Elizabeth Gershoff on the Physical Punishment of Children


Elizabeth Gershoff is a developmental psychologist and associate professor at the Department of Human Development and Family Science at the University of Texas at Austin. One important part of her research is about the effects of physical punishment on children.

Gershoff has a number of interesting publications on the matter, such as:

—> The article “More Harm Than Good: A Summary of Scientific Research on the Intended and Unintended Effects of Corporal Punishment on Children” published in the journal Law and Contemporary Problems (a publication of Duke Law School) back in 2010. Go here for the full article.

—> The article “Corporal Punishment by Parents and Associated Child Behaviors and Experiences: A Meta-Analytic and Theoretical Review” published in Psychological Bulletin in 2002. An online copy can be found here.

—> A report titled “Report on Physical Punishment in the United States: What Research Tells Us About Its Effects on Children” from 2008. Can be found here.

—> The article “The Case Against Corporal Punishment of Children: Converging Evidence from Social Science Research and International Human Rights Law and Implications for U. S. Public Policy” from 2007 in the journal Psychology, Public Policy, and Law. Go here for the full text.

In this overview, I will focus on Gershoff (2008).

What is physical punishment of children?

The Gershoff (2008) report defines physical punishment as “the use of physical force with the intention of causing the child to experience bodily pain or discomfort so as to correct or punish the child’s behavior” and argues that it includes such actions as slapping a child’s hand, hitting children with a paddle, washing a child’s mouth with soap and torture-like actions like forcing a child to sit in painful positions. The report also quickly points the difference between physical punishment and protective physical restraint, such as holding a child to prevent him or her from running into traffic. As far as I can tell, it is common for those who are in favor of physical punishment of children to equivocate these two actions in order to justify physical punishment.

A Long Series of Disturbing Findings

Gershoff (2008) lists a number of results from the research that delivers a decisive empirical blow against physical punishment of children. As far as prevalence goes 80% of American children have received physical punishment from their parents by the time they reach the 5th and over 70% of parents agree or strongly agree with the sentiment that “children sometimes need a good, hard spanking”

1. People who are more likely to use physical punishment on children have themselves been on the receiving end of physical punishment when they were children.

2. Physical punishment is detrimental to long-term compliance.

3. Physical punishment of children leads to less internalization of moral norms, more physical and verbal aggression, physical fighting and bullying, antisocial behavior and less ability for the child to feel empathy. In other words, the more physical punishment the child receives, the more disobedient the child becomes.The results cannot be explained by the fact that aggressive children receive more physical punishment from parents. Longitudinal studies show that more physical punishment the parents uses, the more aggressive behavior the child displays over time, even controlling for initial aggression. A randomized control trial supports this result.

4. The more physical punishment and the stronger it is, the more likely the child is to experience mental health issues as an adult, such as depression and anxiety as well as alcohol and drug use, even into adulthood. The mechanism is thought to be increased psychological distress.

5. All studies carried out on the issue shows that more physical punishment is associated with lower quality of the parent-child relationship.

6. The more a child is physically punished, the more likely he or she is to hit family members, spouse or dating partner.

7. Children who are subjected to physical punishment like spanking has a seven times higher risk of getting severely assaulted.

Why doesn’t physical punishment work?

The Gershoff (2008) report suggests a number of reasons, such as the fact that it does not teach children why their behavior was wrong, it does not teach children what appropriate behavior should have been used instead. it makes parents appear as social models, suggesting that violence is morally acceptable, children fear parents, and this damages the parent-child relationship etc.


Debunker of pseudoscience.

16 thoughts on “Elizabeth Gershoff on the Physical Punishment of Children

  • These people were “spanked” as children by adults, and they “turned out just fine”:

    Man Terrorized After Stopping A Dad From Spanking His Son

    Woman who prompted SEPTA bus shooting pleads guilty

    Terrifying footage shows bus gun attack ‘after man dared to complain about mother spanking her child’

    Parents acting out

  • In light of the Judge Adams video,

    We often hear from those who fight to uphold this practice for those under the age of 18 (even to the blaming of the social maladies of the day on a supposed “lack” of it), but we rarely, if ever, find advocates for the return of corporal punishment to the general adult community, inmate population, military, or college campuses. Why is that?

