The value of physical or corporal punishment is disputed among psychologists; some regard it as harmless, while many others consider it potentially harmful. Some researchers have controversially suggested that parental use of physical punishment may be causally related to the development of aggression. The nature of this relationship, however, has been hypothesized among various writers to be both linear and curvilinear, positive and negative, and even nonexistent. What is lacking is a detailed critical review of the literature examining physical punishment and its causal relationship to aggressive or violent behavior.
An examination of the literature reveals that most studies are supportive of a relationship between physical punishment and aggression. Further, prospective studies suggest that physical punishment may contribute etiologically towards the development of aggressive behavior. Of those few studies which provided the relevant data, a majority favor a curvilinear, rather than a linear, causal relationship. However, the impact of physical punishment on children's aggression levels may vary with the age and gender of the child. The very few studies which examine physical punishment in interaction with other parental factors (such as discussion and use of reasoning) suggest that any noxious effects of physical punishment may be mitigated by other parental disciplinary behaviors. Finally, however, this literature's conclusions are greatly limited by significant methodological flaws, notably control for factors such as child abuse, parental substance abuse, and other parenting behaviors. Prospective studies with appropriate controls are notably missing.
In the United States, interpersonal violence is the leading cause of death among young black males and the leading cause of injury among women (Jones, 1990). A large research literature has sought to elucidate the causes of such violent and aggressively antisocial behavior. One focus of this etiological literature has been the role which various childrearing practices play in the development of aggression.
One childrearing experience which has been an important focus has been children's exposure to parental physical aggression. Such aggression is commonly conceived as existing on a continuum, ranging from very severe parental aggression (i.e., child abuse) to much milder and normative parental aggression, such as use of corporal or physical punishment. Many psychologists are opposed to any parental aggression, even the use of physical punishment (Maurer, 1974). One common concern is that parental use of physical punishment will lead to aggressive behavior in children (Straus, 1983). In acknowledgment of the potential harmfulness of physical punishment, the American Psychological Association announced its opposition to its use in schools in 1975 (Anderson & Anderson, 1976).
On the other hand, surveys suggest that other psychologists believe physical punishment to be an effective and useful socialization tool (Anderson, et al., 1976; Leviton, 1976; Lowenstein, 1977; Maurer, 1974). In response to the hypothesis that physical punishment could cause aggression in children, these professionals can cite studies which note that even abusive parental violence does not always lead to an increase in children's aggression (Spatz Widom, 1989b). Further, recent survey data has suggested that almost all American parents utilize physical punishment at one point or another (see Prevalence data reviewed below). This suggests that despite the psychological controversy over the topic, many, if not most, Americans regard physical punishment as an appropriate childrearing technique, at least under certain circumstances.
In addition to the controversy over whether any relationship exists between physical punishment and aggression, a second controversy exists over the nature of this relationship. Briefly, three types of relationships have been hypothesized to exist between physical punishment and aggression. The first type of relationship is a positive, linear one: some researchers have contended that any parental aggression (even physical punishment) may be positively and causally related to the development of antisocial aggression (e.g., Steinmetz, 1979; Maurer, 1974). In sharp contrast, a second group of researchers have suggested that lack of physical punishment may contribute causally to the development of aggression (e.g., DiLalla, Mitchell, Arthur & Pagliocca, 1988). Finally, a third group proposes a curvilinear relationship: either too little or too much physical punishment may increase the probability of aggressive behavior (e.g., Gelles, 1974; Lefkowitz, Eron, Walder, & Huesmann, 1977).
As Maurer (1974) suggests, what is lacking (among these varying hypotheses) is a detailed critical review of the relevant literature, examining the evidence for (a) the existence of a relationship between physical punishment and aggression; and (b) the nature of that relationship, if existent - - either simple or complex, linear or curvilinear. The purpose of this paper is to provide that review.
This paper has labeled parental physical punishment as a type of parental "aggression." Others have referred to it as parental "violence" (Straus, 1990). These labels are controversial, since "violence" often implies an intent to harm, whereas parents presumably employ physical punishment with the intent to help. For lack of a more neutral term, I have therefore chosen to refer to physical punishment as one type of parental "aggression," since regardless of parental intentions, it inherently involves the use of at least mild physical aggression. Acknowledging the aggressive nature of corporal punishment leads to the question of whether the use of physical or corporal punishment increases a child's probability of developing aggressive or violent behavior.
The scope of the literature on physical punishment is not vast. Typically, more severe forms of parental aggression (i.e., child abuse) are examined and are the subject of comprehensive reviews (e.g., Spatz Widom, 1989). Studies which seek to associate child abuse and the subsequent development of violent behavior have yielded results which strongly suggest that being exposed to abusive parental violence constitutes a significant risk factor for the development of violent behavior (Parke & Slaby, 1983; Spatz Widom, 1989). In contrast to this justifiably widespread focus on extreme forms of parental aggression, fewer researchers have evidenced interest in the cognitive and behavioral consequences of the most common type of physical aggression by the parents --- the socially-sanctioned aggression referred to as physical or corporal punishment (these terms will be used interchangeably throughout this paper, as they are in the literature).
The goal of this paper is to review the research which specifically addresses physical punishment and the outcome of violent or aggressive behavior. This paper is not intended as a review of the literature on child abuse and aggression. Physical punishment differs significantly from abusive parental violence in many ways, including (but not limited to) the degree of aggression used, the potential for injury, how deviant the behavior is, and the typical intention of the parent involved. Due to the enormity of these differences, the literature examining child abuse cannot satisfy questions about the use of physical punishment. However, child abuse is a methodologically relevant issue: i.e., some research which purports to study physical punishment in fact examines abusive parental violence. In addition, failing to control for child abuse can confound relationships involving physical punishment. Therefore, although this is not a review of the role of child abuse, one of the goals of this paper is to more clearly tease apart child abuse and physical punishment and to specifically isolate and examine research which studies physical punishment.
