Estimation of the number of sexual partners for the nonrespondents to a large national survey.

Abstract

For a random sample of 4548 respondents to the British National Survey of Sexual Attitudes and Lifestyles, item nonresponses were assessed by utilizing alternating least squares and optimal scaling. Of men who did not answer the question on number of lifetime partners, 61% were estimated to have had 10 or more partners, whereas 48% of female nonrespondents were estimated to have had no partners. Among the latter, the percentage of those with 10 or more partners was also elevated to the level almost twice that of respondents. However, the overall impact of the estimates on the distribution of lifetime partners was hardly noticeable due to the relatively small proportion of incomplete answers to this question. Those who did not respond to the question of number of lifetime partners were likely to be over 45 years, of lower social class, and poorer education in comparison with those who answered the question fully. A larger proportion of the former were of Asian extraction and non-Christian in comparison to the latter, particularly among women. Not providing a complete answer to the question on the number of lifetime partners was associated with the nonresponse rate in the range of 90% on the questions of STD clinic attendance and drug abuse. Heavy smoking was more prevalent among male nonrespondents in comparison to those who provided a full answer to the question on the number of lifetime partners.