The United States has been a very attractive destination for foreign science and engineering graduate students and postdoctoral scholars for a considerable period of time. Despite the important role of foreign doctoral students in the U.S, relatively little is known about the factors influencing their decision to attend an institution. One factor that is rarely explored is the effect of networks on institution selection. This study aims to provide both qualitative and quantitative information about the role networks play in foreign doctoral students institution selection. This three-part study utilizes different methodologies: (1) focus group interviews conducted with Turkish doctoral students at the Georgia Institute of Technology; (2) a web study of research laboratories in science and engineering; and (3) the estimation of Random Utility Model (RUM) of institution selection. Guided focus group interviews provide important qualitative information about the ways students, alumni, faculty and local community of same nationality influence institution choice. The web study of research laboratories provide evidence that labs that are directed by foreign-born faculty are more likely to be populated by students from the same country of origin than are labs that are directed by native (U.S. born) faculty. The results from RUM of institution selection provide strong and significant evidence for the relationship between the number of existing students from a country of origin at an institution and the probability of attending that institution for potential applicants from the same country of origin. Also, in some of the models there is evidence that the alumni and faculty from the same origin also play a role in student choice. The results of this study have several policy implications related to integration of foreign doctoral students, future enrollments, institutional mismatch , and the role foreign-born faculty play in U.S universities.