Corporal Punishment in Children – What Does It Accomplish?

How Many Parents Spank Their Children?

Gershoff reports that Straus and Stewart (1999) found that 94% of American parents spank their children by the ages of 3 or 4. Primarily they spank children up to age 5, one to two times per month. Gershoff notes too that the US Department of Health and Human Services (2001) indicated that 13 of every 1000 children are abused or neglected. Thinking that spanking leads to abuse, Gershoff investigated the effectiveness of spanking by conducting a meta-analysis. Holden, Baumrind, and Parke respond to her article, and she replies.

Gershoff began by surveying 300+ studies published over a 60 year period. She explains that there are few good studies out there, however, because:

  • Frequency and severity of corporal punishment are inconsistently defined and measured, with some only asking if corporal punishment has ever been used. This makes it hard to draw conclusions across studies.
  • Spanking can be either instrumental (done when the parent is calm and thoughtful) or impulsive (when the parent is distressed and emotional). This is particularly important as more emotional parents are more likely to be impulsive, and are more likely to make negative attributions about their children’s behavior and use corporal punishment believing it is justified and normal in this case. Thus, this distinction in future research is important.
  • Child arousal and mood is seldom assessed. This is important as over-arousal may lead to oppositionality and avoidance, which may decrease the impact of parental teaching and explain the poor positive effects of spanking.
  • Child perceptions of the parent’s wishes matter, as some misbehavior may be due to lack of parental clarity, and some due to child processes. Thus, future research likely should distinguish among kinds of misbehavior.
  • Child attributions about the parent’s actions likely matter too. Attributions that the parent acts in the child’s best interest and supports the child are likely to spell different consequences for physical punishment than attributions of hostility and anger alone.
  • Similarly, child attributions about the reason for compliance – oversight of authority figures versus internal controls and morals – likely plays a role over time too in increasing or decreasing acting out behavior.

She selected 88 studies that were reliable – however, only 44% reported participant ethnicity, 27% reported SES, and only 4 studies reported parenting style. This is of concern given how intertwined ethnicity and SES are, and how intertwined stress and social support are. Further, cultural and community values about compliance could alter the frequency of corporal punishment, as well as the children’s, parents’, and community’s attitudes about it. Religious adherence was a confusing variable, as some studies indicate some religious groups are more likely to use corporal punishment and others are not. Geographic region was tied to corporal punishment as well, with Southern states and Missouri being more likely to use corporal punishment.

What is Corporal Punishment? Does It Work?

Gershoff reviewed 11 variables associated with Corporal Punishment (CP), including:

  • Immediate Compliance – She found that CP is associated with increased immediate compliance.
  • Moral Internalization – She found that CP decreases internalization of moral rules. This is concerning in that parents are more likely to use corporal punishment when they believe the child is at fault for some misbehavior. Thus, using a method that decreases moral internalization to respond to a failure to adhere to internal rules the child should have known is likely to perpetuate the problem.
  • Aggression – She found that CP is associated with increased aggression. This is especially troublesome, she notes, in that parents are more likely to use aggression to stop aggression. However, one study showed that use of corporal punishment to halt aggression increased risk for aggressive behaviors by 50%, regardless of whether the parent or the teacher rated the child’s behavior. Use of aggression after being physically punished for aggressive behavior is likely to be seen as an escalation of misbehavior, which was also associated with greater use of corporal punishment. Thus, corporal punishment is likely to perpetuate the problem.
  • Antisocial Behaviors – She found that CP is associated with increased antisocial behaviors. This was found most strongly for boys, and for children between the ages of 10 and 12. This is also troublesome as boys were more likely to be spanked, and if spanking increases antisocial behaviors, spanking to stop them is likely to perpetuate the problem. Gershoff in fact did find that CP is associated with increased risk of adult criminal behavior.
  • Quality of Parent-Child Relationship – She found that CP is associated with decreased quality of the parent-child relationship. This is more troublesome because most spankings happen between 5:00PM and bedtime, which comprises the majority of parent-child time together for most children. Spankings were also more likely to happen if the child’s misbehavior placed them at some risk for harm, and protecting the child is part of the parent-child relationship. This is also more troublesome, as spanking can lead children to think that aggression is common in relationships with loved ones. Gershoff in fact did find that CP is associated with increased risk of victimization from abusive relationships in adulthood.
  • Mental Health – She found that CP is associated with decreased mental health outcomes. This is concerning, as children ages 5 to 8 are most at risk for severe corporal punishment, ages at which significant emotional, social, and cognitive development happens.
  • Adult Abusive Behavior – She found that CP is associated with increased adult abusive behavior. She reports studies have shown that 2/3s of abusive parent-child incidents begin as an effort to discipline the child and “teach them a lesson.” If this means that adult antisocial behavior is more likely after being spanked as a child, given that other research shows antisocial parents are at greater risk to abuse children, then this could mean that spanking one’s child may increase the risk of abuse for one’s grandchildren.