A second area which this review will not cover is the theoretical mechanism by which physical punishment might lead to aggressive behavior in the child. The purpose of this paper is, again, only to review the evidence for a relationship between physical punishment and the development of aggressive behavior. The issue of mechanism is clearly related (and in fact dependent upon the nature of such a relationship), but also separate. Some mechanisms which have been proposed include modeling (Straus, 1990); attachment theory (Hirschi, 1969 ), lack of reinforcing childrearing techniques (Lefkowitz, Eron, Walder & Huesmann, 1977), the frustration-aggression hypothesis (Dollard, Doob, Miller, Mowrer & Sears, 1939), and many others. The interested reader is referred to these publications for a discussion of different theoretical bases for this proposed relationship. The goal of the current paper is confined to the elucidation of the relationship (if any exists) between physical punishment and child aggression.
Third, the review will be confined to studies which specifically assess violence or aggression as the outcome variable, rather than, for example, "antisocial behavior" or "delinquency." Much research has demonstrated that specifically violent behavior and more generally antisocial behavior often have significantly different etiologies (Magnusson, Stattin & Duner, 1983; Wolfgang, 1983). To combine these two different (albeit related) outcomes may only serve to mask or distort actual relationships.
Finally, this review does not claim to be exhaustive or to comprehend the entire area of aggression, parental aggression, or even physical punishment. Rather, it is intended to address a relationship defined by specific antecedents (physical punishment) and specific outcome variables (violent or aggressive behavior). The resulting review must thus unfortunately exclude some widely cited longitudinal studies on family rearing and antisocial behavior (Glueck & Glueck, 1970; Patterson & Dishion, 1985; Wolfgang, 1983). If such landmark studies are excluded from this review, it is because they fail to address either parental use of physical punishment (not abusive violence) or violent/aggressive behavior (versus the more general category of "antisocial" behavior).
To understand the impact of any variable on any outcome it is essential to understand base rates of the studied phenomenon (Spatz Widom, 1989a). Therefore, I will begin the paper by briefly reviewing the data for baseline rates of physical punishment in the United States. Next, I will examine retrospective, cross-sectional, and prospective research on the relationship between physical punishment and aggression. Finally, I will examine methodological issues and how they affect the interpretation of relevant questions such as:
(1) Is the proposed relationship linear or curvilinear?
(2) When controlling for child abuse, is physical punishment still linked with aggression in children?
(3) Are noted relationships different for male and female children?
(4) What is the relationship between the use of physical punishment and other childrearing behaviors?
I. Prevalence of physical punishment
Prevalence estimates of physical punishment range greatly, depending on the nature of the assessment. Difficulties in estimating the prevalence of physical punishment include discrepancies between referent periods (e.g., one year versus "ever") and discrepancies between the age of the referent child.
When parents of older children are questioned about their use of physical punishment over short referent periods (e.g., over the previous month or year), the percent who admit using such discipline varies between 17% (DiLalla, Mitchell, Arthur & Pagliocca, 1988), 57% (Lefkowitz, Walder & Eron, 1963), and 71% (Gelles, 1978). However, when adult individuals are questioned about their own exposure to physical punishment over their entire childhood, much higher percentages are reported: for example, Deley (1988) found that 89% of his subjects reported that they had experienced physical punishment; similarly, 95% of Bryan and Freed's (1982) subjects recalled experiencing such punishment. Further, studies of toddlers almost always show rates of over 90% (e.g., Sears, Maccoby & Levin, 1957; Straus, 1990). Thus, it seems very likely that a vast majority of Americans are subjected to corporal punishment at one point or another during their lifetime.
The widespread nature of physical punishment has brought into question its relevance in the development of aggressive behavior. Almost all individuals are physically punished, yet only a fraction ever develop deviant, violent behavior. However, this fact does not, by itself, merit dismissal of physical punishment as a potentially important variable. To appreciate the potential contribution of near-universal variables such as physical punishment, it is important to distinguish between necessary and sufficient preconditions. There are many universal or nearly universal conditions which are necessary preconditions for the development of rare events. For example, although sexual intercourse is an almost universal behavior, it is associated with a rare event: cervical cancer. Although intercourse does not "cause" cervical cancer in any simple sense, such cancer almost never develops in the absence of intercourse: i.e., nuns almost never contract cervical cancer, while prostitutes do so much more frequently (Skrabanek, 1988). Thus, the relationship between sexual intercourse and cervical cancer is a casual, but not a simple, relationship.
Likewise, physical punishment may be a contributing factor, but not a sufficient precondition, for the development of violence or aggression. It seems implausible that only a simple, one-on-one causal relationship could exist between any parental aggression and the development of deviant violence (Curtis, 1963; Spatz Widom, 1989b). Researchers have observed that childhood violence experiences appear to be mediated by other developmental factors (Miller & Challas, 1981). (In fact, this was precisely what Spatz Widom (1989) found in the case of child abuse and aggression.) Therefore, it appears to be inappropriate to dismiss the study of theoretically important variables on the basis of a weak one-on-one relationship.
Given the high prevalence of physical punishment, and its admittedly aggressive nature, the potential impact upon behavior is significant. Therefore, I will begin to examine the relationship between physical punishment and child's aggression by reviewing the studies which meet criteria for inclusion here.
II. The Relationship Between Physical Punishment and Aggression
A. Retrospective Research
Six studies have questioned individuals for their retrospective recall of physical punishment experiences during childhood. In a nationwide survey of 1,176 adult respondents, Owens and Straus (1975) found a significant positive correlation between the frequency of interpersonal violence received as a child and approval of the use of violence interpersonally. Their measure of violence received as a child merged physical punishment (e.g., spanking) and abusive violence (e.g., punching, choking). No direct measure of aggressive behavior was made.
