Recent Controversies and News Coverage

Regarding Spanking

A. Chronology of events

1. Feb 1996—Conference on "The Short- and Long-Term Consequences of Corporal Punishment" sponsored by the American Academy of Pediatrics and others
   
   Involved presentations and responses by most of the big names in this arena—Bauman, Baumrind, Cohen, Eron, Graziano, Howard, Hyman, Larzelere, McCord, Polite, Straus.
   
   There was some coverage of the presentations—short articles, columns, etc. on both sides.
   
   The widest coverage was of Baumrind's response in which she said, "A blanket injunction against disciplinary use of spanking is not warranted by the data." This was cited in an AP or UPI news release and was widely quoted. While the title of the presentation was strong, the content of the presentation was considerably less strong—the core of her argument being that "to be maximally effective when a child disobeys, reasoning must be backed up periodically by consequences." She then goes on to equate consequences and spanking with no evidence for spanking being the only or most effective form of consequences.
   
   
   
2. Oct 1996—The proceedings of the conference was published as the October 1996 issue of Pediatrics (Vol 98, No. 4). Very useful collection.
   
   Again there was a small flurry of media reportage—of some of the anti-spanking conclusions, but again mostly of Baumrind's statement (see above) which was used as the title of her short article.
   
   
   
3. Aug 1997—Straus' long-awaited longitudinal study which he had presented as a paper at the 1994 NCFR conference and which he and Larzelere had referred to at the Pediatrics conference and in their articles from that conference. Published in Archives of Pediatric and Adolescent Medicine (151: 761-767, August 1997).
   
   
   
4. Au/Se'97—Very widespread media coverage, most of which took Straus' conclusions (see below) at face value. Straus was flooded by calls for a month or so after the AP news release from the media from all over the world—to the extent that he eventually quit taking calls. Both ABC and CBS had "major" coverage on their evening and night broadcasts. I received 12 or 13 queries from media sources for interpretation—appeared on both Ch. 4 and Ch. 5 simultaneously, had an hour call-in program on MPR, radio in Fargo, and others.
   
   Time (8/25/97, p. 65) had a full-page article mostly criticizing Straus' research (drawing mainly on Den Trumbull, a Montgomery AL pediatrician who is a perennial pro-spanking advocate). The article was a diatribe, not a report. Straus wrote a response letter to the editor (Time, ???? 1997).
   
   Larzelere put a critical review on the FAMSCI network (with very little reaction) in early September. Larzelere, Baumrind, and Polite; Ambati, Ambati, and Rao; and Miller wrote critical letters to the editor of Archives ??? 1997); and Straus wrote a response letter (Archives, ??? 1997.)
   
   

B. Straus' major conclusions and contributions

When parents use corporal punishment to reduce aggression and other anti-social behavior, the long-term effect tends to be the opposite. That is, the more mothers spanked at the start of the study period, the higher the level of anti-social behavior two years later.

This conclusion is not new. Straus' previous research and that of a substantial number of other studies have found this connection. What this study contributes is a methodology that allows for much better assurance that the relationship is real and causal. This was a two-year longitudinal study involving a large sample (910 children) and statistically controlling most of the confounding variables—level of anti-social behavior at the start of the study, family socioeconomic status, sex of the child, and the extent to which the home provided emotional support and cognitive stimulation.

C. The major critical issues raised

1. Age of the mothers
   Trumbull and Larzelere incorrectly assert that the sample consisted predominantly of young single mothers. What they missed is that these women were 14-21 when recruited into the database in 1979. Most did not have children at the time. The data for this study were gathered from these women in 1988 when they were aged 23-30. Most were married at that time.
   
   
   
2. Age of the children
   Trumbull and Larzelere observe that "limiting the study to 6-9 year olds skews the results. Spanking should not be used with kids beyond age six—who are old enough to understand the consequences of their actions. For them frequent physical punishment is likely to be humiliating and traumatic—and might well lead to worse behavior down the line."
   
   Straus' response:
   
   1. The study was not limited to 6-9 year olds. He also has data (with the same results) for 3-5 year olds and 10-13 year olds. He indicated this in the article. But, there was space in the article for treating only one age group. He selected 6-9s for two reasons: The type of ASB that a 5 or 6 year old might be engaged in (at Time 2) would be considerably less problematic and observable than that in which an 8-11 year old might be engaged. Second, the measure of ASB changes when children enter school.
      
