Discussion 2: Self-Selection and Performance-Based Ratings: A Case Study in Program Evaluation

Abstract

I’d like to congratulate Burt on the really brilliant display of graphics and data. I’m rather pleased to see none of those fancy Bell Labs type that always make the rest of us feel envious. Moreover, I can’t think of a graphics package that could do routinely anything like what Burt has done. Burt has presented an enormous amount of data about a complex problem and really covered all kinds of questions. The unusual combination of text and symbols and figures is especially important. Most graphics packages can draw lines and symbols and things, but they generally fall down on the kind of combination shown here. Yet that’s just what you want to see. I do really admire the graphics and display—I think it was excellent.