Last week, I received an email from Dr. Robert Larzelere – a prominent researcher in the field of parenting who, despite the overwhelming research on the negative long-term outcomes, wants the book left open on “conditional corporal punishment” (or corporal punishment that may be “nonharmful . . . under some conditions.”). This email was addressed to all members of the Society for Research in Human Development, urging us to consider both sides of the research before endorsing a ban on corporal punishment. His email is pasted at the end of this blog entry, and I urge you to take a look – though lengthy and technical in parts, it’s critical for us to understand that the spanking debate is alive and well. Dr. Larzelere is a renowned present day researcher in the field of parenting with enormous reach and impact. He influences other scholars, hundreds of young minds through teaching, and families who look to experts for advice. The fact that he is a voracious advocate of conditional corporal punishment in the face of research that opposes the practice is of great concern.
I’d like to first point out where I agree with Dr. Larzelere. Then, I’d like to take a look at fallacies in his scientific arguments. Last, I’d like to revisit the overwhelmingly consistent research on the negative impact of corporal punishment and offer some common sense thoughts as well.
I agree with Dr. Larzelere that it would not be wise to endorse a ban against parental spanking. Education is at the heart of this debate – just as we would benefit from education about gun rights, abortion, texting while driving, alcohol consumption, drug use (the list goes on), our society would benefit from a large scale, national education effort on the topic of corporal punishment. Control by an authority does have an immediate or temporary impact – be it a parent spanking her child or a nation policing our discipline – but that is all it has, as the underlying problems are still present. We would be wiser, as parents and as a country, to focus on long-term goals and outcomes for families. Support through education is the most effective way to create foundational change for the next generation.
Let’s take a look at Dr. Larzelere’s argument for conditional corporal punishment. He contends that conditional corporal punishment is statistically no worse than other methods of non-physical discipline and therefore innocuous. Let’s break it down.
Dr. Larzelere is correct (and the science verifies) that abusive corporal punishment is bad for children, whatever the measured outcome. He rightly calls for isolating the effects of conditional corporal punishment by removing abusive cases from the group. He then compares it to other non-physical discipline methods. Now, this is important: just what are these “other non-physical discipline methods,” and are they an appropriate comparison group? Non-physical discipline is defined by Dr. Larzelere as “grounding, privilege removal, and sending children to their room, and . . . psychotherapy”. Not surprisingly, his results show that “all corrective actions by parents and psychologists appear to increase children’s antisocial behavior” – in other words, conditional spanking is no worse than non-physical (and might I point out, PUNITIVE) discipline. Therefore, Dr. Larzelere claims conditional spanking to be nonharmful, even . . . beneficial.
Though the fallacy is obvious to me, I’m assuming it’s at least a little fuzzy to those who sing the beneficial praises of spanking. To compare spanking to ANY other punitive discipline method is to compare oranges to oranges. It’s not surprising that spanking, grounding, privilege removal, and time-outs are all correlated positively with antisocial behavior. Further, I’m not surprised that psychotherapy and Ritalin (see Larzelere email) aren’t helpful when parents continue to use punitive and controlling tactics with their children alongside therapy and/or drugs. On this, Dr. Larzelere and Dr. Gershoff agree! Where Dr. Larzelere makes the leap away from research is in his interpretation of the data. And this dissonance, I believe, can only stem from personal history, strong rooted belief systems, and a blatant bias.
The research is clear. A vast body of replicated literature links spanking to aggression, anti-social behavior, mental health problems, lower moral internalization, lower empathy, higher defiance, and the lists goes on. Temporary compliance is the only beneficial outcome of even conditional spanking. What is at stake, as Dr. Larzelere says, is the future well-being of children throughout the world. When we find over and over again in research that control through punishment hinders moral development and breeds self-interest, how can it be said that conditional spanking even has a chance at setting our children up to be well-adjusted adults in the long run?
