In 1978, teacher trainees who were undertaking a postgraduate teaching diploma at the Sydney Teachers' College were invited to participate in a longitudinal study, and 156 of the original 170 cohort subjects have provided full data on work, social networks and patterns of illness, and completed self-report measures at 5-yearly intervals since 1978. Cohort subjects were divided into 'teaching' groups (those whom stayed teaching) and 'not teaching' groups (those whom left). Results indicate that the majority of the 'not teaching' group left within the first 5 years of teaching, with self-rated social desirability predicting teacher retention in 1983. Those who were still teaching in 1993 were younger and had a more positive view of teaching prior to commencement in 1978, as evidenced by preference to remain in teaching, seeing it as more enjoyable and greater exposure to a positive role model. The cohort seemed to have preconceived ideas about teaching prior to entering into the classroom and these ideas influenced their decision; however, there is still a need to consider personality style when assessing reasons for teacher retention.