A Comparison of Two Recent Reviews of Scientific Studies of Physical Punishment by Parents
Robert E. Larzelere
June 2002
Summary
Two recent reviews have summarized child outcomes associated with physical punishment by parents.1 2 They arrive at somewhat different conclusions, even though their underlying information is consistent with each other. After documenting this, I will show that child outcomes associated with ordinary physical punishment are also associated with alternative disciplinary tactics when similar research methods are used. Detrimental child outcomes are associated with the frequency of any disciplinary tactic, not just physical punishment. Therefore, it is the excessive misbehavior that is the actual cause of detrimental outcomes in children. Parents realize that excessive misbehavior will hinder their children’s success in life and want to minimize excessive misbehavior with the best disciplinary methods. They need better information about how to discipline their children in the most effective manner. Effective discipline is based on a foundation of a positive, loving parent-child relationship and uses proactive discipline skillfully. In responding to misbehavior, parents need to use milder disciplinary tactics skillfully. The most effective way to use spanking is to back up milder disciplinary tactics, such as reasoning and time out, with 2- to 6-year-old children. Research has shown that this is not only effective in itself, but the child then cooperates with the milder disciplinary tactics, making the spank back-up less necessary as the child gets older.
Contrasting Reviews
Larzelere (2000)2 and Gershoff (2002)1 arrived at somewhat different conclusions in their summaries of scientific studies of physical punishment by parents. Larzelere (2000) concluded, “spanking has consistently beneficial outcomes when it is nonabusive (e.g., two swats to the buttocks with an open hand) and used primarily to back up milder disciplinary tactics with 2- to 6-year-olds by loving parents. . .most detrimental outcomes in causally relevant studies are due to overly frequent use of physical punishment” (p. 215). In contrast, Gershoff (2002) concluded, “Ten of the 11 meta-analyses indicate parental corporal punishment is associated with . . . undesirable behaviors and experiences. . . . Corporal punishment was associated with only one desirable behavior, namely, increased immediate compliance” (p. 544).
Tables 1 and 2 show that their conclusions are based on similar evidence. These tables summarize the most important subset of studies that were considered in both reviews. Those studies focused on 6 outcome variables in children under the average age of 13, including the five child outcomes in Gershoff’s aggression composite (child aggression, delinquency, adult aggression, crime, and abusiveness) plus immediate (or short-term) compliance. Gershoff summarized the studies in terms of effect sizes (Table 1), whereas Larzelere (2000) summarized studies as finding predominantly beneficial or detrimental outcomes or neither or both (Table 2).
Both reviews found beneficial child outcomes associated with nonabusive physical punishment (bottom row of Tables 1 and 2). Both reviews found that the research methods that isolate the causal effect of physical punishment found beneficial child outcomes (Randomized clinical trials are in the right-hand column of the tables. This is the kind of study mandated for confirming the causal effects of drugs by the Federal Drug Administration before drugs can be put on the market.). The differing conclusions are based on different explanations for the overall pattern of beneficial vs. detrimental outcomes associated with physical punishment. The randomized clinical trials of nonabusive physical punishment (in the lower right-hand corner of Tables 1 and 2) differ from the other studies in all of the following ways: They provide more causally conclusive evidence, they limit the physical punishment to two swats of an open hand to the buttocks, their outcomes are limited to immediate compliance with everyday commands and with the time out procedure, spanking is used only when the child does not cooperate with time out, the children are between the ages of 2 and 6 and are so noncompliant that their parents sought help from a clinical psychologist to manage their child’s excessive misbehavior.
Gershoff (2002) considers the type of outcome the most important difference, concluding that physical punishment may increase immediate compliance but that it is associated with detrimental levels of 10 other outcomes. In contrast, Larzelere (2000) emphasizes the evidence of causal effects of nonabusive spanking when used to back up milder disciplinary tactics in 2- to 6-year-olds. Precisely because of the need to sort out these differing explanations, Larzelere (2000) featured a second table that summarized studies in the “partially controlled” column in Tables 1 and 2. These studies made some attempt to take into account the excessive misbehavior that is causing most parents to spank their children more (as well as to use more of all other disciplinary tactics). Gershoff (2002) makes no attempt to find evidence of causal effects from such studies, partly because the guidelines for her statistics cautioned against using evidence from such studies.3 But this advice led her to ignore the most relevant evidence for deciding among the alternative explanations of the overall pattern of findings.
Consider the study by Gunnoe and Mariner (1997),4 for example. Gershoff (2002) summarized their associations between spanking frequency and either aggression or antisocial behavior five years later, resulting in apparently detrimental “effect sizes.” Yet Gunnoe and Mariner themselves concluded, “For most children, claims that spanking teaches aggression seem unfounded” (p. 768). They based this conclusion on statistics that took into account the excessive misbehavior of the child in the first place. Under those circumstances, spanking caused reductions in fighting five years later in 3 subgroups: African-Americans, 4- to 7-year-olds, and girls. In contrast, it caused increases in fighting five years later in 2 subgroups: European-Americans and 8- to 11-year-olds. It also caused increases in antisocial behavior more consistently, but these analyses had the additional problem of being based on the same data source (the mother: Baumrind et al., 2002).5
Similar to Gunnoe and Mariner (1997), Larzelere’s (2002) review emphasized results from 11 such studies that did something to take initial excessive misbehavior into account, thus providing some evidence concerning the causal effects of physical punishment. Overall, those studies found more beneficial than detrimental child outcomes under the following circumstances: when overly severe physical punishment was removed; when spanking was used as a back-up for milder disciplinary tactics, such as reasoning or time out; when the children were younger than 6.5 years old; when the children were extremely defiant; and when the subcultural group viewed spanking as normative discipline (e.g., African-Americans, conservative Protestants). Detrimental outcomes were more likely for long-term outcomes (8 detrimental; 3 beneficial) than for outcomes during the next disciplinary incident or the next day. There was a fairly even balance of detrimental and beneficial outcomes across different types of outcomes (behavior problems, mental health) and by whether positive parenting was accounted for.
