The reproducibility of an interview questionnaire on sexual behavior in Japanese males was estimated with the use of Spearman's rank correlation coefficient (r) and kappa coefficient. The subjects were 151 males aged 50 years or older (mean age = 63.6 years old) in a rural town of Hokkaido Prefecture, Japan, who participated in a consecutive mass screening for prostatic diseases at a one-year interval after their first participation during the period 1988-1993. They were individually interviewed about their sexual behavior by well-trained urologists who were supervised by one of the authors.
A remarkably good reproducibility level at the one-year interval was shown among the questions about age at first marriage (r = 0.92), age at which first sexual intercourse was experienced (r = 0.80), and age at which sexual intercourse was no longer practiced (r = 0.87). The reproducibility of the questions about frequency of sexual intercourse when they were in their 30s, 40s and 50s was fair (r = 0.2-0.6), and that of the questions regarding the frequency in their 60s was poor (r < 0.2). However, there were rare extreme differences in answers of this frequency between the two interviews. In conclusion, the reproducibility of the questionnaire on sexual behavior in Japanese middle-aged or elderly males was considered to be satisfactory.