Premarital contraceptive use: A discriminant analysis approach

Abstract
The ability of seven independent variables to predict accurately, from a sample of unmarried, sexually active students, which ones use reliable contraceptives and which ones use unreliable contraceptives, was investigated. Using discriminant analysis, the seven independent variables were age individuals started engaging in coitus, frequency of coitus, frequency of dating, length of time sex partners knew each other, number of sex partners, having anticipated coitus to occur, and the number of close friends who were thought to use birth control. The results indicated that these variables were accurate in predicting which students were users of contraceptives, with more than 80% of males and females being correctly classified. The number of close friends thought to use contraceptives was the most influential variable for both sexes, followed by length of time the partners knew each other.