Response Characteristics in a Mail Survey

Abstract

A survey of job attitudes and job behavior was mailed to 2,268 workers and their employers. Nonresponse bias was investigated for biographic variables and three substantive variables: pay, job training, and job attitudes. The small significant effects observed suggest only minor bias. The biographic characteristics which did indicate nonresponse bias, e.g., sex, also demonstrated late-response bias. The use of late-responders' characteristics as a proxy for nonresponse bias needs further investigation.