A Problem with Data Consistency

Copyright © 1999 Kevin T. Kilty, All Rights Reserved

As many have heard, no doubt, the University of California changed its affirmative action policy a couple of years ago, and a segment on The News Hour with Jim Lehrer was attempting to illustrate the impact that the new rules had on minority admissions to University of California at Berkeley. Here is the graphic that they placed on the screen.

Ethnic profile of Berkeley entering class
Category19971998
Black6.82.4
Hispanic15.47.6
White33.034.2
Asian35.538.2

Indeed, as commentators point out, the enrollment of black and hispanic students declined, but what they failed to include was a totals row which would show that the enrollments added to only 90.8% in 1997, and 82.4% in 1998. The percentage of students missing from the analysis increased enough to almost account for the drop in black and hispanic enrollments. Who are the missing groups? Are they mixed race? Do they refuse to identify their race? It is difficult to get excited over data that contains such an inconsistency.