    Ask ten unyielding proponents of child/adolescent/teenage-only “spanking” about the “right” way to do it, and what would be abusive, indecent, or obscene, and you will get ten different answers.

    These proponents should consider making their own video-recording of the “right way” to do it.

  • Children and teenagers should have a right to their bodies, and the right to say “No!”

    Currently in the U.S.:

    When an adult does it to another adult, its sexual battery:

    When children do it to adults, its a “deviant sexual prank”:

    When an adult does it to a person under the age of 18, its “good discipline”.

    Research/recommended reading:

    Spanking Can Make Children More Aggressive Later

    Spanking Kids Increases Risk of Sexual Problems

    Use of Spanking for 3-Year-Old Children and Associated Intimate Partner Aggression or Violence

    Spanking Children Can Lower IQ

    Plain Talk About Spanking
    by Jordan Riak

    The Sexual Dangers of Spanking Children
    by Tom Johnson

    “Spanking” can be intentional or unintentional sexual abuse

  • This article is a bit misleading. The way the bulletin points are presented makes it seem as though a causal relationship has been discovered linking physical punishment to “x”. Gershoff acknowledges that she found no causal relationship.

    One thing to note is that a causal relationship can be found between corporal punishment and immediate compliance. She also writes, “Immediate compliance can be imperative when children are in danger.” Based on this we can conclude that in instances where the child is in danger, it can be imperative to use corporal punishment.

    Additionally, the information she uses to establish a correlation is shaky at best. A majority of the information on the use of physical punishment comes from parents’ or adolescent and adult children’s recollection of events. Hopefully I don’t have to explain why that’s not a credible way to get accurate information.

    Another problem is that in the studies she cites there is no clear definition on what constitutes corporal punishment other than punishment that lacks the risk of physical injury. This definition is never given to people they survey so there’s no way to know if they are even talking about the same things. Some parents may think it only applies to spanking, some may think it only applies to a smack on the hand, some may think it applies to both, etc.

    This article appears to be a bit dogmatic and anti-reason. There’s no mention of the studies that show corporal punishment as having positive effects (Larzelere, 1996; Baumrind, 1996a). You also fail to mention any of Gershoff’s criticisms.

    The longitudinal study you mention suffers from the same observer bias and failed methodology as the rest.

    I’m going to post a quote; it’s your job to guess who it’s from:
    “As a field and as a society, we must separate out the emotionally charged aspects of the debate over corporal punishment so that we can knowledgeably and responsibly recommend or discourage parents’ use of corporal punishment with their children.”

    Just to make it clear: I have a daughter. I don’t use corporal punishment unless she’s in immediate danger (smacking a hand away from a boiling pot). I do it for moral reasons though, not for pseudo-scientific reasons. You’re just as bad as the anti-vaccine crowd. You should be ashamed.

  • 1. No, the use of bullet points does not suggest a conclusively demonstrated cause and effect relationship, because the points are phrased in a way that avoids suggesting causation when only a correlation has been found. Furthermore, some of the studies done has controlled for confounding factors that has been suggested by the proponents of physical punishment of children. Furthermore, an independent convergence of evidence is strong evidence for causation.

    2. You make the same flawed argument as Gershoff debunks in the report: restraining a child from hurting him- or herself or others does not fall under the category of physical punishment. I even wrote this explicitly in the article. Clearly, you did not read it very carefully.

    3. Actually, a lot of the data is not based on subjective reporting, but on observable data points. Furthermore, the gold standard of science, a randomized controlled trial supports the general conclusion that physical punishment of children is not effective or healthy.

    4. The physical punishment of children has been clearly defined in the report. The definition used was: “the use of physical force with the intention of causing the child to experience bodily pain or discomfort so as to correct or punish the child’s behavior”. Again, this is something that is clearly written both in the report and this article and the definition is of course used in the relevant studies. Of course scientists are going to make sure the subjects of the study know that the study is about.

