Widening the Research Pipeline Update to NSF/CISE BPC Evaluation Workshop December 7, 2006.

Widening the Research Pipeline

Update to

NSF/CISE BPC Evaluation Workshop

December 7, 2006

Purpose of Project

Broadening participation in the research end of computing

In particular, encouraging underrepresented groups to earn PhDs and develop research careers in academia or industry

This means starting at the undergraduate level to put students on a research track, then continuing to support researchers throughout their careers

Alliance Participants

CRA-W

Computing Research Association Committee on the Status of Women

Focus on women in research

CDC

Coalition to Diversify Computing

Focus on underrepresented minorities in grad school and workforce

Multiple Points of Intervention

Undergraduates Think of grad school and research as a possibility Know how to apply to grad school Have credentials for acceptance to grad school Know how to obtain funding

Grad Students Persist in grad school Find a research area and project Do quality research Complete degree

Career Management Find a job Continue research Manage career

CRA-W and CDC Programs Covering the Research Pipeline

Alliance Activities

Discipline specific mentoring: Connect senior and junior researchers and graduate students within a specialty area to provide discipline specific mentoring and tutelage “best” journals and conferences, hot research areas networking skills and building collaborations

Undergraduate Mentoring: Coordinate and improve existing undergraduate programs (CREU, DMP, DRS)

Extended mentoring: Continue and build relationships after initial research experience Publish results, help with grad school application. Tri-mentoring: one student, two mentors – perhaps one academic and

one industry, or one at student’s home school and one at a research U. Distinguished Lectureship: fine-tuning the Lectureship Series to

focus on recruiting students, especially students from underrepresented groups, into grad school.

First Year Activities

Improve coordination between CRA-W and CDC “weekly” phone calls

Changes and coordination for mentoring programs Develop co-leadership for CREU Define an expanded DMP program for women and

minorities Institute changes in the mentoring programs

Hold first discipline specific workshop

Computer ArchitectureSummer Workshop Area selected because women are even more underrepresented

than in the field overall minority numbers are so low that area representation is meaningless

Workshop participation Held in Princeton, NJ, July 19-21 2006 45 attendees, most grad students or new PhDs or faculty 83% female, 27% underrepresented minority 22 presenters including very senior experts

Workshop content Technical sessions (e.g. What Computer Architects Should Know About

Compilers and Systems Software) Career / research process sessions (e.g. Communication Skills:

Presenting Research and Talking Informally About your Work) Networking opportunities, e.g. poster session, picnic dinner

Workshop Evaluation

Paper surveys to participants at end of workshop Online survey to presenters a few weeks after

workshop Because this was the first discipline-specific

activity and others are to follow, focus was strongly formative: what worked, what didn’t, recommendations for future organizers

General reaction highly positive, praised the discipline focus, the networking opportunities, and the level and variety of content

Future Plans for Workshops General Call for Proposals

First workshop pulled together in a short time after funding was confirmed, but topic had been decided and organizers were waiting

To continue, must identify additional disciplines and workshop leaders and develop a selection process

Significant interest from other disciplines Programming languages planned for May 2007 Operating Systems has a tentative proposal Machine learning held a lunch at AAI with our support

Because of interest, will pursue additional funding

Coordination Challenges Some overlap in existing CRA-W and CDC

programs Needs of women and minorities are similar in some ways

and different in others Need to balance goals and working methods of alliance

members Balancing activities that reach many students less

individually vs. few students more intensively Needs vary by environment

a minority student in a majority institution may need different support/encouragement than a student in a nurturing MSI or women’s small liberal arts school

Coordination Approaches

Regular teleconferencing between leaders of Alliance members

Gradual integration of existing programs, e.g. adding CDC representatives to the selection committees for undergraduate research programs that were previously all-women

Co-leadership of activities

Evaluation Challenges

Evaluate both individual activities and overall program

Plan evaluations to encourage cross-activity comparisonse.g. ask similar questions

Maintain ongoing evaluation of changing programse.g. add the extended mentoring aspect to the

existing Distributed Mentoring Program

Evaluation Approaches Continue evaluation approach we have used in the past

Immediate post-activity evaluations Long-term follow-up by tracking survey

Expand it to include new activities Mix of internal and external evaluation

Ask same questions across activities, when reasonable, to help comparisons

The question is not just “Does it work?” Try to answer “For whom does it work best?” or “Under what circumstances does it work best?”, or “What parts of the experience are key to it working?” Differences between women & minorities Differences between undergrads at large research universities and

smaller schools Participants at different points in their schooling Different aspects of the mentoring experience

Keep in mind the big picture while looking at the individual activities