless, using just the data up to 1994, the decrease from near unanimity to 68% represents an important social change, especially when one views it as a change in support for a cultural norm that has existed for thousands of years and which is the law in every state of the United States.
Chart 17.1 The Belief that Spanking Is Sometimes Necessary Dropped from Almost Everyone in 1968. However, in 1994, Two Thirds Still Thought Spanking Was Necessary
Chart 17.1 originally had two trend lines. The line shown has been adjusted to control for the six demographic variables. There was another line showing the percentages before adjusting for the control variables. Because controlling for changes in the demographic composition of the population made almost no difference and the lines overlapped in a way that made the chart more difficult to read, the line for the raw data is not included in Chart 17.1 More important, the fact that the two trend lines are almost identical, indicates that the decrease from 1968 to 1999 represents a change in cultural norms rather than a change in demographics because the controls adjusted the data to take into account changes in the demographics.
Group Differences in the Trend
Another purpose of analyzing the seven surveys was to find out the extent to which various sociocultural groups in U.S. society differ in the approval of spanking, and if there are differences between groups in the trend over the 31- year period of the study. The evidence from other studies on group differences in the approval of spanking is contradictory. The 1968 survey by Stark and McEvoy (1970) found no difference on any of the demographic variables such as educational level. However, Flynn (1994) analyzed the 1988 General Social Survey and found that approval of spanking varied among different regions of the United States. The West, Midwest, and the South had more favorable attitudes toward spanking than the Northeast. Flynn also found that Blacks are more likely than Whites to favor spanking, and that low education, being male, and being young were also associated with a greater likelihood of approving of spanking. Why did Flynn find differences between regional and other groups, and Stark and McEvoy did not? Perhaps it is because in 1968, almost everyone (94%) thought spanking was necessary. That creates what is called a statistical ceiling effect because there is little room for group differences. The decline in approval of spanking since 1968 may have reduced the ceiling effect. If so, the even further decreases in approval of spanking should allow for more differences between demographic groups. We investigated this by examining the pattern of change over time within each group. Doing that uncovered some clues
to explain the potential differences between groups in approval of spanking. Gender. Chart 17.2 shows that from 1968 to 1994, approval of spanking decreased among both men and women but with a somewhat greater decrease for women. among 1968 almost everyone believed that spanking was necessary (95% of women and 92% men). By 1994 the slightly more women than men had reversed, and fewer women than men approved of spanking ( 61% of women versus 76% of men). This suggests that men have held on to the old values concerning spanking somewhat more than women.
Chart 17.2 The Decrease in Approval of Spanking Has Been Greater for Women than for Men
Ethnic group. Three race or ethnic groups were compared: Whites, Blacks, and other. Overall, in 1968, there was little difference in the percent who approved of spanking. Over 90% of all three groups thought spanking was necessary. The trends in approval of spanking for each ethnic group in Chart 17.3 reveal large decreases in approval of spanking for both, Whites and others over the 26-year period. For Blacks, the decrease in approval of spanking was much less than for the other ethnic groups-14 percentage points for this period, compared with 26 percentage points for Whites and 32 percentage points for the other category.
Region. Chart 17.4 shows that in 1968 there were almost no differences between regions. At that point in U.S. history, almost everyone, regardless of region believed that spanking was sometimes appropriate. Over the 26 years covered by this study, residents in all four regions decreased in the approval of spanking. Nevertheless, Southerners generally had the highest rate of approval of spanking at each time point, and the decrease was less than in other regions from 1968 to 1994. The South decreased only 21 percentage points whereas the Northeast decreased a total of 35 percentage points from 1968 to 1994. The regions thus grew more different from each other in the proportion approving spanking mainly because the South did not decrease as much as other regions.
Chart 17.3 The Decrease in Approval of Spanking Has Been Greater among Whites and Other Ethnic Groups than among Blacks
Age. As expected, the rate of approval of spanking was greatest for older respondents because they were socialized in a cultural climate that was more approving of spanking. However, the differences were not as large as we expected. When age is divided into five categories (from the youngest fifth in each survey to the oldest fifth), the maximum difference between any two age groups was only 6%. All five age groups followed the general downward trend.