    5. You are appealing to false balance. Presenting an accurate view of the science is neither dogmatic nor anti-reason. You have fallen for the single study fallacy. It does not matter that one or two very old studies find something else. What matters is that the vast majority of studies, carried out by different people at different times and in different locations looking at slightly different things all converge to the same general conclusion: the physical punishment of children is ineffective and has both short-term and long-term negative consequences.

    6. You are quoting Gershoff, but it is out of context, so you are performing yet another fallacy. The context is the following:

    establish empirically connections between corporal punishment and potential child outcomes, particularly in longitudinal and prospective studies. It is by examining data and evaluating their conclusions that science progresses and society benefits. The present article is an attempt to analyze systematically the extant data and theory on parental corporal punishment to inform scientific, and ultimately popular, discussion. As a field and as a society, we must separate out the emotionally charged aspects of the debate over corporal punishment so that we can knowledgeably and responsibly recommend or discourage parents’ use of corporal punishment with their children.

    In other words, the “emotional charged aspects of the debate” primarily has to do with the proponents of physical punishment of children, because the results of studies of the physical punishment of children show clearly that it is not effective and that it has many negative effects, such as increased aggression and mental health issues, even as adults.

    You are also performing a straw man fallacy, as I have only reference scientific evidence, not written any “emotional charged” comments.

    7. Since we are on the topic of your own child-rearing practices, they are clearly substandard. The best way would be to explain to your daughter when she reaches the age at which she can reach the stove that it is very hot and that she can hurt herself if she touches it. That way, she will respect your knowledge, rather than fearing you because you punish her physically.

    8. Your last ditch case — a comparison between the opponents of physical punishment of children and anti-vaccine proponents — is flawed, because the evidence is on the side of vaccines and on the side of the opponents of physical punishment of children. Furthermore, it is your side that continues to put forward the same, already debunked, arguments and use the exact same fallacies as anti-vaccine proponents, such as straw man, not understanding the basic science and quoting out of context.

    Still, thanks for commenting, even thought your comment lacked intellectual rigor and overall substance.

  • 1. No where in your original piece did you mention that no causal relationship could be found. That’s why it’s misleading. Someone searching Google and coming across your page would almost certainly think the debate on spanking is over, yet your main source in this article admits that it’s not. In fact, you have an article titled, “There is No ‘Spanking Debate'”. Which isn’t true according to Gershoff. Search for debate in her 2002 paper to see all the instances where she talks about it.

    2. It’s not my argument, it’s from Gershoff’s 2002 paper.
    From page 3:
    “research on learning has confirmed that corporal punishment
    is indeed effective in securing short-term compliance
    (Newsom, Flavell, & Rincover, 1983).”

    She goes on to say on page 12:
    “Immediate compliance can be imperative when children are
    in danger”

    One definition she provides is on page 2: “Behaviors that do not result in significant physical injury (e.g., spank, slap) are considered corporal punishment”. If we put 1 and 1 together, what do we get? Slapping a child’s hand is indeed effective in short-term compliance and can be imperative when the child is in danger. You can’t say this slap is a form of restraint because she specifically use a slap as an example of corporal punishment.

    3. Once again you are arguing against your source. On page 2 she writes: “the
    majority of information on corporal punishment comes from parents’ or adolescent and adult children’s recollections of frequency of corporal punishment.” She brings it up again on page 27: “Because corporal punishment occurs relatively rarely, researchers must rely on parents’ self reports of corporal punishment rather than observations,”

    4. “None of the studies included in the meta-analyses presented above asked parents what they meant by corporal punishment but rather provided parents a definition with which to decide whether their behaviors fit.” She goes on to ask if x, y, or z are considered corporal punishment.