Putting the research aside, there are ethical issues to consider regarding research on corporal punishment. I agree with Dr. Holden’s comment to Dr. Larzelere – “. . . the ethical issues apply regardless of the scientific debate.” So let’s design a prospective study with randomly assigned groups of first time pregnant parents and settle this debate! Because controlled corporal punishment need be compared to discipline methods that are mutually exclusive from spanking on a theoretical level, one group of parents will be trained to conditionally spank, and the other group will be trained in non-punitive discipline methods. (Notice the comparison of oranges to apples.) Here, ethics may override science: Would this proposed study ever make it past the Internal Review Board ethics committee? And even if it did, would participants be OK with random assignment into the spanking versus non-punitive groups? Might that say something about the practice of spanking on an ethical level?
Here are a few other common sense thoughts to consider:
- When children have a problem & we explain that we are going to hit them (and why), is this really the best way to teach them about solving problems?
- When a caregiver is simultaneously the person a child most loves and most fears, might this be confusing at all? Is love conditional?
- For parents who have spanked, are you confident? Did it feel right? How much anger or aggression seeped through? Did you wish an alternative strategy had worked rather than going for your “last resort”?
- Do you want for power and authority to override the lesson?
- Is punitive discipline sustainable? What will you do when your teenager laughs at your threats and “does it anyways”? Is your influence diminished in the long run?
Though I believe the research is sound, sometimes ethics, common sense, and human instinct speak louder than science.
I wonder why Dr. Larzelere holds on to corporal punishment with a death clutch. Is he that afraid to explore non-punitive ideas – discipline that lets go of control, treats children with compassion and understanding, and embraces internal motivation, morality, and perspective taking as long-term ideals? I do not support a ban on spanking, so on that we agree. But to argue so vehemently for the nonharmful (nay, beneficial) effects of conditional spanking is . . . perplexing. Please see Dr. Larzelere’s email below, and please post your thoughts on the topic!
From an open letter to the Society for Research in Human Development by Dr. Robert Larzelere on February 18th, 2012:
“Members of the Society for Research in Human Development (SRHD) were recently asked to endorse a resolution supporting bans against parental spanking. The supporting materials included with the proposal represented only one side of the scientific debate on this important topic. Members are asked to consider the other side of the debate.
A recent APA president compared the anti-spanking and conditional-spanking perspectives, featuring Gershoff’s (2002) meta-analysis and my own literature review (Larzelere, 2000). They called these reviews “two excellent and most current reviews” (p. 208), and said “no conclusions regarding causality can be drawn” (p. 199) and that a research “comparison of spanking with alternative procedures” was a “top priority” (Benjet & Kazdin, 2003, p. 215).
Brett Kuhn and I responded to their top priority by doing a meta-analysis of all the studies in Gershoff’s or my reviews that could be used to directly compare the child outcomes of corporal punishment with alternatives that parents could use instead (Larzelere & Kuhn, 2005). Our meta-analysis revealed that:
- Conditional spanking was linked with less adverse (or more beneficial) child outcomes than 10 of 13 alternatives it had been compared with directly. Conditional spanking is defined as nonabusive spanking when 2- to 6-year-olds defiantly refuse to cooperate with milder disciplinary tactics, such as reasoning or time out. Psychologists trained parents to use it to enforce time out for 2- to 6-year-olds from the late 1960s to the mid 1990s in the best-supported treatments for disruptive-behavior diagnoses (e.g., ODD, CD). No other back-up for time-out has ever been shown to be more effective for these kinds of children. Even the most defiant children quickly learn to cooperate with milder disciplinary tactics so that the spank back-up can be phased out.
- Corporal punishment was linked with more adverse child outcomes than alternative tactics only when it was severe or used as the primary discipline tactic.
- Customary corporal punishment (e.g., frequency) had equivalent child outcomes as all alternative tactics except for predicting less substance abuse than non-contact punishment in one retrospective study.