      
   2. Also disagrees that corporal punishment of toddlers or preschoolers is less humiliating, traumatic, or otherwise harmful.
      
      
3. Trumbull, Larzelere, and other pro-spanking advocates imply that the reduction in use of physical punishment has actually contributed to increases in child abuse. The basis of this "argument" appears to be that by giving up spanking, parents have totally abdicated disciplining. Then, when the child one day goes too far, the parents lose it and resort to severe abusive physical punishment.
   
   First, they phrase this so that it leaves an implication. But, they offer no documentation. Second, if there has been a decrease in spanking, it has been very recent (the past two to four years). This is hardly enough time to have contributed to measurable societal changes in child abuse. Third, Straus' and other data indicate an actual continued reduction in child abuse coupled with a continued increase in reported cases of child abuse. The latter probably results from the same belief— that hitting children is not an acceptable behavior—which led to a decreased belief in not spanking.
   
   
   
4. Ambati et al charge bias, decrying Straus' failure to control what they consider to be some "key variables"—e.g., such as paternal use of CP and violence between the parents. This is a limitation but not a bias. This data is not part of the database Straus used (National Longitudinal Survey of Youth). Many of the variables were controlled in other Straus' studies and CP was still found to be associated with aggression and other behavior problems.
   
   
   
5. No-spanking group probably was really a low-spanking group.
   This criticism by Larzelere has merit. The data as coded did not include a never-spank category, so the "never spank—last week" category could really be a low-incidence of spanking category and demonstrate that infrequent spanking is beneficial. However, this is not compatible with results of other Straus' studies including the Goodhue Study—which found anti-social behavior lowest among the never-spanked, higher for those who had been spanked but not during the past six months, and increased with each increasing frequency of CP in the previous six months.
   
   
   
6. Small effect size
   Given the multiple determinates of child outcomes, a small effect size is to be expected in almost any social science research. But it is an error to automatically dismiss a small effect size. It is a well-established principle in epidemiology that a widely prevalent risk factor with small effect size (such as spanking) can have a much greater impact on public health than a risk factor with large effect size but low prevalence.
   
   Psychologist Patricia Cohen makes an interesting observation about this matter of effect size: ". . . small associations translated into society-wide effects can account for large differences between societies or large changes over historic time. One need only note that the association between a player's batting average and the outcome of a given at-bat is only marginally greater than zero. Nevertheless, we are happy to pay millions of dollars to hire a player whose average is afew percentage poinrts higher. In a society in which punishment of children is so prevalent, it is reasonable to contemplate the possibility of non-trivial improvemnet of the quality of life of children and adults were this practice to be abandoned. Both research in learning theory and the experiences of millions of parents have shown that discipline without corporal punishment can be authoritative, constructive, and successful." (Cohen, Pediatrics, 1996, p. 825)
   
   [For further elucidation of this matter of effect size, see Abelson, R. P. A variance explanation paradox: when a little is a lot. Psychological Bulletin 97:129-133, 1985]
   
   
   
7. Is not an experiment
   Trumbull, Miller, many other pediatricians and some other researchers will accept only controlled experiments as acceptable research, asserting that only randomized, controlled experiments can prove causation. Actually, even such experiments cannot "prove" causation. An experiment permits the greatest confidence in causal inference, but it is still an inference not certainty—artifacts/unknown (thus uncontrolled) variables can also produce erroneous findings in experiments.
   
   Straus merely said his research "allows a stronger causal inference than previous research".
   
   Further many experiments are so artificial that they do not readily translate to the "real world" and raise concerns about ethics (including some of those lauded by Larzelere, 1996).
   
   
   
8. The critics charge that Straus is biased.
   (See Section D)
   
   

D. Conclusion: Biases, paradigms, and "theories"

R. Pitzer
U of MN Ext
11/97

Produced by Communication and Educational Technology Services, University of Minnesota Extension.

In accordance with the Americans with Disabilities Act, this material is available in alternative formats upon request. Please contact your University of Minnesota Extension office or the Extension Store at (800) 876-8636.

University of Minnesota Extension is committed to the policy that all persons shall have equal access to its programs, facilities, and employment without regard to race, color, creed, religion, national origin, sex, age, marital status, disability, public assistance status, veteran status, or sexual orientation.