Analysis of Straus 1997

Spanking by Parents and Subsequent Antisocial Behavior of Children

by Murray A. Straus, Ph.D., David B. Sugarman, Ph.D., and Jean Giles-Sims, Ph.D.

Archives of Pediatric and Adolescent Medicine Vol. 151, August 1997, pp 761-767


The design of the study:

  • The study fails to report on the children under 6 years of age where spanking is more necessary and reasoning is less effective.  Instead, the authors draw their conclusions from analysis of the 6 to 9 year age group, despite the availability of data on children ages 3 years to adolescence.  Spanking after 6 years should be infrequently used and needed.
  • This data for this study comes from interviews of mothers, not actual observation of their children.  The interviews were conducted 2 years apart. Mothers were asked to report how often they had spanked their children in the past week.
  • This study is not a comparison of nonspanked children with those occasionally spanked, but actually a comparison of children who are infrequently spanked with those who are frequently spanked (some as often as 15 times a week and as late in life as 9 years of age).
    • Most of the statistically significant increase in antisocial behavior occurred with those spanked 3 or more times in the previous week(more than 150 times per year).  This is important when considering how often these children may actually have been spanked.
    • The frequency reported probably underestimated the actual spanking frequency for at least two reasons.  (1) Fathers were not included in the interview.  (2) Most parents are not pleased to report spanking their child, esp. with the possibility of being accused of abuse.  Therefore, mothers reporting 2 to 3 spankings in the past week may actually have been spanking daily, not to mention the any additional spanks by the father which were not reported.  Frequent spanking at these ages is a marker for dysfunctional parenting and would correlate with a poor behavioral outcome.
    • The frequency measures are further confounded by the fact that the mothers reporting 0 spanks in the past week were not necessarily nonspankers, only for that previous week. Therefore, many of the mothers considered by the authors to be nonspankers (those in the 0 spanking group) probably were spanking at least occasionally.  The authors concede, “many of these children in the 0 spanking group [were] likely to have experienced frequent corporal punishment, even though not in the week prior to the study.”  This is true since surveys have consistently shown that 70 to 90% of parents use some spanking with their children.
  • The mothers interviewed were relatively young, 21 to 28 years, having an average age at the birth of the child in the study of 19 years.
  • This study fails to differentiate between appropriate versus inappropriate physical punishment. The best research seems to indicate that the context of punishment primarily defines the outcome.  How and when were the children spanked?
  • Even ignoring the previous criticisms, the observed increase in antisocial behaviors attributed to spanking was only 1.3% of the variance (observed difference in behavior).  Most of this change occurred with children reportedly spanked 3 times or more in the previous week.
  • In his opening remarks, Straus defends his case against spanking by citing “a large body of literature”, yet goes on to legitimately criticize these studies for being methodologically flawed, i.e. reliance on correlational data, failing to control for aggression at time-1, confounding spanking with other disciplinary practices, nonspecific age selection, and poor control for ethnic/cultural effects and parental warmth.  These criticisms legitimize his study, but his data have significant deficiencies as well.

Concerning other research reported in the same Archives journal, but not released to the press:

  • In a companion article in the Archives journal, researchers Dr. Marjorie Linder Gunnoe and Carrie Mariner found, ”For most children, claims that spanking teaches aggression seem unfounded.”1  They analyzed data on 1112 children aged 4 to 11 years.  They found that spanking deterred fighting (reduced antisocial behavior) for children ages 4 to 7 years and all black children. In only one subgroup of children (8 to 11 year old white boys in single-mother families) did spanking seem to foster aggression (an age group where frequent spanking is obviously inappropriate).  These findings contradict Straus’ conclusions, and are consistent with the strongest research in this area.
  • An editorial published in this issue, written by prominent psychologists and pediatricians in the field2, is remarkably neutral on the purported link between spanking and antisocial behaviors, like aggression.

Other research on corporal punishment of children by parents:

  • A thorough review of the literature on nonabusive physical punishment of children by Larzelere [Pediatrics Oct. 1996] found stronger evidence for beneficial versus detrimental effects of spanking with children ages 2 to 6 years.  The highest quality studies to date indicate positive short-term effects of its use by parents as a back-up for time-out and reasoning.
  • Some excellent long-term studies, conducted on-site (clinical, not survey-based), have failed to show that nonabusive spanking leads to aggressive behavior in children (Baumrind, Chamberlain, Eron, Simon).  The most significant determinant of developmental outcome is the total pattern of parenting, i.e. communication, consistency, firm control, parental warmth, etc.  A 10-year study (Baumrind) indicates that when judiciously used in the context of love with young children, spanking is associated with optimal developmental outcome.

Concluding thoughts:

  • Disciplinary spanking is a needed disciplinary measure with most children between the ages of 18 months to 6 years, when reasoning and talk alone are ineffective in changing problem behavior.  Milder measures, such as disapproval, time-out and privilege removal, should be used first with spanking serving as a back-up or enforcer of these measures.  Beyond 6 years of age, spanking should be infrequently needed or applied.  More importantly, it is the total pattern of parenting which determines the final developmental result.  The best research shows that young children need both abundant encouragement and consistent behavioral control.  Disciplinary spanking is necessary at times with some children to achieve this behavioral control.
  • Societies and countries where spanking has been banned have not experienced a reduction in violence.  In fact, the reverse effect has occurred in Sweden with a 4-fold increase in child abuse in the past decade.  There is no convincing evidence that ordinary spanking is a root cause of violence in society.
  • Finally, Straus’ findings, weak as they are, do not apply to all children of all ages regardless of the frequency with which they were spanked.  His findings do not even touch the common setting of occasional spanking by parents of children under 6 years (when reasoning is less effective) in appropriate situations.


1. Gunnoe ML, Mariner CL. Toward a Developmental- Contextual Model of the Effects of Parental Spanking on Children’s Aggression. Archives of Pediatrics Adolesc. Med. 1997;151:768-775.

2.  Socolar RS, Amaya-Jackson L, Eron L, Howard B, Landsverk J, Evans J. Research on Discipline. Archives of Pediatrics Adolesc. Med. 1997;151:758-760.