The remaining five studies focused on self-reported aggressive behavior in adults. Bryan and Freed (1982) questioned 170 community college students about their history with physical punishment and their self- reported "problems with aggression." They found that students who reported having received a "high" amount of corporal punishment reported significantly more problems with aggression (among other difficulties).
The remaining four studies specifically examine family violence as an outcome. Parke and Collmer (1975) found that abusive parents often had recollections of "physically punitive childhood experiences." The recollected violence was usually severe enough to be regarded as abuse, rather than as physical punishment. In 1977, Carroll studied 96 adults and found that 36.6% of those who had rated their childhood experiences as "high" physical punishment were violent, compared to only 14.5% of those who reported experiencing "low" physical punishment.
In a similar design, Caesar (1988) found that a sample of 26 wife batterers recalled more parental use of physical punishment than a sample of 18 nonviolent men (58% versus 31%). Finally, Gelles (1974) found that respondents who recalled being hit by their parents frequently (six or more times per year) were far more likely to physically fight with their spouse than were respondents who recalled being infrequently hit.
B. Cross-Sectional Research
Cross-sectional research designs have examined the co-existence of physical punishment and aggression in children. For example, Straus (1983) found, in a nationally representative sample of children (whose ages ranged from 3 to 17 years old), that 15% of children who were not physically punished "repeatedly and severely attacked a sibling," compared to 40% of children who were physically punished (but not abused), and 76% of children who were repeatedly abused.
In contrast to these findings, Yarrow, Campbell and Burton (1968), in a replication of Sears, Maccoby and Levin's (1957) design, found no statistically significant correlation between their measure of parental use of physical punishment and child's aggression. Further, Schuck (1974) conducted path analyses on the data from both Sears, Maccoby, and Levin (1957) and Yarrow, Campbell and Burton (1968). The purpose was to examine the impact of physical punishment on aggressive behavior. He found that the two sets of data yielded similar path results: in both cases, physical punishment was not related significantly to the child's aggression.
One study of adolescents noted a negative relationship. DiLalla, Mitchell, Arthur, and Pagliocca (1988) interviewed the families of 13 court-referred delinquents and found that the more aggressive the parental punishment (parent-reported), the less aggressive the offense of the delinquent. However, this study contains a significant weakness: even the authors strongly questioned the validity of their measure of physical punishment, since interviewers represented the juvenile court system and parents thus frequently failed to report any use of physical punishment.
To summarize, retrospective studies all found that violent and aggressive adults recall receiving more physical punishment than do their nonaggressive counterparts. Of ten cross-sectional studies examining a child's aggressiveness and his/her exposure to physical punishment, seven found a positive, linear relationship, two found no relationship, and one found a negative relationship. Thus, while much of the retrospective and cross-sectional data favors the hypothesis of a positive, linear relationship between physical punishment and aggression, support is not universal.
Furthermore, retrospective and cross-sectional research cannot discern causal direction. Thus, the above literature suggests equally that physically-punished children become aggressive and that aggressive children are more often physically punished (Bell, 1979). In addition, cross-sectional research can only focus on childhood aggression, which cannot be equated with adult violence. Retrospective designs are able to assess adult violence, but have been criticized as potentially inaccurate due to their reliance on human memory (Eron, 1982). Because of these limitations, a few studies have sought to prospectively examine the association between physical punishment and the later development of aggressive or violent behavior. We turn to this literature now.
C. Prospective Research
Seven longitudinal studies have examined parental use of physical punishment and the development of aggressive or violent behavior. The first two of these studies are descriptions of clinical samples and the researchers' observations, involving no statistical analysis or systematic assessment procedures. Rigdon and Tapia (1977) reviewed the clinical histories of eighteen children who were referred to a mental health clinic because of violent behavior towards animals and found that "most" of the children who remained violent two to six years after the initial evaluation were the products of a "chaotic home situation with aggressive parents who administered harsh corporal punishment." Nagaraja (1984) followed 200 boys between the ages of 10 and 15 and reported that physical punishment "was found to increase the occurrence of the target behavior [aggression]." Although both of these studies provide interesting observations, they are of limited use methodologically because of their use of somewhat vague terminology and their lack of systematic analyses.
Quantitative longitudinal studies have also been conducted, ranging from a span of three to twenty-plus years in length. Singer, Singer, and Rapaczynski (1984) conducted a three-year longitudinal study of 55 children and found a statistically significant, positive correlation between "power-assertive child rearing" scores (i.e., child rearing which emphasizes "control and physical punishment") at age six and aggression at age nine years. Lefkowitz, Eron, Walder, and Huesmann (1977) followed 427 children over ten years (ages eight to eighteen). They questioned parents about any use of "spanking" or "slapping" as a punishment. Further, parents were rated on their use of other punitive punishments (e.g., withdrawing love), yielding a "punitive" score. Parents with low scores were termed "very permissive" and were presumed to have utilized positive reinforcement more frequently than punitiveness, although this was not assessed directly. At Time One, the researchers found a significant negative relationship between physical punishment and child's aggression -- - but only in boys who were closely identified with their fathers. In other boys, the relationship was a positive one. Ten years later, at Time Two, Lefkowitz and his colleagues found that physical punishment at age eight was positively correlated with aggression at age eighteen, but again only for boys. However, as Gelles (1974) had found, this relationship was in fact curvilinear: moderately-punitive parents produced the least aggressive boys (versus very permissive parents and very harsh parents).
Finally, McCord (1979) studied 253 boys over approximately 20 years. At Time One, she classified parents as "aggressive" when they either "used little restraint" (e.g., threw objects) or who "regularly used physical force or very harsh verbal abuse" with their sons. At Time Two, McCord gathered official correctional records on the subjects and found that parental aggression significantly accounted for 3.6% of the variance in "personal" crimes (i.e., violent crime such as rape and murder) as an adult.