Finally, the strongest causal evidence for detrimental outcomes of spanking is based on methods that make alternative disciplinary tactics appear equally detrimental in most cases. Of the 11 studies that controlled partially for initially excessive misbehavior, only Straus et al. (1997)6 found uniformly detrimental outcomes. The other 10 studies either found beneficial outcomes (3 studies), a mixture of beneficial and detrimental outcomes (2 studies), neutral outcomes (1 study), or a mixture of detrimental and neutral outcomes (4 studies). But Larzelere and Smith (2000)7 replicated and extended Straus et al.’s (1997) study, using the same publicly available data set. In general, they found similar increases in antisocial behavior two years later for those who used four alternative disciplinary tactics frequently: grounding, removing privileges, docking allowances, or sending the child to his or her room. Further, these apparently detrimental outcomes for spanking and the four alternatives all disappeared after we did a better job of taking the initial level of excessive misbehavior into account.
Only a few studies of physical punishment have investigated recommended alternative tactics using the same methods. In most cases, such studies have found equally detrimental outcomes associated with the recommended alternatives. That is true in major research by Straus,8 myself,9 10 and most applicable studies of children under 13 in Gershoff’s (2002) review. Straus and Mouradian (1998) asked parents how frequently they had used disciplinary reasoning, time out, or privilege removal during the previous six months. They found that the frequency of those tactics was far more strongly associated with antisocial behavior and impulsive child behavior than was corporal punishment. To my knowledge, this is the only study in which Straus investigated recommended forms of nonphysical punishment in addition to spanking. Similarly, my most important original study found that the frequency of spanking 2- and 3-year-olds was associated with disruptive behavior 20 months later,9 about as strongly as the average prospective study of antisocial behavior in Gershoff (2002).1 But the frequencies of all of the following disciplinary tactics were even more strongly associated with higher disruptive behavior 20 months later: nonphysical punishment, reasoning (without punishment), and disciplinary responses that included neither reasoning nor punishment. Of the studies addressing Gershoff’s (2002) aggression composite, only three studies investigated alternative disciplinary tactics for children under the age of 13. None of those three studies found more beneficial associations for nonphysical punishment than for physical punishment.8 11 12
I do not think that these comparisons mean that physical punishment should be preferred over nonphysical punishment or milder disciplinary tactics, such as disciplinary reasoning. Rather, these results confirm my suspicion that the detrimental outcomes associated with nonabusive physical punishment are due to the excessive misbehavior that leads to more physical punishment. Excessive misbehavior is reflected even more strongly in more common types of disciplinary tactics, such as nonphysical punishment and disciplinary reasoning.
We cannot continue to lump all forms of physical punishment together to make an overall assessment that applies to all applications. Rather, we need to discriminate between more vs. less effective ways of using each disciplinary tactic. How parents use any disciplinary tactic is more important than what tactic they use – whether we are considering disciplinary reasoning, nonphysical consequences, or physical discipline. Especially with 2- to 3-year-olds, disciplinary reasoning works better if it is backed up when necessary, first by nonphysical punishment, and then, if warranted, by occasional nonabusive spanking or something equivalent.9 13 This is consistent with the strongest causal evidence about physical punishment: Nonabusive spanking reduces subsequent noncompliance and fighting in 2- to 6-year-olds when used by loving parents to back up milder disciplinary tactics, such as reasoning or time out. Nine studies support this conclusion, and no study contradicts it. To be sure, a brief room isolation has proven to be equally effective as a spank back-up in enforcing cooperation with time out in clinically defiant 2- to 6-year-olds.14 Further, nonphysical punishment works as effectively as physical punishment in backing up reasoning in 2- to 3-year-olds.10 Nonetheless, when two disciplinary tactics are equally effective on the average, they are each more effective for some children some of the time than the other alternative. That was demonstrated in Roberts and Powers (1990).14 Moreover, the most effective way to reduce subsequent fighting in my best study occurred when mothers combined reasoning, a nonphysical consequence, and a nonabusive spanking.9
The evidence to date supports a conditional sequence model of optimal disciplinary tactics.13 For 2- to 6-year-olds, parents should establish a solid foundation of a positive, loving parent-child relationship. They should emphasize proactive teaching – an ounce of prevention is worth a pound of cure. When misbehavior occurs, they need effective responses, beginning with verbal correction and reasoning. Disciplinary reasoning becomes more effective by itself when backed up periodically with nonphysical punishment. When a 2- to 6-year-old refuses to cooperate with nonphysical punishment, such as time out, it needs to be enforced with something like a two-swat spank to the buttocks. Yes, there are alternatives that work better for a few children all of the time and for all children some of the time. But when spanking is used in this way at these ages, the evidence to date indicates it is effective, especially in getting children to cooperate more with the milder disciplinary tactics. In this way, parents can reduce the need to use spanking at all as the child gets older. Parents need more disciplinary options, not fewer ones. They also need to know optimal strategies for all aspects of discipline. Hopefully future research can build on these two reviews to provide parents with that information.
people.biola.edu |
