Using Verbal and Paralinguistic Behaviors to Explain Variation in Responses to Self-Reported Health Questions

The self-reported health item summarizes a wide range of information on health status across several domains of health, and is widely used to measure health because it predicts mortality well (Idler and Benyamini, 1997). We examine whether the characteristics of the interviewer-respondent interaction during the administration of the self-reported health item reflect in part actual health status and thus may be useful indirect measures of health status when limited data on health are available. We find that the odds of several characteristics of the interaction are significantly lower for those with better self-reported health. We also find that reporting a larger number of health conditions is associated with higher odds of each of the characteristics of the interaction we examine, and that many of these relationships remain significant after controlling for the respondent’s self-reported health and cognitive ability. Finally, we find some evidence that the characteristics of the interaction may be able to explain variation in health status among respondents with the same self-reported health, in that most of the characteristics of the interaction we examine are associated with the number of health conditions reported among respondents whose self-reported health is “excellent.” Overall, we find promising evidence that the characteristics of the interaction might reveal important information about respondents’ health status not captured by the selfreported health question. We recommend further investigations of whether researchers using self-reported health to measure health status would benefit from coding the characteristics of the interaction as an additional, indirect measure of health status.