Document Type


Publication Date



Since 1973, the McDonnell Douglas framework has been a key analytical structure in employment discrimination law. Academic debate regarding the framework has alternately sounded its death knell, posited its irrelevance, or asserted its continued vitality. What has gone unnoticed in this discussion is the gradual weakening of the framework over the past two decades. Rather than casting this test into oblivion, courts are slowly chipping away at its preeminent place as a proof structure.

Little by little, courts are gradually eroding the McDonnell Douglas test's power through both procedural and substantive means. Procedurally, courts have questioned, rejected or diminished the use of McDonnell Douglas at the pleading stage, in jury instructions, and when considering post-verdict motions. Substantively, a growing number of courts have created alternate structures, de-emphasized McDonnell Douglas as the primary proof structure, or openly called for the test's demise. These retrenchments from the burden-shifting framework established by McDonnell Douglas constitute powerful criticisms of the test's larger applicability.

This Essay demonstrates the substantive and procedural turn away from McDonnell Douglas. It argues that courts and scholars should alter how they discuss the framework to accurately reflect the test's current place in discrimination law. Current ways of describing the test overstate its importance by failing to recognize the rise of other structures by which to analyze employment discrimination and the narrow procedural window in which the courts now use the burden-shifting test.

Part I explains the McDonnell Douglas decision and its progeny. Parts II and III demonstrate how courts have diminished the importance of the framework substantively and procedurally. Part IV discusses major critiques of the framework. Part V argues that courts and scholars should alter the language used to discuss McDonnell Douglas to better reflect the procedural and substantive retreat from the test.


This article was published in 34 Berkeley J. Emp. & Lab. L. 257 (2013).