My colleagues and I later replicated the strongest causal evidence against customary spanking, namely that spanking predicts slightly greater antisocial behavior two-to-four years later after controlling for pre-existing differences in antisocial behavior (Gershoff, Lansford, Sexton, Davis-Kean, & Sameroff, 2012, in press; Straus, Sugarman, & Giles-Sims, 1997). Our two studies, however, were the first to compare these results with alternative disciplinary tactics that parents could use instead (Larzelere, Cox, & Smith, 2010; Larzelere, Ferrer, Kuhn, & Danelia, 2010) www.biomedcentral.com/1471-2431/10/10. Unfortunately, both of our studies showed similar apparently adverse effects of nonphysical punishment. Even psychotherapy and Ritalin looked as harmful as spanking when analyzed in the same way. Additional analyses in our two studies indicated that these “effects” are all statistical artifacts due to selection biases from child effects that have been reduced but not eliminated by these statistical controls. Any professional can replicate those findings from the NLSY data.
A more complete summary of the scientific evidence for the conditional-spanking viewpoint about spanking bans is available in Larzelere and Baumrind (2010) at http://www.law.duke.edu/journals/lcp/lcptoc73spring2010.
When I asked George Holden, the author of the proposed resolution, to consider a constructive dialogue about the scientific evidence, his reply was that the ethical issues apply regardless of the scientific debate. I do not doubt the good intentions of anyone supporting spanking bans. But without reasonably conclusive scientific evidence, my question is this: When is it ethical for one group to impose a set of absolutist values on others by law rather than by persuasion?
We need the best science to reduce child abuse without undermining appropriate parental authority. After 33 years, there is still no unbiased evaluation of spanking bans, and the available evidence suggests that appropriate authority has been reduced more than child abuse (e.g., Larzelere & Johnson, 1999; Patterson & Fisher, 2002, p. 74).
I am raising these issues not just for the welfare of children and parents, but also for the identity of the SRHD as a scientific society. At SRCD’s recent Themed Conference on Developmental Methodology, Margaret Burchinal’s plenary address commended development researchers for moving beyond what econometricians call “naïve” analyses to produce stronger causal evidence, so that developmental research on early child care could influence policy rather than be discounted by econometricians’ criticisms about valid causal inferences. Econometricians use the term naïve to refer to superficial statistics that do nothing to take selection bias into account (Angrist & Pischke, 2009, p. 72). By that definition, Gershoff’s (2002) effect sizes against corporal punishment are all naïve effect sizes, because they are merely a mathematical transformation of naïve correlations (mostly cross-sectional, longitudinal at best), without trying to do anything to take into account the selection bias due to child effects. I assume she did that for consistency, even for the few studies with stronger causal evidence. Because our field relies too much on naïve associations, we keep opposing most actions parents use to correct problems in their children, including helping them with homework (Hill & Tyson, 2009) and warning adolescents about the dangers of smoking and unprotected sex (de Leeuw, Scholte, Sargent, Vermulst, & Engels, 2010; Deptula, Henry, & Schoeny, 2010) as well as all disciplinary corrections.
Of course, everyone knows that corporal punishment is used too frequently and too severely by too many parents and that positive parenting needs to be encouraged as much as possible, consistent with the long-term beneficial effects of authoritative parenting (Baumrind, Larzelere, & Owens, 2010). Without improvements in causal evidence, however, developmental scholars will have difficulty helping spanking bans be effective for children and their parents, because our methods make nonphysical punishments (e.g., time out: Gershoff et al., 2010), psychotherapy, and Ritalin look as harmful as customary spanking.
Therefore, I consider it premature for SRHD to approve George Holden’s resolution advocating spanking bans. Instead, I call for a forum to compare and contrast the competing scientific evidence to clarify what both perspectives agree on, what they differ on, and to collaborate on research to make the types of discriminations needed for science, for parents, and for the future well-being of children throughout the world.”