Gender differences in students’ self-efficacy in introductory physics courses in which women outnumber men predict their grade

Students’ self-efficacy in physics classes can play a key role in shaping course outcomes. Prior research has shown that women have a lower self-efficacy than men in calculus-based introductory physics courses. We administered a validated survey to 564 students to investigate the gender differences in self-efficacy and how it predicts student grades at the end of a mandatory two-semester introductory physics course sequence primarily for bioscience majors in which women outnumber men. We used structural equation modeling to investigate how students’ self-efficacy predicts female and male students’ grade at the end of the course. We found that women had a lower self-efficacy and grade than men and that the students’ self-efficacy played a major role in predicting students’ grade even though women outnumbered men in this physics course. This study hints at the fact that numerical representation alone is not sufficient, e.g., to mitigate the effects of deep-rooted societal stereotypes and biases related to who belongs in physics and can excel in it. Thus, it is important for physics instructors to be intentional about creating equitable and inclusive learning environments in which all students, particularly those from traditionally marginalized groups such as women, have high self-efficacy and can thrive.

Figure 1

Schematic representation of the model and how self-efficacy mediates the relation between gender and grade in the physics 2 course for bioscience majors in which women are not underrepresented. From left to right, all possible paths were considered (including the one from gender to grade in physics 2), however, only some of the paths are shown here for clarity.

Figure 2

Result of the path analysis part of the SEM between gender and physics 2 grade through students’ self-efficacy for model 1. The line thickness indicates the relative magnitude of $\beta$ values. The dashed lines indicate covariances between constructs. All $p$ values are $p<0.001$. Gender does not directly predict the physics grade.

Figure 3

Result of the path analysis part of the SEM without SAT math between gender and physics 2 grade through students’ self-efficacy for model 2. The line thickness indicates the relative magnitude of $\beta$ values. The dashed lines indicate covariances between constructs. $p$ values are indicated by no superscript for $p<0.001$ and superscript “a” for $p=0.008$. Gender does not directly predict the physics 2 grade.

Figure 4

Result of the path analysis part of the SEM without high school GPA between gender and physics 2 grade through students’ self-efficacy for model 3. The line thickness indicates the relative magnitude of $\beta$ values. The dashed lines indicate covariances between constructs. The $p$ values are indicated by no superscript for $p<0.001$ and superscript “a” for $p=0.002$. Gender does not directly predict the physics 2 grade.

Figure 5

Result of the path analysis part of the SEM without high school GPA and SAT math between gender and physics 2 grade through students’ self-efficacy for model 4. The line thickness indicates the relative magnitude of $\beta$ values. The dashed lines indicate covariances between constructs. $p$ values are indicated by no superscript for $p<0.001$ and superscript “a” for $p=0.007$.

Figure 6

Result of the path analysis part of the SEM without grade in physics 1 between gender and physics 2 grade through students’ self-efficacy for model 5. The line thickness indicates the relative magnitude of $\beta$ values. The dashed lines indicate covariances between constructs. All $p$ values are for $p<0.001$.