Research note: proxy responses in health surveys: a methodological issue.

Abstract

The practice of using proxy respondents in health surveys has been identified as a possible cause of confusion in the literature concerning sex differences in morbidity. Specifically, it has been suggested that proxy respondents (usually female) tend to under-report illnesses of absent persons (usually male) and hence the observation that women suffer from more illness than men may simply be an artifact of this method of data collection. In order to investigate proxy effects, 50 married couples were asked to report symptoms recently experienced by their spouses as well as themselves. A high level of agreement was found between husbands and wives regarding the number and severity of symptoms experienced by each individual, despite considerable differences in health status between sub-groups. Where there was a discrepancy, it was in the direction of the proxy respondent (of either sex) over-reporting symptoms. It follows from this that, if anything, the over-use of female interviewees would tend to minimise observed sex differences in illness. Thus, the excess female morbidity found in health surveys is not attributable to proxy effects but is more likely to result from differences in the way in which men and women experience symptoms.