Income and education. No relationship was found between income and approval of spanking. When income, like age, was divided into five categories (lowest fifth to highest fifth within each study), all five income groups followed a similar downward trend in approval of spanking from 1968 to 1994. Education, on the other hand was strongly related to less approval of spanking. The respondents in each survey were divided into quartiles: the quarter of each survey with the least education, the two middle quarters in education, and the quarter with the highest education. All four educational groups decreased in approval of spanking over the 26-year period, but the highest education group decreased the most. As with the other demographic characteristics, in 1968, almost everyone approved of spanking and education made little difference. By 1994, the highest education group had a lower rate of approval than the other education groups.
Chart 17.4 Difference between Regions in Approval of Spanking Increased because in 1968 Almost Everyone Approved
Trends in Approval of Spanking since 1994
The General Social Survey continued asking the same question on whether spanking is sometimes necessary in surveys up to 2010. Chart 17.5 extends the trend line to that date. We expected that the decline shown previously in Chart 17.1 would continue or even accelerate. Instead, the percentages in Chart 17.5 for the years since 1994 show a remarkable cessation in the downward trend of public opinion about the necessity of spanking. Instead of a continued large decrease with each survey, there has been almost no change since the mid 1990s. Just over 70% of the U.S. adult population continues to believe that "a good hard spanking is sometimes necessary." Research is needed on why the decrease has not continued. A similar cessation of the rapid decline in approval of spanking was found by Gallup/ABC News surveys (Pinker, 2011, p. 436). We suggest that part of the explanation follows from the results showing that almost all parents continue to hit toddlers, even though as shown in Chart 17.7, they now hit toddlers less often. It may be an example of the principle explained in the introduction to this chapter: cultural norms both influence and also reflect what people actually do. That is, because over 90% ofU.S. parents spank toddlers, the cultural norms and beliefs about spanking continue to reflect that aspect of what parents do. Or putting it another way, it may be that U.S. parents are reluctant to believe that what they did was not necessary.
Chart 17.5 Percent of U.S. National Samples Who Agreed or Strongly Agreed that "A Good Hard Spanking Is Sometimes Necessary" (General Social Survey for all years except 1968. Moving Average 3 per year.)
Trends in Spanking
Almost three quarters of U.S. adults may believe in the necessity of spanking sometimes, but what percent actually spank and has that changed? There are many indications that the actual use of spanking is decreasing. There are two studies that provide data on the extent to which this has occurred. The first are the annual surveys conducted between 1988 and 1999 by Prevent Child Abuse America. Those surveys found that during just that decade, the percentage of parents who spanked in the past year decreased by one third (Daro & Gelles, 1992). This decrease is consistent with the theory outlined in the first part of this chapter-that parents tend to socialize children in ways that will help them fit into the social environment that the parent thinks the child will experience as an adult. Specifically, these children face a post-industrial society where selfdirection and cognitive and interpersonal skills are needed for more and more of the available jobs. However, the Prevent Child Abuse America surveys did not provide data on the age of the child. As a consequence, we do not know if the decrease was for across the board or more for children of certain ages. Data on trends for each age group is important because, as found in our study about the use of corporal punishment in the United States (Chapter 2), the percent of spanking ranges from almost all parents of toddlers to a small percentage of parents of older teenagers.
Fortunately, we were able to examine the trend in spanking for children of different .ages using data from three studies of nationally representative samples of parents: the National Family Violence Surveys conducted in 1975 and 1985 (Straus, 1990d) and the survey conducted for us in 1995 by the Gallup Organization that provided the data about the prevalence of corporal punishment in the United States (Chapter 2). The three studies used three somewhat different versions of the Conflict Tactics Scales to measure the use of corporal punishment. However, the question on spanking was the same in all three surveys. This enabled us to test the hypothesis that spanking has decreased over these three time periods. The question asked the parents whether they had spanked or slapped the child with their open hand in the past year and how often. The percentages who spanked will be lower than in other reports on these three studies, including the Gallup study (Chapter 2), because those reports were based on the full Conflict Tactics Scales, which also asked about methods of corporal punishment such as use of a belt or hairbrush.
