Husbands and Wives as Survey Respondents

Reviews the literature of survey response error and husbands/wives as respondents and tests the findings against RAND's Seattle recession survey data (R-1352). Unlike previous studies, RAND found (1) practically no difference between spouses in response patterns on financial adjustments and work history; (2) practically no difference in response consistency between reports by actors and other informants; (3) little difference by wife's employment status; or (4) little difference by similarity of spouses' educational level. Confirming earlier work: (1) response consistency depends on the level of generality of the question asked, and (2) more highly educated couples showed more response consistency. A cost-effective method of choosing respondents is to accept "any knowledgeable adult" in the household. In practice, husbands are interviewed more often in less-educated households where consistency is lower, and wives in the higher-educated households where consistency is higher and husbands are harder to contact. However, an even better approach would be to determine who in each family actually handles finances, instead of using sex role assumptions. (Survey description and questionnaires are appended.)