Our study adds to prior work on Indiana's Twenty-first Century Scholars(TFCS) program by focusing on whether participating in--rather than completing--the program affects the likelihood of students going to college and where they initially enrolled. We first employ binary and multinomial logistic regression to obtain estimates of the impact of the program without controlling for self-selection into the program and find that signing up for the TFCS program has a relatively small effect on whether students go to college and where they enroll. We then used alternative statistical approaches to reduce the bias due to self-selection into the TFCS program. Although the estimated program effects after accounting for self-selection were larger than in the earlier models, they were still notably smaller than reported in prior studies of the program.