Chart 17.6 shows that in 1995, 74% of the parents of2- to 4-year-old children said they had spanked or slapped the child in the past year. This is only a small decrease (12%) from the 83% who spanked in 1975. On the other hand, spanking 9- to 12-year-old children decreased by 31% from 1975 to 1995, and for teenage children it decreased by 56%. In the next chapter, we suggest an explanation for how it can be that, for older children, approval of spanking and actual spanking decreased to one third or one half of what it had been in 1975, whereas during this same period almost all parents continued to hit toddlers. Why was there so little change in spanking younger children during the same period that cultural norms supporting spanking declined so precipitously, and -when actual use of corporal punishment with older children also decreased at what may be a historically unprecedented pace? That chapter also suggests what would be needed to produce the same decrease for young children as has occurred for older children.
On the other hand, quite a different picture emerges when the criterion for change is how often parents who spanked did it. Chart 17.7 shows that the biggest decrease was for toddlers and the smallest decrease in how often parents spanked or slapped was for teenagers. What this boils down to is that although there was little change in the percent of parents who hit toddlers, how often they did it dropped by 51%, whereas the opposite was found for parents of teenagers. There was a big decrease in the percent of parents who hit teenagers, but those few who did it in 1995 hit almost as often as in 1975
Chart '17.6 Spanking and Slapping Has Decreased Tremendously since 1975 for "Tweens" and Teens, but Very Little for Toddlers and Younger Children
Chart 17.7 How Often Parents Spanked and Slapped Also Decreased since 1975, the Most Decrease Was for Toddlers
Data from national surveys conducted by Murray A. Straus and colleagues. University of New Hampshire (//pubpages@unh.edwmas2).
Spanking in 2006
Although we were not able to analyze a more recent national survey of parents using the same question on spanking and slapping as was used in 1975, 1985, and 1995, the 2006 Harris Youth Survey included a question we suggested. This provided recent and nationally representative data based on the responses of 1 ,213 children age 8 to 17. The participants were asked:
"In the past year, that is, since about this time last year, about how many times did your mother or your father (or whoever was taking care of you that year) use physical punishment, like spanking, slapping or hitting you? Once; twice; 2 to 5 times; 6 to 10 times; 11 to 20 times; more than 20 times; not in the past year, but it has happened before; never."
The data from the Harris Youth Survey cannot be used to infer change from 1995 to 2006 because the measure was not the same as in the surveys in Chart 17 .6. In addition, the participants in the study were children rather than parents. There may be differences between parents and children's willingness to disclose or ability to remember instances of spanking. If children disclose or remember more instances, the positive side is that the Harris Youth Survey data probably provides a more complete picture of the extent to which cor~ poral punishment is used than do parents. Chart 17.8 shows that over 40% of children age 8 to 10 said that they had been spanked or hit during the previous year. This is a high percentage, but it is much lower than the approximately 65% of children that age in the 1995 survey (see Chapter 2, Chart 2.1 ). Unfortunately, because of the differences between the Harris survey and the 1995 survey, it does not necessarily mean a decrease in corporal punishment. For older children, the percentages are also lower than a decade previously, but still very high. In 2006, almost one quarter of early teenage children and about one out of seven children age 17 and 18 were hit by their parents during the year of the study.
A study by Vittrup and Holden (2010) provides further evidence that, despite decreases, spanking remains both the statistical and the cognitive norm. They studied the parents of 108 children aged 6 to 10 living in a large city in Texas and also the children. According to the parents, 71% spanked their children at least once in the past week. According to the children, 82% of the 108 had been spanked in the past year. This is about the same percentage as shown in the Gallup study on the prevalence of corporal punishment in the United States in 1995 (Chapter 2). Of the 6-year-old children, 70% reported being spanked at least once a week, 36% were spanked 1 to 2 times per month or once a week, and 20% spanked 2 or more times per week. A study by Taylor and colleagues (Taylor, Lee et al., 2010) found that two thirds of3-year-old children were spanked at least once in the previous month by one or both parents.