Most results indicated spanking increased negative behaviors and decreased positive ones. The median effect size was .25, which by Cohen’s standards is good. Results were also highly tolerant to “file drawer studies” meaning that the results of this analysis would not change even if you added many new studies that showed the exact opposite conclusions. Gershoff was careful to note that this does not mean that a child spanked once is likely to grow up, beat people, steal, and go to jail. Rather, as the frequency and severity (abusiveness) of corporal punishment increases, so too does the risk for negative outcomes. Clearly, she notes, a majority of children are spanked within some limit we define as “normal,” and the majority do not grow up with such negative outcomes. However, for a significant group, they are physically disciplined in a way that goes beyond whatever limits constitute “normal” and are thus at increased risk.

While some have studied child characteristics that seem to lead to abuse, Gershoff focused more on adult characteristics, and found corporal punishment was more likely when parents were younger, female, more aggressive, depressed, inconsistent and ineffective in their parenting style, or reliant on authoritarian techniques. These are the same risk factors for abusive behaviors. Thus, Gershoff argues that the risks of “normal” spanking turning into abusive discipline deserve serious consideration.

What Do Experts Say is Good Discipline?

  • There is a difference between discipline and abuse
    • Holden (echoing Gershoff’s concerns) notes that because so few studies really accurately differentiate between abuse and non-abusive corporal punishment, between for example an open-handed smack on the butt or hand and a closed fist to the face or chest, it is hard to draw conclusions about spanking as opposed to abuse based on these studies. Even in studies where a researcher does try hard to ask needed questions to separate normal spanking from abuse, parents may not be honest in their reports.
    • Baumrind echoes this same concern about abusive parents “sneaking,” so to speak, into the studies Gershoff includes, but goes further. She argues that Gershoff’s inclusion criteria were unfair, and Gershoff added into her analysis of corporal punishment quite a few studies that were really about physical abuse. She cites for example that 16% of her included studies allowed hitting with an object (like a belt or a stick) as corporal punishment. She notes one study including prison inmates indicated 49% reported physical punishment from their fathers, and 44% reported physical abuse. She seems to say that only that 5% (49% – 44%) received punishment without abuse. Other studies used the term “beat” to describe what the parent did to the child, which would seem to clearly indicate the study was focused on abuse.
    • Gershoff replies that while these criticisms are valid from a scientific standpoint, they likely do not diminish the study results. A 1995 Gallup survey of 900 parents found 28% used objects to spank their children from ages 2 to 8, and of the 23 States that allow corporal punishment in the school, most recommend a paddle. She also notes common language includes the term “beat” and “whip”, and so the use of such language and the use of an object to spank a child that Baumrind objects to, while possibly abusive, are also far from uncommon.
  • Spanking is used with others kinds of discipline
    • Spanking is also a “packaged variable” according to Parke, and as such it is difficult to separate it from other discipline techniques. Holden notes that corporal punishment might occur after other kinds of punishment have failed to work. Holden and Baumrind note corporal punishment might also serve to strengthen other parenting techniques as a kind of “back up” intervention. Baumrind notes that corporal punishment might also serve as a short-term parenting technique, designed to end noncompliance (which Gershoff showed it likely does), so that other, non-corporal, parenting techniques can be used over the long-run.
    • Thus, it is difficult to separate the simple results of corporal punishment from the simple results of other discipline methods, as well as from the complex results of the interaction of corporal punishment and other discipline methods. Beyond this, Parke notes, parenting styles, timing, parent-child relationship quality, quality of verbal explanations paired with spanking… all come into play as well. The impact of a spanking from a warm and normally flexible parent followed by a reasonable explanation the child can understand, is likely quite different from the impact of spanking from a cold and strict parent with no explanation. Gershoff agreed in her response, but notes that not all children have a warm and loving relationship with their parents. Thus, this last point carried little weight with her.
  • Cultural issues make a difference too
    • Baumrind also points out that while some studies did not include the ethnicity of participants, some did. Those that looked at minorities often did not find that spanking lead to increased aggression.
    • This is not a new criticism of the research. A number of studies have found this, and attributed the differing outcomes from corporal punishment in minority parents to the differing environments in which they have to parent. Thus, corporal punishment in a high crime neighborhood as part of a controlling parenting style is more likely to be part of the parents’ efforts to assure their children’s safety, and less a part of the parents’ need for control and authority. Her point though, is that aggregative findings from studies of mostly majority parents may not be applicable to minority parents.
  • Parenting is many things
    • Parke quotes Bronfenbrenner (1974), as saying that studies of parenting behavior are “the science of behavior of children in strange situations with strange adults.” Studies that were more true to “real life” could examine situations in which parents disagree about corporal punishment, are experiencing relationship distress, or must parent amid more real world distractions and contradicting demands. Some studies could look at the difference between the results of corporal punishment in families that do and do not have a clearly stated philosophy about its use.
    • This might also be helpful in creating a developmental perspective of corporal punishment, one that included both its differing impacts on children of different ages, as well as its differing use across parents of different ages and after life changes, like divorce or remarriage. Finally, newer research has revised many theories by including physiological differences and processes alongside the psychological ones. Understanding the physiology of parent-child discipline might expand our understanding greatly.
    • Baumrind agreed with this, offering for example that spanking a 3 year old may lead to fear and greater compliance, while spanking a 10 year old might lead to anger and oppositionality. Further, parents appear to change their views over time. She notes one study that showed that adults, after having children, generally decreased their opinion of the effectiveness using corporal punishment. Those that did not reported that they used corporal punishment only after other techniques had failed. Thus, underlying changes in parental views would be a useful addition to a developmental model.

How Reliable are Studies on Corporal Punishment?

