Much attention is fittingly paid to race- and income-based achievement gaps in K-12 schools. But research has also documented similar and worrying gender-based gaps in college classes on high-stakes science tests. Analysts have attributed these to variations of “student deficit,” such as unequal K–12 preparation for college-level science, and to “stereotype threat”—the idea that women are led to believe they don’t have the same ability as men to succeed in STEM fields and thus perform poorly at the most stressful moments, like when taking exams. If these gender gaps and their underlying causes remain unaddressed, important and lucrative job paths in fast-growing STEM fields could be closed to many women.
A recent study by Sehoya Cotner and Cissy J. Ballen of the University of Minnesota proffered a different theory for these observed gaps: a “course deficit” model, wherein course structure leads to performance gaps, specifically instances in which high-stakes midterms and finals are the main components of final grades.
Cotner and Ballen looked at nine high-enrollment introductory biology courses at an unnamed large public university with varying mixes of high-stakes and low-stakes assessments comprising their final grades. They analyzed summative course grades and performance on midterm, final, and various lower-stakes assessments as a function of gender and of incoming preparation as measured by composite ACT scores.
Their findings were just as they had predicted: As the percentage of overall grades determined by midterm and final exams increased, the performance gaps on those tests between female and male students increased, too; and, as that percentage decreased, the gaps disappeared and sometimes even reversed. They surmise that, “[F]or some individuals, performance on exams may not reflect a student's actual content knowledge.”
Cotner and Ballen also conducted three case studies to dig down into specific course-based variables. All three reinforced the general findings, leading them to conclude that to better support women pursuing degrees in STEM field, universities should offer more courses that deemphasize high-stakes exams and rely more on active learning techniques, such as group projects, low-stakes quizzes and assignments, class participation, and in-class activities such as labs.
Even before conducting this study, Cotner and Ballen were long-time proponents and researchers of active learning techniques and mixed assessments. And although their conclusions here are compatible with the data tested and with their previous research, they leave too many questions unanswered. Wouldn’t using ACT science scores instead of less specific composite scores better determine incoming students’ preparation? After noting that performance patterns appear to support stereotype threat, why not test for it? Why not also control for the instructor’s gender? They also boldly proclaim that “the lower-value exams assessed the same content knowledge as the high-value exams, a finding that should assuage concerns that low-stakes testing means a watering-down of expectations.” But if a test constitutes a smaller percentage of a student’s grade than a midterm or final, wouldn’t the less-weighted exam be shorter and less extensive? Cotner and Ballen should have at least investigated test rigor as a factor rather than assuming it was the same in all cases.
“Why so few women in science?” asked a 2010 report from the American Association of University Women (AAUW). Stereotype threat was at the top of its list, followed by gender bias. And today’s research continues to argue strongly that stereotype threat is still at play today in STEM fields. Any study looking for an alternate theory must therefore clear a higher bar. This study needs expanding and deepening to approach that bar.
Source: Sehoya Cotner and Cissy J. Ballen, “Can mixed assessment methods make biology classes more equitable?” PLOS ONE (December, 2017).