L'utilisation du telephone dans une enquete de sante mentale: analyse de l'impact du rang d'appel, des donnees manquantes et de l'effet "enqueteur"

Abstract

Use of telephone sample surveys on mental health has increased in France in recent years. As part of an epidemiological survey of the long-term psychological consequences of a natural disaster-the 1992 floods in the Vaucluse-this article begins by examining the relationships between call rank and the population's socio-demographic characteristics and mental health status; the relationship between missing data, socio-demographic characteristies and personality disorders is then examined; finally an attempt is made to evaluate the interviewer effect on the mental health data collected by telephone. The results are consistent with those in the literature for the relationship between call rank and socio-demographic variables. By contrast, no clear relationship is identified between call rank and the various aspects of mental health considered. Lastly, non-response on questions of mental health depends on both the socio-demographic characteristics and the neurotic tendencies of the respondent. The non-response rate and the anxiety level reported by respondents increase when the interviewer is female. The implications of these results for telephone collection of data on mental health are discussed.