  • It depends on how the study is designed
    • Baumrind also points out that even when the study is well designed, the problem remains that many of these studies are cross-sectional rather than longitudinal. Cross-sectional studies look at samples of people now, say 5-8 year olds, 9-12 year olds, and 13-16 year olds, and try to draw conclusions about how children develop over time. They assume that the 13-16 year olds used to be just like the 9-12 year olds, and they used to be just like the 5-8 year olds. While faster, this kind of study is likely more problematic, as the 16 year olds today grew up in different homes, times, and attitudes about parenting than most 5 year olds today. Thus, a better design is to study the same group of children when they are 5-8, again when they are 9-12, and again when they are 13-16.
    • Gershoff acknowledged this, but again pointed out that the research we have, even if flawed, is all we have to go on. Her conclusions are based on the available research.
  • It depends on who participates
    • As noted above, parents reported on their own behavior in most studies of corporal punishment. While this means some abusive parents might have minimized their actions, Baumrind points out that this means too that findings will be convoluted because those who administer the spankings are likely to see the need for spankings. In other words, a parent who uses corporal punishment to respond to a child’s aggression may report high levels of aggression in order to justify the continued need for spanking. Or, parents who use excessive corporal punishment minimize their use to normal levels, and report continued aggression from the child despite their efforts to end it. Either case would artificially inflate the apparent link between spanking and aggression. Baumrind re-computed some of Gershoff’s findings, examining differences between the findings of longitudinal versus cross-sectional studies, separating studies she believed were really about abuse from studies of what she believed were really about non-abusive parenting, and studies that use parent and teacher reports versus parent reports only. She found all of Gershoff’s statistical results shrank.
    • Gershoff responded to this by noting that parental awareness of the legal obligations of psychologists to report child abuse, as well as simple feelings of shame, together create a kind of “Don’t Ask: Don’t Tell” policy in parenting research. Thus, clearly distinguishing between abusive and non-abusive parents will always be difficult.
  • It depends on how the conclusions are stated
    • Baumrind also points out these findings are really correlational. While it may be that corporal punishment leads to increased aggression, it is also possible that increased child aggression leads to increased corporal punishment (as some studies have shown) or that some third factor causes both parental use of corporal punishment and childhood aggression.
    • Gershoff responds by pointing out this is a double-edged sword. Suppose that results had shown that spanking children was associated with a decrease in aggression. While it may be that corporal punishment leads to decreased aggression, it is also technically possible that decreased aggression in children leads parents to spank (perhaps parents spank because the children are not aggressive enough?) or that some third factor stimulates aggression in parents while repressing it in children. No one would seriously entertain such hypothesis. Gershoff offers that arguing that parental actions actually may not impact children’s actions, especially when parents act with the desire to change children’s actions, seems equally silly.
