This year, a group of faculty members (and one soon-to-be faculty member) from multiple departments at Vanderbilt are joining together to consider how to promote persistence in STEM at Vanderbilt. We plan to use Talking about Leaving Revisited (TALR) as a guiding text, and we have identified several goals:
- Develop a more holistic and research-based understanding of the reasons that students leave STEM
- Tease out the factors that are common across STEM as well as those that are more discipline-specific
- Develop a common vocabulary for talking about the educational practices that impact student success
- Identify steps we can implement in our own classes to promote a better student experience
- Identify non-classroom experiences that can influence students’ educational experience (e.g., undergraduate research, access to timely support) and related policy recommendations
In our second meeting, we focused on reviewing Chapter 2 in TALR to understand the trends observed in that study, consider the degree to which similar trends occur at our institution, and identify questions and actions they suggest for us.
The chapter begins with very good news: data from the Higher Education Research Institute that represents a very large national pool of students, persistence in STEM has improved significantly since 1991. Specifically, TALR says that “switching from STEM majors into non-STEM majors declined dramatically from 44.1% to 28.1%….persistence in original STEM majors improved by 13.9% from 46.% to 59.9%….for women, persistence increased substantially by 23.6% from 37% to 61.3%.” The tables below show rates of persistence in several STEM disciplines.
This is, of course, great news. We speculate that this change may be the result of faculty adopting evidence-based teaching practices, such as increasing course structure and incorporating active learning. As LC member Richard Haglund noted, we also know that some students should switch from their intended major as they encounter the smorgasbord of options that college offers and find options more aligned with their goals and interests, and several faculty members wondered whether this level of switching is persistent across majors and thus is just something that we should expect for students who are new to college.
As the chapter goes on, however, the findings become less rosy. While national persistence in STEM majors has increased, the study of the six institutions in TALR indicates that students in different groups do not leave at the same rates. Specifically, women leave STEM majors at great rates than men at all six institutions studied, and underrepresented students leave at greater rates than well-represented students at three of the six—and have numbers too low to measures at a fourth (figures below).
Low math scores predict switching. Women switch at greater rates than men across all math performance levels in both underrepresented and well-represented groups, and underrepresented students switch at higher rates than well-represented students across all math performance levels. In other words, low performance on standardized math tests predicts switching—but performing well is not fully protective if you are underrepresented or a woman.
First-semester GPA also appeared to have a different impact on different groups. Mirroring the observations for math scores, women switch at greater rates than men across all GPAs in both underrepresented and well-represented groups, and underrepresented students switch at higher rates than well-represented students across all GPAs. Again, low first semester GPA predicts switching—but performing well is not fully protective if you are underrepresented or a woman.
Most students switch by the third term, with 50% of students who switching by the end of the first year. This suggests that we need to focus our efforts on students’ very earliest experiences in STEM courses—an observation consistent with our own experience and other reading.
The group discussed both why we observe these persistent patterns that suggest that women and underrepresented students switch at a greater rate even when GPA and math scores suggest they should be, and are, succeeding. Is it due to deciding that the path is not attractive, and if so, why? Could it have to do with a fixed mindset, where even a relatively small challenge to one’s self-identification as a “good” student could be offputting—but if so, why would we see a difference between groups in this parameter? Could it have to do with Claude Steele’s idea that some students have a larger pool of good experiences to draw from to bolster confidence? Guil Gualda noted that these patterns suggest that our response has to be a more concerted institutional effort rather than one based just on individual faculty actions.
We also discussed a few ways to help students when they are facing what may be their first big academic challenges in college. One idea, suggested by Kathy Friedman from Biological Sciences several years ago, is the development of STEM study center where students gather for study groups, tutoring, office hours, and informal interactions. A related idea that seems particularly suited to the Covid area was suggested by Adriane Seiffert: Develop a virtual help desk for different disciplines, where students could log in to a Zoom session at a variety of times to get help in different classes within that discipline. Physics has a help desk that works this way, and it is typically staffed by Physics majors who know that teaching others is a great way to prepare for the GRE; the School of Engineering has something similar.
We also discussed the importance of normalizing “failure” and seeing academic challenges as a normal part of college life, and considered the possibility of having panel discussions for first-year students, where both faculty and more senior students would talk about their academic challenges and how they addressed them. This too seems as though it would work relatively well in a socially-distanced time. The group discussed that a key element is helping students develop a larger perspective, helping them see beyond a single test or even course to a larger picture. We acknowledge, though, that this is hard both because first-year students at Vanderbilt are quite young, and because we have selected a group of students who are very focused on academic success.
Finally, we discussed the importance of students seeing a diverse group of scientists represented in their courses. Adriane Seiffert does this by varying her homework assignments to represent a broad range of people (not unlike Jeff Schinske’s Scientist Spotlight homework assignments), while Lars Plate does so by highlighting scientists who have been instrumental to the work that they are discussing in class.
We are eager to dig into the qualitative findings from the study to flesh out the quantitative patterns detailed in this chapter. We are going to read Chapter 3: Why Students Leave STEM Majors for an overview for our next meeting, but will then turn to Chapter 8, which describes students’ perceptions of good STEM teaching.