A large amount of controversy has surrounded the use of Straus and Gelles' (1990) Conflict Tactics Scale to measure domestic violence, most of which suggests that this scale is likely to underestimate the true prevalence of domestic violence (Johnson, 1995). If these criticisms are valid, the prevalence of domestic violence among study non-respondents will be higher than among respondents. The Expectation-Maximization (EM) algorithm makes use of distributional assumptions about variables in the data to maximize the likelihood of parameters for missing data given the observed data (Schafer, 1997; Allison, 2002). This allows the researcher to exploit relationships between variables in the data to estimate maximum likelihood statistics for unobserved cases. This research makes use of the E.M. algorithm applied to data from the Project on Human Development in Chicago Neighborhoods to obtain less biased estimates of the prevalence of domestic violence and child abuse in Chicago. The estimates of prevalence of the use of any parental violence against children increased from 54% to 76% (p<0.05) when E.M. was used to estimate the means. Estimates of partner violence also increased significantly for minor violence and for severe violence by the partner of the respondent. The findings provide some support for the contention that the prevalence of domestic violence and child abuse is higher among non-respondents. The results suggest that social work research on the prevalence of child abuse and similarly sensitive topics must make use of newer techniques for handling missing data, such as E.M. or Markov Chain Monte Carlo. While adequate freeware programs exist, the use of these techniques is illustrated using programs written by the author in the R statistical package.

Allison, P. Missing Data, Sage Publications, 2002.

Johnson, M. “Patriarchal Terrorism and Common Couple Violence: Two Forms of Violence Against Women” Journal of Marriage and the Family 57:283-294, 1995.

Schafer, J. Analysis of Incomplete Multivariate Data, Chapman and Hall, 1997.

Straus, M. & Gelles, R. Physical Violence in American Families, Risk Factors and Adaptations to Violence in 8,145 Families Transaction Publishers, 1990.