Other prospective research has failed to find any relationship between physical punishment and later aggression. Sears (1961) followed 160 children over seven years but failed to find any significant relationship between physical punishment at age five and self-reported attitudes towards aggression at age twelve. Similarly, Johannesson (1974) studied 212 children and found no relationship between physical punishment at nine to 24 months of age and aggression at ages ten to twelve years. He found that nonaggressive children were "smacked" by their parents just as often as children who were rated as "aggressive" by their teachers.
In summary, then, five of the seven prospective studies offer a quantitative design. The two qualitative studies of clinical populations found that most aggressive individuals had histories of parental physical punishment. In contrast, the five quantitative studies offer mixed results. Two (Johannesson, 1977, and Sears, 1961) found no evidence for a relationship between physical punishment and aggression or attitudes towards aggression. Three others (Lefkowitz, Eron, Walder, & Huesmann, 1977; McCord, 1979; Singer, Singer, & Rapaczynski, 1984) found a positive relationship between physical punishment and later aggressive behavior. No studies noted a negative relationship.
These mixed results may be partially attributable to the different ages of the children assessed; Johannesson (1977) and Sears (1961) studied the physical punishment of younger children (aged 9 months to five years), while Lefkowitz, Eron, Walder, and Huesmann (1977) and Singer, Singer, and Rapaczynski (1984) studied six- and eight-year-olds. McCord's subjects, similarly, ranged in age at Time One between 5 and 13 years, with a mean age of 10.5 years (SD-l.6). The physical punishment of older children is certainly more unusual and more deviant (Straus, 1983), and thus the positive results noted in these studies may be due to the detection of more deviant parental rearing practices in general (including, but not necessarily limited to, the use of physical punishment).
An alternative hypothesis for the difference noted is differential assessment of aggression. Singer, Singer, and Rapaczynski (1984) had the mothers report on their sons' aggression, and Lefkowitz, Eron, Walder, and Huesmann (1977) utilized a peers' rating of aggressive fighting in school. McCord (1979) utilized official records of violent crimes. In contrast, Johannesson (1977) utilized a teacher's rating of aggression which included "quarreling"; therefore, children who demonstrated no physical aggression may have been included in the "aggressive" group. Similarly, Sears (1961) did not measure aggressive behavior at age 12, but rather, attitudes towards aggression. In summary, it appears that the studies which concentrated on the physical punishment of older children and their subsequent physically aggressive behavior were those which noted significant and positive relationships.
D. Summary of Retrospective, Cross-Sectional, and Prospective Studies
Most, but not all, of the cross-sectional and retrospective research reviewed above found a positive relationship between physical punishment and aggressive behavior. Correlations between physical punishment and aggression in the best-designed studies range from .21 to .32 (with stronger correlations for males). Prospective research examined how well this association is maintained longitudinally, and suggested that when physically aggressive behavior is directly measured in older children (over five years old), corporal punishment may be included in the group of variables which potentially contribute to aggressive behavior.
One issue which limits all research in this area is the definitions used for "violence" and "aggression." Frequently children are rated on a variety of aggressive acts, which are then summed to yield a total score. Children with low scores are compared with those at the high end of the scale (e.g., Larzelere, 1986); unfortunately, this method weakens the consistency between studies, since no standardized measures appear to be widely used. Further, raters of aggression range from peers (e.g., Lefkowitz, Walder, & Eron, 1963) to teachers (e.g., Sears, 1961), to parents (e.g., Straus, 1983). The rating of physical punishment appears to be more uniformly assessed via self-report on the part of parents (except, of course, in retrospective studies).
The remainder of this paper will be devoted to examining more specific findings and methodological issues in this literature. For example, is the nature of the positive relationship noted above linear or curvilinear? Can low correlations possibly be masking a curvilinear relationship? How do males and females compare when examining the association between physical punishment and aggression? And finally, what are the methodological issues which bring into question the strength of the conclusion that physical punishment may be causally related to the development of aggression and/or violence?
III. Methodological Issues
A. The Relationship between Low Physical Punishment and Aggression
Eight studies reported or utilized data analysis methods which can help clarify the relationship between low or no use of physical punishment and the development of aggression. All eight studies agree that severely-punished children were among the most aggressive (Larzelere, 1986; Straus, 1983; Eron, 1982; Bryan & Freed, 1985; Gelles, 1974; Lefkowitz, Eron, Walder, & Huesmann, 1977; Sears, 1961; Lefkowitz, Walder, & Eron, 1963). However, some studies found that "low" physical punishment was associated with higher aggression (Gelles, 1974; Lefkowitz, Eron, Walder, & Huesmann, 1977; Sears, 1961), while in contrasting studies, "low" physical punishment was associated with lower aggression (Lefkowitz, Walder & Eron, 1963; Larzelere, 1986; Straus, 1983; Eron, 1982; Bryan & Freed, 1985).
Central to this issue is the assessment of "low punishment" conditions. Unfortunately, the data reported (or not reported) in these studies makes it very difficult to address this issue here. In some studies, the "low punishment" condition consisted of parents who literally used no physical punishment (e.g., Straus, 1983), while in other studies, the "low punishment" condition included parents who used some physical punishment (usually to an imprecise extent) (e.g., Bryan & Freed, 1982). Other studies do not report how they formed their comparison groups (e.g., Sears, 1961; Eron, 1982), while Lefkowitz, Walder, and Eron (1963) based their comparison groups on the number of types of physical punishments used, rather than on their frequency. These inconsistencies make it impossible to discuss the linear versus curvilinear issue without a great deal of speculation about unreported research methods; however, it seems plausible that the direction of the relationship between low physical punishment and aggression may be determined by related and interacting factors (e.g., other parenting factors). The importance of additional variables and their impact on the physical punishment - aggression relationship will be discussed below.
