Scholars are increasingly using administrative data to determine whether multi-channel college admission system benefits the upper middle class. However, administrative data lack information about family background, and such research usually replaces it with the regional characteristics. One problem is that regional characteristics and family background operate at two different levels. A measure supportive of rural students does not necessarily benefit the students from low socioeconomic-status families. Moreover, even if an admission channel favors a certain type of student, it shouldn't automatically be considered unfair. A bias favoring upper middle class students should mean that they are admitted due to overestimation for the academic ability or matching degree. It is necessary to use social surveys to collect information on the family backgrounds of students. This paper uses Sun Yat-sen University survey data to demonstrate the logic and methods for clarifying these issues.

Keywords: multi-channel admission, socioeconomic status, urban-rural differences, academic performance