    5. “Psychologists and other professionals are divided on the question of whether the benefits of corporal punishment might outweigh any potential hazards;” “Despite this controversy and the hundreds of scientific studies invoked on either side of the debate,”

    6. The context doesn’t change anything.

    7. It’s imperative according to Gershoff.

    8. I suppose it was a faulty comparison. The anti-vaccine crowd has no evidence on their side. Your side does have evidence. You just dismiss all of the other side’s evidence. Which does exist according to Gershoff.

    • Sorry if I sound combative. After re-reading my posts I noticed they may come off that way. I want to thank you for the dialogue though.

    • 1. It is assumed that you understand that formulations such as “the more X, the more Y” is a correlation.

      2. Now you are confusing two different things. First you spoke of restraining (which does not fall under the definition of physical punishment), now you speak of short-term compliance to physical punishment. Yes, but short-term compliance can be attained by other, non-violent methods, and the short-and long-term consequences of physical punishment of children far, far outweighs any short-term beneficial effects. The difference is in the intentions and effects. Is it a slap to get the hand away from a hot pot on the stove, then that is a form of restraining. Are you using a slap to try and correct behavior afterwards, when there is no immediate danger, then that is a form of physical punishment. Notice that the definition given is not strictly about behavior only, but about intention.

      Do you really believe that short-term compliance (which can be obtained by other, non-violent methods) is worth the increase in risk of mental health issues? Aggression? Broken relationship between parents?

      3. It does not matter, since many different studies come to the same general conclusion. For your position to be valid, you would have to assume that every single parent has a net bias to remember more spanking than occurred, but the bias is actually the other way around. Memory falters over time and you remember less, not more.

      4. So you agree that a definition was provided? Good.

      5. More quotes out of context. The quotes come from the introduction of the 2002 article and therefore describes how the situation was prior to the impact of the 2002 review. However, this is 2012 and, as we saw, the evidence is quite convincing against the physical punishment of children and the situation does not represent the current view among psychologists.

      6. Yes, the context changes everything seeing how the article is a searing indictment of the physical punishment of children.

      7. No, on the contrary. I am sure Gershoff would prefer non-violent methods to violent methods any day of the week.

      8. “Your side” has no evidence. Again, you are portraying the situation as if there were two sides, but there is not. There is only one side, which is the side with the scientific evidence. There is no scientific controversy over this issue. Short-term compliance does not justify the increased risk of severe negative outcomes, especially since it is no better than non-violent methods.

    • Also check out this new study looking at physical punishment and mental disorders recently published in pediatrics. This study also notes that:

      However, there is evidence that supports the validity of accurate recall of adverse childhood events and that psychopathology is not linked to less reliable or less valid self-reported data on adverse childhood experiences.

      If physical punishment is not a contributing factor to the development of mental disorders, why do we see this association?

  • I fixed an ambiguous reference error in the original blog post.

  • I didn’t come here to take a stance on the morality, efficacy, or proposed legality of corporal punishment of children. I came here because a blog titled, “Debunking Denialism” is doing the very thing it’s purports to debunk. Those seeking to make a blanket statement about physical punishment, as well as the studies that try to show causation, ignore some key factors that need to be included in the debate. Examples would be, but aren’t limited to, the frequency, severity, locality, its use in conjunction with other methods, and other environmental variables. Even one of the studies 18andsafenow cited as “research/recommended reading” shows that there is no statistically significant difference between spanking once or twice a month and not spanking at all. That is to say, according to Catherine Taylor’s longitudinal study, no correlative evidence could be found between spanking once or twice per month and an increased risk in aggressive behavior. If our goal is academic integrity, we can’t ignore that fact.

    Taylor’s study can be found here:

    Before I exit the discussion I’ll leave some links to show that the debate does exist.


  • If you did not come to discuss the efficacy (I assume you mean effectiveness; efficacy is about how well a treatment does against placebo in a clinical trial), then why are you here?

    You also do not seem to understand the term “statistical significance”. It only tells you something about how likely the obtained evidence, or more extreme evidence, is on the null hypothesis. It tells you nothing about the practical relevance of the results. Also, statistical significance is a poor way to evaluate results. Instead, look at the effect size and confidence intervals.

