I'd say that makes out a prima facie case that university hiring practices are having a disparate impact. The burden of proof is therefore on the university to show why this disparity exists.Three problems. First, without greater knowledge of the sample sizes and a number of other characteristics, we don't know whether the change from a 39% left-liberal, 34% right-conservative split in the 1984 survey to a 72%-15% split in the 1999 survey is the result of chance. Second, there are changes in the information faculty who responded to the survey acted upon. In 1984, President Reagan's tax cuts, Paul Volcker's monetary policy, and Yuri Andropov's crudity were on peoples' minds. In 1999, the religious wing of the Republican Congress impeached but failed to convict President Clinton. Third, there are changes in the faculty itself. A more vibrant private sector coupled with retrenchments and recisions in the universities might have done more to divert potential conservative scholars to the think tanks or away from the research life completely than the PC atrocities that had been coming to light. The residual effect of self-selection among Marxist historians or nihilist literary critics or socialist physicists might be small.
DISPARATE IMPACT, QUANTIFIED? Professor Bainbridge has two posts addressing a Chicago Tribune West forum on the propriety, or not, of some dissident UCLA (motto: On! Wisconsin!) alumni providing beer money for evidence of leftist proselytizing masquerading as teaching. This quote, from the earlier post, shows a danger in statistical inference.