Chart 17.8 The Percent of School-Age and Teenage U.S. Children Spanked or Hit in 2006 Remains High
Finally, the study by Zolotor, Runyan, Chang, and Laskey (20 11) described in Chapter 2 found that 82% of the parents of children age 3 to 5 had spanked or slapped the child with an open hand that year. This is less than the 94% reported in Chapter 2 for our 1995 national Gallup survey, but not a lot less. Zolotor et al. themselves compared their results with our study on corporal punishment in the United States (Chapter 2) and with our 1975 and 1985 national surveys. They concluded that between 1975 and 2002, 18% fewer children were slapped or spanked by a caregiver and that this decrease was primarily for the older children.
Trends in Europe
Several nations in Europe have enacted legal prohibition of corporal punishment by parents. Bussman, Erthal, and Schroth (2011) studied changes in five ofthose nations. In Sweden in the 1950s, 94% of parents spanked, and one third did it at least daily (Stattin et al., 1995). One can infer from this that almost all Swedish parents did it at least once a week. By 1995, the percentage of parents who had spanked decreased to 33%. A 2003 study found no further decrease in the percentage who had ever spanked but found a tremendous reduction in the percentage who spanked once a week or more-to only 4% (Durrant, Rose-K.rasnor, & Broberg, 2003). In Germany, surveys of nationally representative samples of children aged 12 to 18 in 1992 and 2002 found large decreases, especially in the most severe forms of corporal punishment (Bussman, 2004).
• Light slap in the face decreased from 81% to 69%
* Severe slap in the face decreased from 44% to 14%
* Beaten on the bottom with a stick decreased from 41% to 5%
* Beaten to the point of bruising decreased from 31% to 3%
Decreases of that magnitude are not likely to have occurred in response to a law passed in 2000, especially because the law was not known to 70% to 75% of the population. These remarkable decreases probably reflect changes in German culture and social organization as much or more as the law banning spanking. Nevertheless, analysis of the trends in Sweden led Bussman to conclude that an unambiguous legal prohibition accelerates the change.
Summary and Conclusions
Approval of Spanking
In 1968, 94% of the U.S. population believed that spanking was necessary. Except for the obligation to provide for the basic physical needs of a child, it is hard to think of any other aspect of, child rearing on which there was such consensus. The percent who believe that spanking is sometimes necessary decreased steadily until the late 1990s. However, since then, it has remained at just over 70% until the most recel?:t survey (2010).
Decline in Consensus on Spanking
One thing that is consistent in the recent General Social Survey results and other results in this chapter is that there are important differences between sectors of the population in believing that spanking is sometimes necessary: more men than women, more low education than better educated, and more Blacks than Whites. These groups also differed in the rate of decrease. Because some groups in the population have changed faster than others, it has ended the former national consensus on the necessity of spanking. The different rate of change has produced a widening gulf in attitudes toward the appropriateness of spanking children. For example, approval of spanking by men has decreased, but less than approval by women. Therefore, in contrast to the results for 1968, men are now stronger supporters of spanking than women. As a result, there are likely to be more couples who disagree about the necessity of spanking, and that disagreement could affect marriages as well as children.
Although approval of spanking in the South decreased since 1968, the decrease was smaller than in other regions of the United States. As a result, even though support for spanking declined in the South, the South has remained more favorable to spanking than other regions. Similarly, approval of spanking by Blacks declined, but less than in other ethnic groups. As a result, the Black-Wliite difference in the belief that spanking is necessary has become larger than ever. Approval of spanking declined less for people with lower education and for older age groups. These differences in the rate of change have resulted in subcultural group differences in U.S. society in respect to spanking that did not exist or were minimal only a generation ago. In 1968, there were no differences because almost everyone (94%) believed in the necessity of spanking. The findings of the research in this book suggest that this change will have great benefits for children and for society. At the same time, the discrepancies created by differences in the extent to which different parts of the population have embraced the change can create or increase conflict over whether to spank, such as exists between fathers and mothers and Blacks and Whites. It is an example of the principle that all social change, no matter how beneficial, also has costs.
Is There a Backlash?