    Taylor et. al. (2010) actually concludes the exact opposite of you:

    The current findings suggest that even minor forms of CP, such as spanking, increase risk for increased child aggressive behavior. Importantly, these findings cannot be attributed to possible confounding effects of a host of other maternal parenting risk factors

    Even minor physical punishment increases the risk of child aggression and these results cannot be explained by the confounders you claim. Try reading the study before putting it forward as evidence for your cause. If you look at the effect size in terms of OR and confidence intervals, it was 1.17 (0.94–1.44) for the spanked 1 or 2 times. When it comes to confidence intervals, them most plausible values is near the effect size, with plausible rapidly declining as you move towards the edges of the confidence interval. This finding should not be ignored by saying “not statistically significant”.

    Also, listing a bunch of articles is not evidence of a genuine scientific debate. You are deploying the denialist tactic of manufactroversy:

    Tactic: Manufactroversy.
    Description: Tries to make it look like there is a genuine scientific controversy where there is none.
    Countermeasure: Explain how strong the science is. Also see “Confusing Mechanism With Fact” and “False Balance”.

    For instance, there are many (flawed) studies on homeopathy that purports to show an effect. However, what needs to be looked at is the overall message of the literature, not the result of single studies.

    The first study you link complain about not eliminating confounders, but this is precisely what most newer studies have been doing, such as Taylor et. al. The second study relies on a single study not showing negative long-term effects of physical punishment, thereby performing the single study fallacy. The third study (really an opinion-piece) complains that a ban on spanking “shows a serious lack of cultural sensitivity by imposing one set of Euro-centric disciplinary values on underrepresented cultures and subcultures”, which has no bearing on science. The fourth “study” (an opinion piece in a law journal) lacks any kind of substance and does not address the actual research. The firth study is the exact same as the first, indicating that you (1) have not read the studies you link and (2) post duplicates to artificially boost the count. The sixth study make use of questionable subgroup analysis (“data dredging”). The seventh study is a 9 year old article calling for better elimination of confounders (which have been done, se the Taylor study). The eight “study” is just a comment on the 2002 metaanalysis, which has no relevance to the majority of evidence presented post-2002. The ninth study only looks at studies between 1990 and 2000, and says that “there was insufficient data about a number of the moderator variables to conduct meaningful analyses”, which makes you shoot yourself in the foot. Surely, a problem with confounders goes both ways. The tenth study is not a study, but a description of the problem of selection bias and does not address the research.

    You are right about one thing though. You are exiting this discussion.

  • Entering this debate rather late, but as it appears that all of these studies are correlation studies. Has anyone asked the question, “maybe it’s the aggressive behavior in children that cause their parents to use corporal punishment?”

    Obviously it would be impossible/unethical to do a randomized study to arbitrarily apply corporal punishment to establish a causal effect, but has anyone at least done a twins study to rule out genetic sources of aggression?

    To me, there is a possibility of the tale wagging the dog here. As anyone that works with children knows, there is a lot of variability in the degree of aggressiveness, dominance, and disobedience between different children. If a parent were inclined to use corporal punishment, wouldn’t they only apply it to the children that are the the most aggressive/disobedient?

    If that effect were real, then it would definitely show up in any correlation study, and it’s unclear to me how one would filter the effect out in a causal analysis of corporal punishment.


    • Researchers have replicated the general findings while controlling for a number of confounders, such as initial level of child aggression (thereby torpedoing your reverse causation scenario), demographics and a number parental risk confounders (child neglect, intimate partner violence, depression, alcohol and drug abuse etc.)

      I discussed one such study, Taylor et. al (2010), earlier in this comment section:

      Despite American Academy of Pediatrics recommendations to the contrary, most parents in the United States approve of and have used CP as a form of child discipline. The current findings suggest that even minor forms of CP, such as spanking, increase risk for increased child aggressive behavior. Importantly, these findings cannot be attributed to possible confounding effects of a host of other maternal parenting risk factors.

Comments are closed.

%d bloggers like this:

Hate email lists? Follow on Facebook and Twitter instead.