To investigate the effects of demographic variables on changes in rates of depressive symptoms in a non-clinical population, a 4-year follow-up study was conducted on a total of 6378 employees of a Japanese electrical company. Data were collected five times at 1-year intervals (T0-T4) using a questionnaire that included the Zung Self-rating Depression Scale (SDS). Of the non-depressed (i.e. having a SDS score of 47 or less) at baseline (N = 4857), 14% were found depressed at least once during tj-t4. Younger, never married, female and blue-collar workers were significantly at greater risk for becoming depressed during the follow-up period (P < 0.05). Of the depressed at baseline (N = 505), 20% were depressed every time during T1-T4Younger workers were significantly at greater risk for persistence of depressive symptoms during the follow-up period (P < 0.05). The results suggested that age is associated with both occurrence and persistence of depressive symptoms, while gender, marital status and occupation are associated only with the occurrence. © 1995, Cambridge University Press. All rights reserved.