Far from random? The role of homophily in student supervision

Giulia Rossello & Robin Cowan

#2019-024

The paper studies racial and gender homophily in student supervision relationships in a context of social transformations, South Africa academia. We develop a technique to separate choice homophily from that induced by the system. Comprising two permutation tests repeated at two levels of aggregation, system and departments. We find clear evidence of homophily in student supervision, along racial lines in particular. Roughly half of the observed homophily is induced by the departments composition and stays constant over time. Overall, choice homophily has similar magnitude along racial and gender dimensions. Further, we ask where choice homophily originates in the demographic groups of students and professors. We find that white (male) students have high tendency to form same-type relations, while among professors it is black (female) who display the higher frequency. Group differences show that choice homophily is likely to originate from students in the former majority.

JEL Classification: A14, D71, D85, I23, I24, J15, J16

Keywords: Academia, South Africa, Student supervision, Induced homophily, Choice homophily, Segregation, Assortativity mixing, Permutation test, Social Transformations, Social Change, System of Organisations, Institutional constraints, Gender ties, Racial ties, University System emerging countries, Racial and Gender Homophily

  


UNU-MERIT