    • While Gershoff does not offer this point, it seems reasonable to me to offer that correlational research also does not tell us what to do. While it may be equally likely that parental spanking causes aggression as that aggression causes parental spanking, children have limited ability to change their environments and behaviors compared to adults. Any intervention that reduces aggression is most likely to be targeted at adults, regardless of the direction of the relationship. Further, even if childhood aggression does cause parental spanking, this does not mean that parental spanking does not cause other problems, or that parental spanking does not make childhood aggression worse.
    • Holden goes on to point out that, regardless of the findings above, corporal punishment with children likely does not work because, speaking from a learning standpoint, four criteria would need to be met in order for it to be effective.  It is unlikely that parents meet these four criteria, and thus unlikely that spanking works as a physical punishment that changes behaviors in the desired direction:
      • It should be immediately after the infraction
      • It should occur after every infraction
      • The punishment of the first offense should be more severe than the punishment for later offenses
      • There should be no warning of the impending punishment (like yelling or warnings of punishment if behavior does not change)

Closing Thoughts on Spanking

Baumrind returns at the end of her response to the “blanket injunction” against spanking, offering the analogy of a horse moving. At a walk, the horse’s feet strike the ground slowly and independently, making for a four-beat rhythm. At a trot, the horse moves in a different way, making a two-beat rhythm. At a gallop, the horse moves in yet a third way, making for a three-beat rhythm. While these three kinds of movements can be seen as the horse’s speed increases, the three different movements can not be reduced to speed alone without losing something. Similarly, while use of no corporal punishment, use of appropriate of corporal punishment, and use of abusive punishment might be three different points on a continuum, that does not mean all three are reducible to one variable, such as frequency for example. As a result, a blanket injunction for all parents against corporal punishment, based on the fact that some parents physically abuse, is unreasonable. It is like suggesting that no one should drink alcohol, based on the fact that some people drink excessively.

Gershoff, at the end of her reply, returns to this point as well. She notes that the possibility of a link between corporal punishment and abuse means that corporal punishment, as a parenting technique, must be held to a higher level. While it may not be harmful when used in a non-abusive manner by 94% of parents, as Baumrind believes, that does not mean that it is helpful. Given that violence against adults, animals, and criminals to change their behavior is illegal, Gershoff questions why violence against children to change their behavior should be socially sanctioned, especially without clear and compelling evidence that it results in desirable outcomes.


This review is based on five articles published in 2002:

  1. John Bessa says:

    I think the three stages of spanking is important, with the least stage being benign. I am against it but I know that kids (and adults) play at spanking, so in many cases I don’t think it is really corporal punishment as there is no actual pain–just the message.

  2. [...] with increased risk of victimization from abusive relationships in adulthood” (Niolon, 2010, http://www.psychpage.com/family/disc.html).  In fact, many of the women that I know who were spanked “lovingly” became victims of [...]

  3. [...] child’s development and impaired child development was not discovered, Gershoff was quick to point out that the reliability of many of these studies is questionable for a number of reasons, including [...]