Summary
Self-efficacy beliefs are essential for developing and employing academic competencies. They help to set and achieve goals, increase motivation. The study was dedicated to identifying the differences of self-efficacy in STEM fields based on gender, domain and course time. The research questions of the study were:
RQ1) Does student self-efficacy vary by STEM academic domain? If so, does the variation depend on gender or subject?
RQ2) Does current class self-efficacy and STEM self-efficacy change over the course of two semesters of physics classes? If so, is this change moderated by gender or academic performance?
RQ3) Is the gender difference in current class self-efficacy observed in Study 1 present very early in the class?
RQ4) Does current class self-efficacy change as a result of class feedback? If so, is this change moderated by gender?
The students from courses of the STEM disciplines in large eastern US university were included in the research. The self-efficacy questionnaire was based on the Bandura’s social cognitive theory. The study identified three tiers of STEM self-efficacy: current class self-efficacy, STEM self-efficacy, professional self-efficacy. The results show that neither type of self-efficacy changes or its change is moderated by gender or academic performance in STEM classes, although the change was noticed for the physics class.

Relevance for Complex Systems Knowledge
There are many definitions of self-efficacy, but in general it can be determined as capability to execute actions that lead to a designated goal. The previous research show that men tend to express higher self-efficacy towards the generally male-dominated areas, such as mathematics, physics, and engineering. Thus, the aim of this research was to explore self-efficacy across multiple STEM domains and describe its development in general and with respect to gender.
The Bandura’s social cognitive theory was employed in the theoretical framework of the research. The study was conducted in the eastern USA university. Approximately 3000 students participated in the study (RQ1 and RQ2) related to the students’ self-efficacy based on gender and subject. Another study (RQ3 and RQ4) was conducted with approximately 900 students in multiple time points during the course. The majority of students were male (~75%) for both datasets. The self-efficacy was discussed in three levels:
- STEM self-efficacy: general expectations to performed in STEM classes formed by the previous academic experience in college or earlier.
- Current class self-efficacy: self-efficacy beliefs toward the specific formed by new STEM academic experience.
- Professional self-efficacy: functioning in student’s planned profession.
The survey was designed as a questionnaire with 5-point Likert scale (“strongly agree” -> “strongly disagree”) with modifications specific to the discipline. The data about students’ gender, race/ethnicity, and college grade point average were also included in the research. Linear mixed effects modelling was applied in the analysis to compensate the imbalanced samples size caused by a small number of female representatives in math and physics classes.
The analysis showed the highest self-efficacy towards the students’ intended profession and the lowest to the current class. STEM self-efficacy did not differ significantly based on gender, but women expressed lower self-efficacy in current classes of mathematics and science. The time (early-stage vs late-stage in the course) was not significant to self-efficacy although slight changes related to the students’ test average were identified.
The authors conclude that the student start the STEM courses with self-efficacy based on their previous experience outside the classes and the analyzed courses do not produce self-efficacy differences. As the self-efficacy level changes throughout the course, the self-efficacy differences remain approximately the same. Therefore, the courses should institute policies to reduce the existing differences.

Point of Strength
The study identifies self-efficacy levels and performs research on how they are related to academic domain and gender, and the dynamics of self-efficacy levels in multiple time points during the course. The analysis showed that self-efficacy differences show very early in class and do not change significantly during the STEM course. Thus, the instructional approaches should pay more attention in reducing the existing differences.

DOI
https://doi.org/10.1007/s10956-020-09853-5

Keywords
Gender, STEM course self-efficacy, Impact of feedback, Introductory math and physics