Summary
Grading strategies applied in STEM courses often lead to lower grade point average (GPA) compared to non-STEM courses. This can result in students changing their major, choosing non-STEM courses as preference, or leaving STEM programs. The difference between students’ grades in STEM courses and non-STEM courses is called STEM penalty as student usually receives lower grades in STEM courses.
Three research questions were addressed in the paper:
1) What is the net attrition out of STEM? Is the exodus from traditional STEM majors counterbalanced by an influx from other majors?
2) Among successful (eventual graduates) and high‐performing students who enroll in STEM coursework, what are the patterns of relative grading as measured by one's STEM‐GPA and non‐STEM‐GPA across the first semesters in college? Do these averages vary by gender?
3) Is relative grading—a “STEM grading penalty”—negatively associated with one's chances of STEM retention over and above demographic factors, academic preparation, and early course selection?
The paper focuses on identifying relationship between grading patterns in the early semesters and resulting STEM attrition in the later semesters. The data consists of students’ interviews towards the end of their first, third, and sixth year of college together with other information from all attended colleges. The analysis show that STEM courses are graded tougher compared to non-STEM courses throughout all the semesters for both non-STEM and STEM graduate groups.

Relevance for Complex Systems Knowledge
The paper compares grading of STEM courses in relation to grading of non-STEM courses and students’ reasoning to choose STEM or non-STEM courses. The analysis was performed between the groups of students who have chosen their major in STEM disciplines against non-STEM disciplines. Students’ grades in STEM and non-STEM courses were compared throughout the semesters. This type of analysis helps to understand how performance in course contributes to leaving STEM education, especially for those who are interested in STEM disciplines.
The research data consists of students’ interviews and transcripts about students’ progress semester by semester. the data was collected for the study directed by National Center for Education Statistics. The data shows that approximately 9000 students registered to their first semester and only about 60 percent graduated within 6 years. Out of the graduates, about 20 percent graduated from the STEM programs, such as computer and information science, mathematics and statistics, engineering, physical sciences, biological and biomedical sciences. The logistic regression models were employed to test hypothesis about the multiple paths to STEM discipline, variation in balance between grade averages of STEM and non-STEM courses for students who take STEM courses early, negative influence on receiving lower relative grades and therefore leaving the program.
The results show that those students who take STEM courses during the first semesters and graduate in non-STEM programs, have a significantly higher STEM penalty in the first semesters in comparison to those who graduate in STEM programs. In later semester students from non-STEM programs take less or easier STEM courses, therefore the STEM penalty reduces. Moreover, students who graduated in STEM had higher scores in non-STEM courses than students who graduated in non-STEM field.
The study also compares STEM penalty with respect to gender. The multivariate analysis shows that although women get a higher STEM penalty, it does not affect their chance to graduate in a STEM field.
The authors state that lower grades of STEM courses result in students moving away from STEM fields. They recommend revising the grading process of STEM courses and consider shifting or expanding STEM grading distribution. This should help to give higher grades for the best performances and could result in higher graduation rate for STEM disciplines.

Point of Strength
The paper provides analysis on the grade difference between STEM and non-STEM courses. The authors state that there is STEM penalty – a lower grade point average than in non-STEM courses. The results show that students, who ultimately graduate in STEM, have an academically stronger record than non-STEM graduates in STEM and non-STEM courses.
The research suggests an approach to analyze group of “STEM-actives” as separate group. This group consists of students who take STEM activities (courses and credits) in the first semester. This group has a potential to become interested in STEM and change their major. Such changes could result in increased number of STEM graduates.

DOI
https://doi.org/10.1002/sce.21580

Keywords
college major, higher education, relative grading, retention, STEM