Interviewing older adults: validity of self-reports of satisfaction

Respondents in a personal interview study were asked to evaluate several domains of their lives using several different response scales that included both verbal and pictorial response alternatives. Structural equation models were specified to permit estimation of the validity of these survey measures. Across all of the items, the concept factors explain an average of about one half of the total variance, method factors explain about one tenth of the variance, and the remaining one third or so of the variance is unique to each item. Comparisons of the validity of reports by respondents of different ages indicate that method factors explain almost twice as much of the variance in the responses of older respondents as in those of younger respondents.