At the University of Connecticut, we screen at various levels:

1.  For competitions where only one proposal (investigator
initiated) can be submitted, we request abstracts and relevant CV, and
form a faculty screening committee.

2.  For graduate student competitions, again investigator
initiated, we also screen in the same way.

3.  For a state-wide high tech competition, where we were told we
had to rank proposals by priority, we not only requested
abstracts, but had faculty come before a faculty review panel
for an oral 5 minute presentation.  Faculty opinion was that this
would take less time of the applicant than a "pre-
proposal."

4.  For other types of proposals, such as NSF facilites or large
equipment competitions, the ability of the particular department
or school/college to cost share; consideration of those areas
targeted for growth in the University (aren't many of us
targeting high tech areas?);strategic plan of the University,
etc.; are all considered.  Final ranking is by the Director of the
Research Foundation and the Provost.  Again, an abstract is
requested, using guidelines from the funding agency (University
commitment, availability of equipment or facility to other
researchers, funding of listed researchers).

5.  Oddly enough, we don't have many objections from faculty
using this process. We try to be very honest, and give each
applicant reasons why others more closely match the
evaluation criteria spelled out by the funding agency, (e.g.,
more closely fit the agency's priorities, cost sharing in place,
industry involvement clearly spelled out)

6.  This review takes much more work on our part; ideally,
however, one would have time to fluff up the final proposal.  After
all that, a poorly written proposal hurts.