A decrease from 94% who believe spanking is sometimes necessary to the current 70% indicates both a major social change and persistence in the beliefthat spanking is necessary. It means that the cultural norm specifying that parents should spank when necessary, although no longer nearly unanimous, continues. In addition, what was a consensus on the necessity of sometimes spahking has become a widening gap between sectors of society. A change in a cultural norm about a family behavior that is as deeply embedded in peoples' beliefs about what is right can be a problem. It is one element in the much talked about culture wars. It may have given rise to a backlash against the movement to bring up children more humanely. An indicator of the backlash may be that child protective service agencies, which have responsibility for enforcing the legislation intended to protect children from physical and other abuse, have come under attack on the grounds that they prohibit spahking, even though that is not the case in any state in the United States. Legislation placed before the Florida legislature in the spring of 1994 would have prohibited child protective services from ruling that a child had been abused if the only evidence was that the parents used corporal punishment, and the child exhibited welts or bruises. The bill was passed overwhelmingly but was vetoed by the governor. However, the reason for the veto was not a disagreement with the purpose of the bill; it was because the governor believed it would create administrative problems (Hollis & Currie, 1994; Neal, 1994).
The seeming backlash and the widening gulf between socioeconomic status groups with respect to spanking is a source of concern. Ignoring these problems could interfere with the transition to new cultural norms that forbid any spanking. It could exacerbate class and race conflict if efforts to end spahking are seen as an attempt to control and limit the lives of the underprivileged. Such a perception would be a cruel irony because as we have previously noted in this book and in a previous study (Straus & Gimpel, 2001), violent socialization in the form of frequent spanking reduces the probability of upward social mobility. The greater adherence to spanking by low socioeconomic status groups suggests that the shift away from spanking among the majority, and the lesser change among low socioe,conomic status groups, could have important consequences for class and race differences. We believe spanking by low socioeconomic status parents has and will continue to retard efforts to create a more equal society because, as this book and much other research shows, spanking is one of the reasons that the children of the poor have lower IQ, have a higher probability of being physically aggressive, have more psychological problems, and have more crime as adults. The social class gap could increase if low socioeconomic status parents continue to be frequent spankers and higher socioeconomic status parents continue to spank less and less.
In evaluating these conclusions, it is important to keep in mind that we said "increases the probability." Most children who are spanked a lot will not experience lower IQ than they would have otherwise, but an avoidable percentage will. This is the same sort of relationship as the link between smoking and dying of a smoking-related disease. As we pointed out in the introductory chapter to this book (Chapter 1) and elsewhere, one third of heavy smokers will suffer that ultimate harmful side effect. But that same statistic means that most heavy smokers will not die of a smoking-related disease. (See the section in Chapter 1 on Risk Factors: The Real Meaning of Social Science Research Results.)
The additional handicap imposed on children of low socioeconomic status suggests that although the efforts to end spanking need to be directed at all parents, there needs to be an even stronger effort for the low-income and loweducation sectors of society. That includes many Blacks. Leaders of the Black community, including Black academics, need to be at the forefront of that effort. For academics, a needed focus is research on the most effective way to communicate the information about the effectiveness and side effects of spanking to different socioeconomic and cultural groups.
Variations in Trends in Spanking
The trend in respect to the use of spanking in the United States, like the trend in approval, presents a mixed picture. Among teenagers, the percent hit by parents is about one half of what it was in 1975. For children age 9 to 12, it has declined by about one third. However, spanking and slapping toddlers remains almost universal. That is a cruel age difference becausethe study of the adverse effects on spanking on IQ in Chapter 10 found that the harmful side effect of spanking is greatest at that age. However, that may not be true for all side effects of spanking. Finkelhor's (2008) Developmental Perspectives Model of child victimization argues that the age of maximum sensitivity to a form of victimization can depend on the nature of the developmental task. It may be different for attachment, control of aggression, peer relationships, or learning to defer gratification in favor of presumably larger later rewards. The most vulnerable age for each of these is unknown. But what is known is that the advice to confine spanking to preschool-age children, whereas an improvement over no age limitation is based on folk-wisdom rather than research evidence.
