M entor (n): A trusted adviser. In Greek mythology, Mentor was Odysseus' trusted teacher and counselor.

F ellow (n): A student in training appointed to a position granting financial aid and providing for further study in exchange for performing academic duties.

As I near the end of my term on the National Institutes of Health (NIH) Fellows Committee ( Felcom), I continue to be astounded by how much work Felcom does for NIH's Intramural training program. Currently, there are 13 subcommittees, each serving at least one of two major purposes: facilitating communication between fellows and the NIH community, or working with the administrative programs that affect the training experience. The Felcom Mentoring Committee serves both.

At any given time, NIH has over 3000 fellows in the NIH Intramural Research Program, each of whom depends on effective mentoring for a successful training experience. Recognizing this, in 1998 Michael Gottesman, the deputy director for Intramural Research (DDIR) at NIH, published a booklet, "A Guide to Training and Mentoring in the Intramural Research Program at NIH". In it, he defines a mentor as "a person who has achieved career success and counsels and guides another for the purpose of helping him or her to achieve like success." The guide is specifically for supervisors, mentors, and trainees, and it provides insight on the general foundations on which training programs should be based.

Shortly after the publication of the booklet, the Mentoring Committee conducted a preliminary Web-based survey to establish a baseline of the NIH fellows' assessment of the quality of their mentoring. The 20% response rate was disappointingly low, and there was no way to assess the demographics of the responders. The DDIR and the scientific directors encouraged Felcom to repeat the survey, using statisticians and epidemiologists as consultants to ensure the validity of the results.

When I joined Felcom a year ago, I chose to serve on the Mentoring Committee, where preparations for a second, more robust survey were already under way. At this point, the committee's objectives are to gather information regarding the fellows' feelings about the mentoring they have received and then to make recommendations to the NIH about how the mentoring programs can be improved.

To this end, we have prepared a second survey for distribution to a randomly selected group of 750 fellows in the spring of 2001. Our primary goal is for a 70% response rate--hopefully, the lessons we learned from the first survey and the steps we've taken as a result will increase the likelihood that we'll achieve this higher response rate.

There are a few things that make this survey much different from the last. We made a couple of changes to ensure confidentiality, because we believe this issue contributed to the last survey's low response rate. We decided against a Web-based survey and sent the hard copies of the questionnaire by Federal Express. Although this entails more work and expense, we felt that the Web-based survey left open questions about a possible e-mail trace between the survey participants and internal NIH personnel. Additionally, we have arranged for a consulting company to administer the survey *. All mailings will be handled by the company, and the results will be available to our committee--and to the NIH--in summary form only.

For this survey, we have included incentives to respond to the survey. Overall, 10% of the sample will either have lunch with the NIH director, or receive a gift certificate to the Foundation for Advanced Education in the Sciences Bookstore. Lastly, based on our consultants' suggestion, we are conducting a random sampling of 750 of the 3000 fellows. This approach allowed us to focus our resources to ensure a higher response rate and a more effective sample. One of the problems that we encountered is that there are a couple of institutes within NIH that are very large (> 200 fellows), and some that are very small (< 50 fellows). We designed the survey to over-sample the smaller institutes to make sure that we would have enough responses from them to still be able to work with. This design is such that the proportion of the fellows sampled from each institute will be represented equally.

The survey contains 28 questions covering five major categories that we think are important for a good mentor/fellow relationship: 1. Defining a mentor(s), 2. Finding a balance between independence and guidance in research projects and future goals, 3. Access to the mentor, 4. Credit for work in authorship and presentations, and 5. Scientific networking and career development. Included with the questionnaire will be a cover letter stating our goals and purpose, as well as the incentives to respond to the survey. The nonresponders will be reminded at various times throughout the course of the survey, and if necessary, they will be sent a second questionnaire.

Once the summaries of the responses are received, we intend to calculate summary statistics on the responses, e.g., mean, median, confidence intervals. We will use simple statistical approaches to compare one group to another, e.g. by institute, gender, U.S. citizens vs. non-U.S., etc. After the final analyses, we will summarize the results to identify areas of concern and address possible solutions. We will make our report to the DDIR and to the scientific director of each institute. Through our efforts, we hope that the mentors will learn to mentor better and the fellows will learn to help the mentors help them.

In Part II, we'll summarize what we learned from our survey, including details on how to implement a survey like this at your institution; to help with that, we'll provide a link to the questionnaire that we designed. But don't hold your breath. ... We do not expect the results to be ready before the winter.

*Dr. Joni L. Rutter is a fellow at the National Cancer Institute in the Division of Cancer Epidemiology and Genetics. She serves as a Fellows Committee representative and is chair of the Mentoring Committee. She received a Ph.D. from the department of pharmacology and toxicology at Dartmouth Medical School and is currently working on the genetics of breast, ovarian, and prostate cancer susceptibility in the Laboratory of Population Genetics at NIH.

*The Office of the Director at NIH is going to great lengths to improve the mentor/fellow relationship and is funding this project.