Abstract
We analyse if and how the characteristics of grant research panels affect the applicants' likelihood of obtaining funding and, especially, if particular types of panels favor particular types of applicants. We use the award decisions of the UK's Engineering and Physical Sciences Research Council (EPSRC). We show that not only applicants' but also panels' characteristics matter. Panels of higher quality, in terms of prior research performance, for instance, as well panels that include more female members or members of Mongoloid origin, are tougher than others. Our main results indicate that panel members tend to favor more (or penalise less) applicants with similar characteristics to them, as the similar-to-me hypothesis suggests. We show, for instance, that the quality of the applicants is more critical for panels of the highest quality than for panels of relatively lower quality, that basic oriented panels tend to penalise applied-oriented applicants, and that panels with less female members tend to penalise teams with more female applicants.