Using 414 business undergraduates at two universities in the US, a 28-item computer self-efficacy scale is validated and used to examine the relationship between (a) training and computer self-efficacy and (b) user attitudes and computer self-efficacy. Survey responses were collected both at the beginning and end of an introductory computer course. A principal factor analysis of the computer self-efficacy scale supported a conceptually meaningful four-factor solution with high alpha reliabilities. Results suggest that training significantly improved the computer self-efficacy of males and females in this study for all factors. Training programs seemed more effective for male and female respondents with positive attitudes toward computers. Training programs seemed less effective for respondents with negative attitudes toward computers. Implications of these findings are discussed and research opportunities described.