BACKGROUND Understanding more about student decisions to leave engineering may lead to higher retention. This study builds on the literature and focuses on the experiences of a cohort of students who aimed to complete their undergraduate work in 2007. PURPOSE (HYPOTHESIS) This paper presents the outcomes of the longitudinal administration of the Persistence in Engineering survey. The goal was to identify correlates of persistence in undergraduate engineering education and professional engineering practice. DESIGN/METHOD The survey was administered seven times over four years to a cohort of students who had expressed interest in studying engineering. At the end of the study, the participants were categorized as persisters or non-persisters. Repeated measures analysis of variance was used, in conjunction with other approaches, to test for differences between the groups. RESULTS Persisters and non-persisters did not differ significantly according to the majority of the constructs. Nevertheless, parental and high school mentor influences as a motivation to study engineering, as well as confidence in math and science skills, were identified as correlates of persistence. Intention to complete an engineering major was also a correlate of persistence; it appears to decline sharply at least two semesters prior to students leaving engineering. The findings also suggest that there might be differences among non-persisters when they are further grouped by when they leave engineering. CONCLUSIONS Facilitating higher levels of mentor involvement before college might increase student motivation to study engineering, and also constitute a mechanism for fostering confidence in math and science skills. Since the intention to complete an engineering degree decreases well before students act, there may be opportunities for institutions to develop targeted interventions for students, and help them make informed decisions.