was treated more as a colleague than as an outsider. This was facili-
tated by the welcoming attitude that program officers and panel
chairs manifested toward me from the onset. During the course of
deliberation, I tried to be as inconspicuous as possible as I sat quietly
at the end of a table. I took ample written notes andavoided betray-
ing my reactions to arguments made by evaluators.
Reactivity and Confidentiality
I acknowledge that reactivity (the impact of the identity of the inter-
viewer on the object of study) probably characterized both the dy-
namics of the panels I observed and the content of the interviews I
conducted. It is likely that in anticipation of being interviewed, pan-
elists were more reflexive about the criteria they used and about their
behavior as evaluators. They may have been more aware of possible
biases, and more likely to go the extra mile to try to limit their im-
pact. Thus, their emphasizing that “cream rises” was certainly accen-
tuated by the interview situation. But the in-depth interviews I con-
ducted convinced me that this view is so pervasive (as illustrated by
the data quoted in this book) that it cannot be explained solely by
the communicative act that the interview represents.
One of the panels I observed in Year 1 was quite contentious. The
funding organization refused to let me observe deliberations in Year
2 in part because the chairperson was concerned about reactivity,
and also perhaps because of the discomfort panelists experienced
from having heated debates in front of an outside observer.
The relationships I had with the panelists had a small impact on
the interviews. I had close connections with only two respondents
(one is a personal friend; the other, a close colleague in my field). I
had a more distant relationship with three other panelists; their areas
of research being somewhat related to mine. This relatively low de-
gree of connectedness probably weakened the impact of reactivity on
my results.
254 / Appendix