SOURCES OF SAMPLING BIAS IN SEX STUDIES

In: Public opinion quarterly: journal of the American Association for Public Opinion Research, Band 21, Heft 3, S. 388-394

Abstract

In their STATISTICAL PROBLEMS OF THE KINSEY RE- PORT, Cochran et al point out that the sampling procedures employed by Kinsey may have introduced biases into his % statistics. These procedures involve using R's who: (a) volunteer to be interviewed; & (b) are members of org's included as sampling units. Data collected by Burgess & Wallin permit tests of whether these aspects of sampling procedure do, in fact, bias enumerative findings concerning sexual behavior (SB). These data are obtained from 1,000 Coll educated, Ur Me, engaged couples willing to fill out questionnaires. An intensive follow-up study was made of 666 of the couples who had actually married & were still available to res. After filling out the engagement questionnaires some couples volunteered for 'Kinsey-type' interviews regarding SB. Others refused to disclose sexual information until after marriage. In the follow-up study detailed m data on SB was obtained from all couples. To test for the biasing effect of voluntarism, marriage questionnaire .items on pre-marital sexual intercourse with spouse & others, on strength of sexual desires, &, for women, on orgasm experience were cross-tabulated for 159 couples who had refused. To test for the biasing effect of org'al membership & participation, data of this sort were gathered in the enment & marriage questionnaires &, similarly, cross-tabulated with questionnaire items on SB. In general, the results of these 2 tests do not support the notion that the indicated aspects of Kinsey's sampling procedure signif'ly biased his % descriptions. Only 1 of 7 % diff's in responses to sex items between volunteer R's & non-volunteers attained statist (X2) signif at .05 level (p=.33). None of the 11 cross-tabulations between sex responses & org'al membership & participation attained the .05 level of signif. W. Delany.