Advancement Via Individual Determination (AVID) is a college readiness program that attempts to address both the academic and informational barriers to postsecondary education. AVID is designed to prepare students in the academic middle for postsecondary enrollment and success through a combination of support services and increased expectations. The AVID program, which serves more than 425,000 youth in the United States alone at the middle, high, and postsecondary levels, has been the subject of significant research. But prior studies on AVID have ignored student exit from the program despite evidence that exit rates of 50% or higher are common. In this dissertation, I use a mixed methods approach to study the causes and effects of AVID exit in an urban school district in the Midwest. For the qualitative component of my study, I interviewed AVID coordinators and teachers about their perspectives of student exit and analyzed interview data to assess the alignment between AVID exit and role exit theory, which focuses on disengagement from a role that is central to an individual's identity. For the quantitative component, I used summary statistics to describe patterns of student exit. I also used Propensity Score Matching to create matched groups of AVID exits and AVID persisters in order to estimate the impact of exiting AVID. Overall, I find that role exit theory aligns very well with AVID exit, and that teachers generally believe that ex-AVID students struggle post exit. The estimated effects of AVID exit resulting from the Propensity Score Matching procedure are larger for attendance and behavior, smaller for Grade Point Average, and negligible for standardized test scores, although even the largest effects on attendance and behavior are in the range traditionally considered to be small. In addition, I find that some ex-AVID students achieve success post-exit and that teachers can identify many situations where AVID exit is beneficial for the student and the AVID program. I argue that AVID exit is common and should receive explicit attention in future AVID research, and that AVID exit is not necessarily something that should be avoided in all cases.