A final difficulty in studying the impact of low or no physical punishment is accuracy. Given the large percentage of the population which utilizes physical punishment, the possibility of any study locating a group of parents who literally utilize no physical punishment seems low. Although the number of these parents who deny using any physical punishment probably varies greatly given the length of the referent period, it is very plausible that parents often use mild physical punishment which they later fail to recall (e.g., slapping the hand of a toddler). In summary, therefore, any study which attempts to measure "no" parental use of physical punishment may be flawed in this respect.
B. Child Abuse and Physical Punishment
As stated earlier, parental aggression may be conceived of as occurring on a continuum, ranging from abusive violence to mild physical punishment. One of the greatest difficulties in studying physical punishment is the "teasing apart" of the effects of abusive violence versus the effects of more common and mild parental aggression.
It is difficult to separate these two parental behaviors because it appears that they may be related to one another. The use of physical punishment seems to increase the likelihood that child abuse will occur (Kosky, 1983; Maurer, 1974). In an attempt to lower the incidence of child abuse, Sweden passed a 1979 civil law which forbids parental use of physical punishment (Feshbach, 1980). The association between corporal punishment and abuse, taken together with the high prevalence rate of child abuse measured in a representative sample (Straus & Gelles, 1990), makes it probable that child abusers were included among the parents studied in the physical punishment literature.
Because child abuse is causally related to aggression (Spatz Widom, 1989), the parents who are abusers (rather than the parents who are only physical punishers) may be responsible for the associations with aggression noted above. Thus, one important factor to control for, in the study of physical punishment and aggression, is child abuse. When the presence of child abuse is controlled for, does the association between physical punishment and aggression remain significant? Does parental violence have to be extreme before it is associated with the development of deviant violence in the child?
Unfortunately, only two studies in this area directly address this issue by deliberately assessing both child abuse and physical punishment. In one study of a clinical sample comparing wife batterers to non-batterers, Caesar (1988) questioned subjects about their exposure to parental violence in their families-of-origin. They assessed parental violence at three levels: (1) "spanking"; (2) "use of a switch, belt, razor strap, paddle, etc."; and (3) "beating." Batterers had experienced, relative to non-batterers, more parental violence at the second level. However, batterers and non-batterers did not differ in the proportion exposed to "spanking." Whether the second level of violence here is considered to be abusive or not, spanking alone was not associated with wife-battering. However, this finding is of limited generalizability because, in a (nonrandom) clinical sample, spanking may have been equally related to battering (in the index subjects) and to other psychological difficulties (evidenced by the controls).
In addressing this issue, the differential effects of child abuse and physical punishment have also been investigated in a nonclinical, nationally representative sample. Straus (1983) considered, cross-sectionally, a group of several thousand children ranging from three to 17 years of age. He interviewed their parents as part of the 1975 National Family Violence Survey, which assessed for both physical punishment (e.g., slapping) and child abuse (e.g., hitting with a closed fist). He then compared the rates of aggression between children whose parents had only used physical punishment and children whose parents had used physical punishment and abusive violence. The results noted that aggression was much more common in abused children, when compared to physically punished children. However, physically punished children (while not as aggressive as abused children), were more aggressive than children whose parents used neither physical punishment nor abusive violence.
While only these two studies directly assessed for child abuse, other studies may offer revealing information, particularly those which separated parents into "low", "moderate", and "severe" punishers. It is most probable that abusive parents make up part of the "severe" category but are less likely to be included in the "moderate" or "low" categories. Therefore, another way of examining the impact of child abuse is to ask, does only moderate physical punishment increase aggression? If no, then it remains possible that abuse accounts for the relationship between physical punishment and aggression. If yes, then it is more likely that physical punishment may be related to aggression independently of abuse.
Lefkowitz, Walder, and Eron (1963) compared mean aggression scores between children whose parents refrained from using physical punishment, children whose parents had only used physical punishment once annually, and children whose parents used more than one type of physical punishment annually (resulting in more frequent physical punishment). He found that the crucial distinction in mean aggression lay between the "no physical punishment" children and the other children, implying that any (even very infrequent) use of physical punishment serves to increase aggression in children. Similarly, Eron (1982) found a linear relationship between low, medium, and high physical punishment and peer-nominated aggression in children, implying a difference between the "low"- and "medium" -punished children.
Other studies have implied the opposite: namely, that moderate punishment does not serve to increase aggression in children. Gelles (1974) compared children's aggression between parents who did not use physical punishment, those who used it less than six times a year, and those who used it on a daily to monthly basis. He found that the moderate users of physical punishment had reared the least aggressive children, even less aggressive than the "no" punishment parents (or the "high" punishment parents) Gelles' methodology makes it unlikely that child abusers in the sample were placed in the "moderate" physical punishment category (although there is no way to be certain of this). Thus, his work implies that the use of moderate physical punishment in the absence of abusive violence may not increase aggression at all; in fact, it may decrease aggression. Two others studies (Lefkowitz, et al., 1977; Sears, 1961) support this hypothesis by finding similarly that the children of moderate users of physical punishment were the least aggressive subjects.
Bryan and Freed (1982) questioned college students about their own aggression and their recollection of their parents' aggression. They then divided the parent scores on both severity and frequency into "low", "medium", and "high" Corporal Punishment (CP) groups. Their "high CP group had a much higher average rate of severe violence (e.g., strapping, whipping, punching, kicking, beating up, tying up) than the other two CP categories, making it likely that this category represented the child abusers in the sample. Results noted that although the students exposed to "high CP" had more problems with aggression than the two other groups, the "medium" and "low" CF groups did not differ significantly from each other. Like Gelles' (1974) results, this study does not support the hypothesis that moderate or low use of physical punishment serves to increase aggressive behavior in children.
Larzelere's (1986) results also suggest support for this conclusion. He used Straus' (1983) data to compare children's aggression based on the frequency with which their parents used spanking. He found that regardless of the age of the child, parents who spanked their children frequently (20 plus times a year) had children who were more aggressive with their siblings than did parents who spanked their children moderately (6 times a year) or parents who spanked their children minimally (once a year). Unfortunately, mean values were not reported to aid comparisons; but it appears (from graphed results), that among preschool-age subjects, the significant difference lies between the very frequent spankers and the minimal spankers. Moderate spanking did not appear to significantly elevate the risk of aggressive behavior in very young children.
