EDUCATION REPORTED IN INTERVIEWS: AN ASPECT OF SURVEY CONTENT ERROR.

This paper examines differences in level of educational attainment as reported for the same individuals in two questionnaires administered at approximately a three-year interval. The use of education as a research control variable and major indicator of socioeconomic status needs no amplification. Because of the importance of respondents' educational achievement in surveys and polls, this paper is presented as a contribution to the understanding and possible improvement of interview data reliability. [ABSTRACT FROM AUTHOR]