A periodically administered survey of the postdoc population can provide accurate information that would help policy-makers establish priorities and also set a baseline against which changes in science postdocs' conditions can be measured. In many cases, postdocs have taken this task upon themselves, designing and administering surveys  at the departmental and institutional levels. For local change, this process can be effective. However, without policy intervention at the national level, Freeman and colleagues  argue that "market incentives will perpetuate the current structure, benefiting senior investigators at the expense of new entrants."
Clearly more data are required to craft national policy. Stakeholders believe that means either coordinated local surveys or a national survey. Yet even without coordination, local efforts have contributed to the national dialogue. For example, the results of several local postdoctoral association (PDA) surveys corroborate the findings of the Committee on Science, Engineering, and Public Policy ( COSEPUP ) (2000) and the AAU Committee on Postdoctoral Education  (1998), both of which identify career development, salary, and benefits as main concerns of postdoctoral scientists. However, beyond these general statements it is difficult, if not impossible, to compare existing PDA surveys.
The Postdoc Network has, nevertheless, compiled a summary of lessons learned and promising practices derived from a meta-analysis  of 12 surveys undertaken by different PDAs. Here?s some more of what we found.
Top Five Survey Topics
The following survey topics crop up frequently:
Career goals. Changing postdoc aspirations before and during appointments
Career advising. Becoming aware of the job market, career options, and requirements
Professional development. Identifying appropriate development of skill sets and competencies (e.g., developing research-based instructional methods, writing proposals and technical papers, giving oral presentations, and networking) and receiving performance evaluations
Salary. Initial pay and current pay
Benefits. Health care, dental, and vision insurance; single and family coverage; disability and retirement insurance; leave policies; and child care assistance.
PDA surveys have posed numerous questions about each of these topics. Experienced PDAs have begun to limit the questions asked to gather only the data that they have found to be the most valuable (e.g., most often quoted). The Stanford University Postdoc Association ( SUPD ) is a case in point. SUPD has conducted surveys each year since 2000. Its most recent effort (available on its Web site) is only one page, with questions on demographics, career goals, health benefit plans, satisfaction with the postdoctoral experience, housing, and SUPD?s role.
PDAs have found the following demographic information to be critical when developing executive summaries and action points for institutional and PDA attention:
Number of months as a postdoc at the current institution
Number of months as a postdoc at other institutions
Ethnicity and race
Citizenship and visa status
Partnership status (e.g., single, married, domestic partnership)
Dependents (child and nonchild, and number of each)
Source of funding.
PDAs have found that using demographic categories to analyze survey responses can enrich understanding and help determine their significance. Such focused analysis permits PDAs and administrators to target resources and programs to particular populations.
For example, the Postdoctoral Scholars Association of the University of California, Davis  detected statistically significant differences between the experiences and opinions of international and domestic (U.S. citizens and permanent residents) respondents. International postdocs were twice as likely as domestic postdocs to be in labs with eight or more trainees (graduate students and postdocs). Although the international postdocs also indicated that they had recently presented research outside their group at roughly the same proportion as domestic respondents, a significantly larger portion of international respondents stated that they had never presented outside their group.
Just as bench research does, survey-based research profits from technique being refined, including increased precision in wording the questions. To minimize ambiguity, the University of Chicago Biological Sciences Division Postdoc Association (BSD-PDA) solicited expertise  from those experienced in survey-based research. Several disciplines (e.g., social scientists in demographics, or business and agriculture researchers doing needs analyses) have the unique technical knowledge and skills to assist in survey construction, implementation, and data analysis.
Most PDA surveys effectively use a variety of question formats--open-ended text, multiple choice, and Likert scales  for opinions and evaluations. Feedback on several surveys has demonstrated the need to use multiple-choice instead of open-ended questions to reduce the number of questions yielding no response. For example, the Case Western Reserve University Postdoctoral Researchers Association  discovered that questions concerning job title and funding source were answered by more respondents if possible options were listed in a multiple-choice format.
As in all survey research, the results must represent a sample large enough to be credible. One concern with many PDA surveys is the low response rate. The lack of established lines of communication has made it necessary for PDAs to do a lot of preparatory legwork. A survey campaign may even require a lab-to-lab census to identify postdoctoral fellows, as the Brown University Postdoctoral Association  discovered. In addition, the variety of postdoc job titles begs the question: Who is a postdoc? That is, who should be included in the survey population?
Foreign nationals represent about 50% of the U.S. postdoctoral community. Thus, for a representative profile of the postdoctoral experience, international postdocs must be encouraged to complete the survey. It is important to stress that participation is voluntary and individual responses are confidential. PDAs have made special appeals to foreign scholar associations to help publicize their surveys. Printing promotional flyers in other languages also can boost response rates.
Identifying the survey population is absolutely necessary and is one of the most challenging aspects of a postdoc study. Without knowing the size of the overall postdoc population, it is impossible to state the response rate and thus to estimate the degree to which the responses are representative. PDAs generated a buzz about their survey effort by placing print copies in mailboxes, putting advertisements in newspapers, and hanging posters and flyers in labs. Electronic versions of the survey instrument--such as that made available at the Baylor College of Medicine Postdoc Association  Web site--appear to be the easiest to complete, increasing the possibility of a larger response rate. But to assure that the greatest number of eligible postdocs receive and complete the form, PDAs have found it necessary to distribute the survey through a number of means, primary among them e-mail and faculty advisers.
The Postdoc Network is in the process of developing a survey template with a set of standard questions for use in local surveys. Using standardized questions will permit the results to be compared nationally. In the interim, we encourage you to study other PDA surveys before conducting your own: Build upon these emerging promising practices!