In summary, then, nine studies address the issue of moderate or low use of physical punishment versus child abuse by assessing for severity and/or frequency of use of parental aggression. Of these nine, six failed to support the hypothesis of a linear association between physical punishment and aggression: three found that moderate physical punishment and low/no physical punishment conditions did not differ in their association with aggression, and three found that moderate physical punishment was actually associated with the lowest levels of aggression. The three remaining studies (Eron, 1982; Straus, 1982; Lefkowitz, Walder, & Eron, 1963) found that moderate physical punishment did increase aggression in children (over low/no physical punishment).
Thus, two-thirds of the studies reviewed here did not support a linear relationship between physical punishment and aggression. If future research confirms the validity of a curvilinear relationship, then the direction of the relationship may be clarified. Recall that one hypothesis states that a positive association between physical punishment and aggression may reflect only the fact that aggressive acts provoke more physical punishment. A curvilinear relationship may refute this hypothesis by its distinction of a "moderate punishment - low aggression" group of subjects. Among these subjects, is it more plausible to assume that low aggression provokes moderate physical punishment, or that moderate physical punishment results in low aggression? This author believes the latter to be the more plausible interpretation. However, it is also possible that these studies detected a group of families with low-aggression children who transgress in other ways, and it is these other transgressions which may provoke the physical punishment. In other words, the relationship between physical punishment and aggression may interact with other parenting factors, or may even be spurious among some families (possibly those who use other methods with physical punishment), as suggested by several authors reviewed here (see below).
C. Males versus Females
Here, results tend to be consistent. Five studies compared correlations between males and females; four of these found that the correlations for male subjects were stronger (Owens & Straus, 1975; Becker, Peterson, Luria, Shoemaker, and Hellmer, 1962; Eron, 1982; and Lefkowitz, Eron, Walder, and Huesmann, 1977). Correlations for females were generally not statistically significant, while for males they varied between .21 and .35 (and were statistically significant). Johannesson (1977) found no correlation between physical punishment and aggression for males; however, this is probably a reflection of his lack of findings in general. Like the other studies, he found no significant correlation for females.
D. Physical Punishment and Other Parenting Behaviors
Another issue which is important in evaluating the strength of the connection between physical punishment and the development of aggression is raised by Parke and Slaby (1983), who note that other adult characteristics may be associated with the use of physical punishment, and that it may be these behaviors which are responsible for associations noted (rather than the use of physical punishment per se). Research has implied that adults who choose to use physical punishment may be different from adults who prefer other methods of punishment or behavior modification. For example, one study found that "close-mindedness" and "neuroticism" were highly correlated with the use of corporal punishment by public school teachers (Rust & Kinnard, 1983). Another study found that parental anger-proneness was related to the use of "harsh" parental punishment (Engfer & Schneewind, 1982). Becker, Peterson, Luria, Shoemaker, and Hellmer (1962) found that "hostile" parents tended to use more physical punishment than less hostile parents. Family factors have also been tentatively related to the use of corporal punishment; such factors include socioeconomic status (Magmer & Ipfling, 1973), cultural values (Escovar & Escovar, 1985), larger families (Wagner, Schubert & Schubert, 1985), a high degree of intrafamilial conflict (Engfer & Schneewind, 1982), and marital satisfaction (Kemper & Reichler, 1976).
The relationship between personality and parenting factors and parental choice of punishment method must be very complex: these few studies can only be suggestive. Before we can definitively tease apart the impact of physical punishment from the impact of other parental and family factors, we must have a much clearer idea of what factors are most strongly associated with the use of physical punishment. Nevertheless, four of the studies on physical punishment and the development of aggression have attempted to control for other parenting and family factors. Although these can only be considered suggestive and by no means conclusive, they will be presented below.
One retrospective study (Carroll, 1977) investigated the interrelationship of parental "warmth", use of physical punishment, and the development of aggression. All subjects were questioned about their recall of their parents' warmth and nurturance using items from the Bronfenbrenner-Devereux Parental Activity Inventory. The subjects were then divided into four groups: low warmth/low punishment, low warmth/high punishment, high warmth/low punishment, and high warmth/high punishment. These four groups were then compared on their mean frequency of violent behavior, and, as predicted, the low warmth/high parental punishment families had the highest mean frequency. However, the only statistically significant effect was for parental warmth; no other main or interactive effects achieved significance. Thus, although the data is in a direction which implies the importance of physical punishment above and beyond other parenting factors, physical punishment may not be related to aggression when parental warmth is accounted for.
Larzelere (1986) studied the interaction of physical punishment and parental use of discussion in his cross-sectional study. He found that, in younger children, parental discussion did not interact with physical punishment in the association with aggressive behavior. In older children, a different picture emerged. In preadolescents and adolescents, "the combination of frequent spanking and minimal discussion was particularly associated with frequent aggression"; in addition, when parents frequently used discussion, no association remained between use of physical punishment and aggression.
Finally, in their study, Eron and his colleagues noted that physical punishment increased aggression only in boys who lacked close identification with their fathers. When such identification was present, physical punishment appears to decrease aggression in boys (Eron, Walder, Huesmann, & Lefkowitz, 1974). Lefkowitz, Eron, Walder and Huesmann (1977) noted the same result. These findings may help explain why moderate spanking was associated with the least aggressive children in some studies, and more aggressive children in others (see sections above).
These studies, while only suggestive, imply that other parenting factors (especially the warmth and intimacy of the relationship between boys and their punishers) may be very important when considering the impact of physical punishment on children's aggressive (or lack of aggressive) behavior. These studies suggest that in the context of affectionate and effective parenting, physical punishment may not increase a child's risk of aggressive behavior.