The huge and almost entirely negative reaction described in Chapter 1 to a Time magazine article in April 2010 on The Long-Term Effects of Spanking could be interpreted as a backlash against the growing professionaLadvice to avoid spanking when possible. An alternative interpretation of what seems to be a backlash is that it may be more a dying gasp by those who have always believed in the necessity of spanking. In the past, they did not have to defend the use of spanking because, as shown in Chart 1 7.1, the necessity of spanking was a belief shared by close to 100% of the U.S. population. Not only is that no longer true, but the advocates of spanking may eventually become a beleaguered minority. An indicator of the emerging cultural shift was a 2006 episode of the immensely popular television show Desperate Housewives. It pilloried one of the characters for spanking. A similar manifestation of the changing cultural landscape is the TV show, Supernanny. In her first book, Jo Frost; the supernatiny of the show did not rule out spanking. But late in 2006, the U.S. edition of Supernanny devoted an episode to the effort of Supernanny to help a mother stop spanking.
The Confusion of a Culture in Transition
The cultural norms and practice of spanking are in a state of transition. Part of the inconsistency between studies and between the trends for younger and older children probably reflects that. The question used for this chapter to measure the approval of spanking requires a more complete rejection of spanking than has been achieved so _far in the United States. It asks study participants whether they agree or disagree that spanking is sometimes necessary. To answer "disagree" means that the person believes children should never be spanked. Our guess is that, if the survey question were about whether spanking should be avoided as much as possible, most people in the United States, and especially the college educated, would now agree. The era in which spanking was thought to build character and prepare the child for the hard knocks of life is mostly gone. But an era in which children should never be spanked has not yet arrived in the United States. It has arrived in Sweden and, as will be shown in the final chapter, which discusses the notion of a world without spanking, the process is well underway in other nations. Our estimate of the current situation is that, in the United States, most people including most professionals concerned with children such as pediatricians, psychologists, and parent educators, are opposed to the use of spanking. However, they also believe it is sometimes necessary. An example we hear over and over is when a child runs out into the street or repeatedly commits the same misbehavior. Although never spanking is not currently the norm in the United States and most other nations, we believe it will eventually become the norm in all nations. Of course, there will always be some parents who spank, just as there will always be some parents who do not use seat belts.
As pointed out at the start of this chapter, the change is part of the transition to a post-industrial economic order, which as argued in the concluding chapter, is part of a process of social evolution that, 1since the late middle ages, has reduced person-to-person violence (Eisner, 2003; Elias, 1978; Pinker, 2011; Straus, 2001d). In our opinion, the social organizational and cultural evolution that has reduced other types of violence will eventually make spanking or slapping a child as offensive, morally and legally, as slapping a wife or husband.
In the final months of his life Professor Straus handed out free copies of this book and took other actions to indicate that he wanted to make it available, presumably because he recognized that his work could help teach how to reduce violence if made available freely. Therefore I'm posting the rest of the book in four parts.
The Primordial Violence 2014-1
The Primordial Violence 2014-2
The Primordial Violence 2014-3
The Primordial Violence 2014-4
Did the Disturbing Philosophy of To Train Up a Child Lead to Hana Williams’ Death? 10/03/2011
Spare the Kids: Dr. Stacey Patton
Cross-cultural comparative studies show that societies that are high in use of spanking also tend to be high in frequency of warfare and violent interpersonal relationships (Lansford, 2010; Otterbein, 1974; Russell, 1972) p.215
Kohn and others (Kohn, 1969; Kohn & Schooler, 1983; Pearlin, 1967; Straus, 1971) found that frequent use of spanking is associated with an emphasis on obedience rather than on reasoning, negotiation, and problem solving. Those parents also have less interest in the child going to college (Pearlin, 1967). p.251
We believe spanking by low socioeconomic status parents has and will continue to retard efforts to create a more equal society because, as this book and much other research shows, spanking is one of the reasons that the children of the poor have lower IQ, have a higher probability of being physically aggressive, have more psychological problems, and have more crime as adults. The social class gap could increase if low socioeconomic status parents continue to be frequent spankers and higher socioeconomic status parents continue to spank less and less. p.267
No comments:
Post a Comment