Doctoral Education and the Faculty of the Future

7. Undergraduate STEM Research Experiences: Impact on Student Interest in Doing Graduate Work in STEM Fields

Does participation in undergraduate research in science, technology, engineering, and mathematics (STEM) affect the likelihood that participants will enter STEM graduate programs and succeed? This is one of many questions addressed in the literature on the effects of undergraduate research (UGR). Typically such studies examine the impact of UGR on a number of factors: the completion of a bachelor’s degree in a STEM field; the probability of transitioning to graduate school after earning the bachelor’s degree; changes in the depth of commitment to or level of interest in continuing into graduate school (because many UGR participants have already planned to attend graduate school); the decision or plan to pursue a career in STEM, especially in research; the research skills learned (operating lab equipment, applying appropriate statistical analyses, reading and synthesizing prior research papers, finding key prior research papers in a library, working independently, writing in appropriate scientific styles); the ability to think like a scientist; and the acquisition of communication skills (presentation; working effectively with faculty, graduate students, and undergraduate peers on teams). Some studies also have considered the costs and benefits of UGR as well as its limits, such as the percentage of majors who can be effectively provided with quality UGR experiences given department size. 7 Undergraduate STEM Research Experiences Impact on Student Interest in Doing Graduate Work in STEM Fields Myles Boylan 110 Doctoral Education and the Faculty of the Future The Limitations of Empirical Studies The quality of the extant research is affected by the problem of a relatively heterogeneous treatment group, of finding high-quality, affordable assessment data, and of obtaining long-term data on actual decisions as undergraduates transition to postbaccalaureate studies and full-time employment. Short-term assessment studies rely heavily on survey questionnaires. Fortunately, some long-term data exists, and the majority of actual UGR experiences studied were of fairly high quality and effectively designed for their clients, undergraduate students. At many colleges offering UGR experiences, the students selected to participate were already high-achieving and oriented toward graduate school. For example, this is typically the case for National Science Foundation Research Experiences for Undergraduates (REU) projects. Only a few studies look at the impact of UGR on average or typical STEM students. Broadening the evident range of student capability would be a more useful and powerful test of the impact of UGR on student achievement and continuation into graduate programs. The lack of a control group matched to the treatment group forces most studies to use a pre- and postexperience evaluation of UGR experiences , relying on student and faculty mentor responses to survey questions . In addition, most studies do not actually observe student choices long enough to determine the possible effects of UGR directly. Direct observation avoids the problem that students’ plans to attend graduate programs and their career intentions will not actually be realized in a number of cases. We know from data provided by decades of surveys conducted by the Higher Education Research Institute at the University of California–Los Angeles and published in the annual American Freshman: National Norms that the formative intentions of first-year students for attending graduate programs—particularly doctoral study— are quite optimistic when compared to what actually happens years later. (This indicates the value of surveying the depth of student commitment to graduate school or embarking on a science and engineering career.) Larger-scale studies such as program evaluations have incomplete controls and data on the nature of the UGR experience itself, other than some built-in quality-control procedures (such as peer review in the case of REU awards). However, while the experiences and the disciplines covered do vary, most experiences covered by published studies appear to be carefully constructed. Virtually all of the data are survey based. When questions are not about factual events, student and faculty respondents code their answers using 5-point (or 4-point) Likert scales. Item construction can be critical in this context. The difference between a 5 and a 4 on a 5-point scale can Undergraduate STEM Research Experiences 111 be influenced by the way these numbers are represented verbally, and there is little hope that the studies in this review are fully comparable. The mix of undergraduate disciplines seems to matter a great deal. In the absence of UGR opportunities, bachelor’s graduates in mathematics are much less likely to attend graduate school in the near future...