Of the studies which separated parents on frequency of use of physical punishment, most found that no or "low" use of punishment was associated with less aggression in the child. However, a substantial minority (more than one-third) found the opposite: "low" or no use of physical punishment may increase a child's aggression.
Of the two studies which directly assessed and controlled for child abuse, the methodologically stronger study found that even when abuse was controlled for, physical punishment was associated with an increase in the child's aggression. However, of the nine studies which teased apart moderate punishers from severe (and possibly abusive) punishers, three observed that moderate physical punishment produced the least aggressive children, and three more found no difference in aggression between children of low and moderate users of physical punishment. Three others found that moderate physical punishment did increase aggression. Thus, two-thirds of these studies imply that moderate physical punishment does not increase aggression which weakens support for a hypothesized linear association between physical punishment and aggression. In addition, as detailed above, studies which support a curvilinear relationship tend to weaken the hypothesis that more aggressive children provoke more physical punishment.
One more consistent finding was that severe or frequent parental use of physical punishment was associated with an increase in the child's aggression. Longitudinal, prospective research tended to support this: these studies found that when physical aggression is studied directly (not, for example, via attitudes towards aggression) in older children, there is a positive association between use of severe physical punishment and aggression in the child. This suggests that severe physical punishment may contribute etiologically towards the development of aggressive behavior.
It should be noted that the relationship appears strongest among school-age and older males. Finally, the four studies which examined other parenting characteristics in interaction with physical punishment suggested that family factors such as identification with parents, high parental warmth, or use of reasonable discussion may eliminate any noxious effects which occur as a result of the use of physical punishment.
All of these conclusions can only be termed tentative, because the literature of this area suffers from significant methodological problems. For example, the child development literature makes it clear that any study examining parental aggression must account for child abuse; however, only two studies directly assessed for this. Further, despite the noted relationship between alcohol and drug abuse and parental violence, no research has considered these factors when studying physical punishment. The issue of causality also requires further work. It would be best addressed by prospective research which begins at an early enough age to note, at Time One, the presence of physical punishment without any presence of aggressive behavior. Finally, such studies must examine the development of aggressive behavior in light of the full scope of parental violence and relevant family factors. Ambitious as such a project would be, it is necessary to ascertain any possible deleterious effects of a childrearing practice as widespread as physical punishment. Given the current literature it appears that a positive, curvilinear, causal relationship between physical punishment and subsequent aggression plausibly exists, but cannot be termed conclusively true at this time.
Anderson, K., & Anderson, D. (1976). "Psychologists and spanking". Journal of Clinical Child Psychology, Fall, 46-49.
Becker, W., Peterson, D., Luria, Z., Shoemaker, D., & Hellmer, L. (1962). "Relations of Factors Derived From Parent-Interview Ratings to Behavior Problems of Five-Year-Olds". Child Development, 33 509-535.
Bell, R. Q. (1979). "Parent, Child, and Reciprocal Influences". American Psychologist, 34(10), 821-826.
Bryan, J., & Freed, F. (1982). "Corporal Punishment: Normative Data and Sociological and Psychological Correlates in a Community College Population". Journal of Youth and Adolescence, 11(2), 77-87.
Caesar, P. (1988). "Exposure to Violence in the Families-of-Origin Among Wife-Abusers and Maritally Nonviolent Men".Violence and Victims, 3(1).
Curtis, G. C. (1963). "Violence Breeds Violence". American Journal of Psychiatry, 120, 386-387.
Deley, W. (1988). "Physical Punishment of Children: Sweden and the USA." Journal of Comparative Family Studies, 19(3), 419-431.
DiLalla, L., Mitchell, C., Arthur, M., & Pagliocca, P. (1988). "Aggression and Delinquency: Family and Environmental Factors". Journal of Youth and Adolescence, 17(3), 233-246.
Dollard, J., Doob, L., Miller, N., Mowrer, 0., & Sears, R. (1939). Frustration and Aggression. New Haven: Yale University Press.
Engfer, A., & Schneewind, K. (1982). "Causes and Consequences of Harsh Parental Punishment: An Empirical Investigation in a Representative Sample of 570 German Families". Child Abuse & Neglect, 6(2), 129-139.
Eron, L. (1982). "Parent-Child Interaction, Television Violence, and Aggression of Children". American Psychologist, 37(2), 197-211.
Escovar, L., & Escovar, P. (1985). "Retrospective Perception of Child-rearing Practices in Three Culturally Different College Groups". International Journal of Intercultural Relations, 9(l), 31-49.
Feshbach, N. (1980). "Tomorrow is Here Today in Sweden." Journal of Child Clinical Psychology, 9(2), 109-112.
Gelles, R. (1974). The Violent Home: A Study of Physical Aggression Between Husbands and Wives. Beverly Hills, CA: Sage Publications.
Gelles, R. (1978). "Violence Towards Children in the United States". American Journal of Orthopsychiatry. 48(4), 580-592.
Glueck, S., & Glueck, E. (1970). "Working Mothers and Delinquency". In M. E. Wolfgang, L. Savitz, & N. Johnston (Ed.), Sociology of Crime and Delinquency. New York: Wiley.
Hirschi, T. (1969 ). Causes of Delinquency. Berkeley: University of California Press.
Johannesson, I. (1974). "Aggressive Behavior Among School Children Related to Maternal Practices in Early Childhood". In J. de Wit, & W. W. Hartup (Ed.), Determinants and Origins of Aggressive Behavior (pp.413-425). The Hague: Mouton.
Jones, J. (1990). "Violent Statistics". American Psychological Association Monitor, October.
Kemper, T., & Reichler, M. (1976). "Marital Satisfaction and Conjugal Power as Determinants of Intensity and Frequency of Rewards and Punishments Administered by Parents". Journal of Genetic Psychology, 129(2), 221-234.
