US College Enrollment: What Matters Most

In a well-conceived 2008 article, Su Jin Jez examines the factors affecting whether a given US student will attend college.  It is no small feat to establish cause and effect relationships when factors of interest -- race, academic preparation, and family wealth -- can all so easily confound one another.

Jez astutely uses statistical analysis (regression) to disentangle these complex relationships.  Working out of a UC Berkeley think tank, the author draws on a large-scale, nationally-representative study of data from the Integrated Postsecondary Education Data System (IPEDS).  Her research shows, in a nutshell, that

• when you control for wealth, race is no factor in college enrollment;
• when you control for academic preparation, wealth is no factor either!

So students from wealthier families will more likely attend, regardless of race, and this connection is true only because wealth is paired with better academic preparation.

The article is long on methods and findings and a little short on discussion of implications for K-12 or higher education.  But even so it constitutes a good example of the way one can use sequential regression, in planned stages, to clarify otherwise puzzling relationships among variables.