Stocks And Gelles Family Violence

A Comparison of Stocks' and Gelles' Family Violence Surveys

Straus and Gelles (1986) compared results of 1975 and 1985 surveys that attempted to measure intrafamily violence. They reported declines in the rates of child and wife abuse during this period. They attributed their findings to "a combination of changed attitudes and norms along with changes in overt behavior" with respect to intrafamily violence toward wives and children.

Stocks (1988) reanalyzed the data and found the results questionable. He found what he called "serious" problems of method (selection and statistical conclusion threats to validity) and that interpretation of the survey data allows for a different conclusion. He further found mathematical errors in the presentation of the results that lessen the statistical significance of the findings.

A major difference between the 1975 and 1985 surveys was the method of sample selection. The 1975 survey was conducted in person (65% completion rate), while the 1985 survey was conducted via telephone (85% completion rate). Gelles and Straus proposed that the higher completion rate of the 1985 survey meant it was the more representative of the two.

Stocks attempts to argue that there are significant demographic differences between households with phones and those without. According to him, the 1985 survey contained an over-representation of households with phones. I find this to be specious nonsense. The percentage of homes in the United States without phones is so insignificant that it can be ignored. Even if it is true, as Stocks states, that households without telephones are more likely to belong to low-income, minority, and younger groups, he presents no evidence that these groups have a higher incidence of intrafamily violence.

Finally, Stocks recomputed statistics from the reported rates for