Kosky, R. (1983). "Childhood Suicidal Behavior". Journal of Child Psychology & Psychiatry & Allied Disciplines, 24(3), 457-468.
Larzelere, R. (1986). "Moderate Spanking: Model or Deterrent of Children's Aggression in the Family?". Journal of Family Violence, 1(1), 27-36.
Lefkowitz, M., Eron, L., Walder, L., & Huesmann, L. (1977). Growing Up to Be Violent: A Longitudinal Study of the Development of Aggression. New York: Pergamon.
Lefkowitz, M., Walder, L., & Eron, L. (1963). "Punishment, Identification, and Aggression". Merrill Palmer Quarterly, 9, 159-174.
Leviton, H. (1976). "The Individualization of Discipline for Behavior Disorder Pupils". Psychology in the Schools, 13(4), 445-448.
Lowenstein, L. (1977). "The Value of Corporal Punishment". News and Views, 3(4), 14-15.
Magmer, E., & Ipfling, H. (1973). "On the Problem of Specific Social Class Punishments". Scientia Paedagogica Experimentalis 10(2), 170-192.
Magnusson, D., Stattin, H., & Duner, A. (1983). "Aggression and Criminality in a Longitudinal Perspective". In K. T. Van Dusen, & S. A. Mednick (Ed.), Prospective Studies of Crime and Delinquency (pp. 277-302). Boston: Kluwer-Nijhoff.
Maurer, A. (1974). "Corporal punishment". American Psychologist, August, 614-626.
McCord, J. (1979). "Some Child-rearing Antecedents of Criminal Behavior in Adult Men". Journal of Personality and Social Psychology, 37(9), 1477-1486.
Miller, D., & Challas, G. (1981). "Abused Children as Parents: A Twenty-five Year Longitudinal Study". National Conference on Family Violence Research, University of New Hampshire, Durham, New Hampshire.
Nagaraja, J. (1984). "Non-Compliance: A Behavior Disorder". Child Psychiatry Quarterly, 12(4), 127-132.
Owens, D., & Straus, M. (1975). "The Social Structure of Violence in Childhood and Approval of Violence as an Adult". Aggressive Behavior, 1, 193-211.
Parke, R., & Collmer, C. (1975). "Child Abuse: An Interdisciplinary Analysis". In E. M. Hetherington (Ed.), Review of Child Development Research (pp. 509-590). Chicago: University of Chicago Press.
Parke, R., & Slaby, R. (1983). "The Development of Aggression". In P. Mussen (Ed.), Handbook of Child Psychology: Volume IV (pp. 547-621). New York: Wiley & Sons.
Patterson, G. R., & Dishion, T. J. (1985). "Contributions of Families and Peers to Delinquency". Criminology, 23, 63-79.
Rigdon, J., & Tapia, F. (1977). "Children Who Are Cruel to Animals: A Follow-up Study". Journal of Operational Psychology, 8, 27-36.
Rust, J., & Kinnard, K. (1983). "Personality Characteristics of the Users of Corporal Punishment in Schools". Journal of School Psychology, 21(2), 91-98.
Schuck, J. R. (1974). "The Use of Causal Nonexperimental Models in Aggression Research". In J. de Wit, & W. W. Hartup (Ed.), Determinants and Origins of Aggressive Behavior, (pp. 381-389). The Hague: New Babylon.
Sears, R. (1961). "Relation of Early Socialization Experiences to Aggression in Middle Childhood". Journal of Abnormal and Social Psychology, 63(3), 466-492.
Sears, R., Maccoby, E., & Levin, H. (1957). Patterns of Child Rearing. New York: Harper & Row.
Sears, R., Whiting, J., Nowlis, V., & Sears, P. (1953). "Some Childrearing Antecedents of Aggression and Dependency in Young Children". Genetic Psychology Monograms, 47, 135-234.
Singer, J., Singer, D., & Rapaczynski, W. (1984). "Family Patterns and Television Viewing as Predictors of Children's Beliefs and Aggression". Journal of Communication, Spring, 73-89.
Skrabanek, P. (1988). "Cervical Cancer in Nuns and Prostitutes". Journal of Clinical Epidemiology, 41(6), 577-582.
Spatz Widom, C. (1989a). "Child Abuse, Neglect, and Violent Criminal Behavior". Criminology, 272, 251-271.
Spatz Widom, C. (1989b). "Does Violence Beget Violence? A Critical Examination of the Literature". Psychological Bulletin, 106(1), 3-28.
Straus, M. (1983). "Corporal Punishment, Child abuse, and Wife-beating: What Do They Have in Common?" In D. Finkelhor, R. Gelles, G. Hotaling, & M. Straus (Ed.), The Dark Side of Families: Current Family Violence Research (pp. 213-234). Beverly Hills, CA: Sage. Also in Straus and Gelles (1990) Physical Violence in American Families: Risk Factors and Adaptations to Violence in 8,145 Families. New Brunswick, NJ: Transaction Pub. (Former title: "Ordinary Violence, Child Abuse, and Wife Beating: What Do They Have In Common?").
Straus, M. A., & Gelles, R. (1990). Physical Violence in American Families: Risk Factors and Adaptations to Violence in 8,145 Families. New Brunswick, NJ: Transaction Pub.
Wagner, M., Schubert, H., & Schubert, D. (1985). "Family Size Effects: A Review". Journal of Genetic Psychology, 146(l), 65-78.
Wolfgang, M. E. (1983). "Delinquency in Two Birth Cohorts". American Behavioral Scientist, 27, 75-86.
Yarrow, M., Campbell, J. D., & Burton, R. V. (1968). Child Rearing: An Inquiry Into Research and Methods. San Francisco: Jossey-Bass.
The owner of this paper is Professor Murray A. Straus, The Family Research Laboratory, University of New Hampshire. He has graciously permitted its presentation on The No Spanking Page.
To visit the FRL Home Page:
Family Research Laboratory
To contact the FRL :