Friday, August 20, 2010

The strange dynamics of faculty merit pay

In many US universities, faculty performance is rewarded with merit bonuses or increases, the initial idea being to prevent other universities from poaching the best performers. But seeing how this is approached in a very heterogeneous way across institutions, one sometimes wonders whether the merit process is done optimally.

Finn Christensen, James Manley and Louise Laurence had access to much relevant data in a large public university. There, merit pay is distributed in each department by a committee of tenure faculty. The outcome is that about two thirds of all faculty reach the highest merit scale, three fourths among tenured faculty. Now this outcome could be justifiable with additional data, which the authors have for a particular college in this university in the form of various output measures that should matter for evaluation. It turns out that untenured faculty is as productive, but still gets less merit pay. Even worse, it appears that the output measures explain about 10% of the variation in merit.

What is going on? Christensen, Manley and Laurence show theoretically that given the institutional structure, one should not be surprised by these results. I can believe that, and this is probably compounded by compression: as new faculty commands higher pay than old, the old compensate with higher merit. The authors find little evidence of this. Internal politics also certainly play a role, as you want to avoid offending someone who may be determining your merit later. But what is clear is that there is, at least at this place, very little transparency in merit attribution.

2 comments:

Kansan said...

In my department, the chair does all the merit based on vague criteria set by the department. I have found merit outcomes rather puzzling at times, but as I have not been on the losing end, I am not complaining.

James said...

Thanks for posting our research! Education is notoriously difficult to evaluate- lately the trend has been to see how much teachers improve children's test scores, but hopefully we want educators to do more than just teach to some test. In college, often it's student evaluations that are used, but there are many ways to get students to say they like a teacher! To motivate teachers, institutions like the one we studied impose merit pay. Far too often that ends up captured, a political prisoner that functions only to force teachers to cozy up to their supervisors. Until a better index is created for evaluating teachers, there can be no functioning market, which you and I know is how we get to an efficient outcome! Again, thanks for calling attention to our research.