In-Class Participation and Performance Gender Gaps in Introductory Biology Courses: The Instructor Gender Effect
Faculty Mentor: Liz Bailey, Biology
It is well known that undergraduate women are underrepresented in STEM disciplines,
which includes the sciences, technology, engineering, and mathematics (NSF 20011). However,
the explanation for this phenomenon is not completely understood. Much research has been done
recently and suggests that the “why?” is very complicated and multi-faceted (Smith 2012, Gayles
2014). Reasons can include female students’ sense of belonging, perception of ability, access to a
positive mentor, and others (Smith 2012). While biology programs are unique in that female
enrollment often exceeds that of males, gender inequities are still seen as postgraduates enter
their careers (Handelsman et.al 2005). Because the discrepancies are not as obvious, differences
in male/female achievement in biology courses are not as widely researched as those in other
STEM disciplines. Because in-class participation has been shown to increase a student’s sense of
belonging in a class and improve their achievement, differences in male and female in-class
participation is also worth investigating (Eddy et.al 2014).
Thirty-four life sciences classes over five different semesters were observed three times
each at a large four-year university. To improve reliability and reduce gender bias in data
collection, one male and one female student researcher attended each observation. The student
researchers silently observed the class without interacting with any students or the instructor and
recorded the following data: the male/female ratio in attendance that day, the instructor’s gender,
the number of times the instructor invited class participation, all student participation events, and
the instructors’ responses to each participation event (positive, negative, or neutral). Participation
events, for this study, were defined and limited to the following categories: raising a hand;
voluntarily asking a question; voluntarily answering a question posed by the instructor;
voluntarily offering a comment; volunteering for an activity related to the subject material;
performing a task not related to the subject material; students calling out an answer, question, or
comment without raising their hand; being randomly selected by the instructor to ask a question,
answer a question, or comment; and being selected non-randomly by the instructor to ask a
question, answer a question, or comment. Observers determined the difference between the last
two categories (random versus non-random selection) by observing instructor behavior prior to
the event (such as using a random number generator or pulling a name from a stack). Student
researchers only recorded a participation event if it was loud enough for them to hear and
identify the speaker and if it was directed toward the instructor. If some of these events took
place as part of the same event (i.e. they raised their hand AND were called on to ask their
question), these were both recorded and indicated that they happened together. In this way it was
possible to determine both how many students raised their hands and how many were actually
called on to participate. To ensure that all data was collected consistently and accurately, all
student observers used the same standard form to collect the data. Gender gaps are calculated as percent participation attributed to a female student – percent attending students who are female.
Variables did not differ by semester except for class size, so class size is included as a covariate
in all analyses.
For each of the classes that was observed, achievement data was collected from the
registrar’s office in the form of de-identified final letter grades. Student gender and ACT score
(in an attempt to control for student ability) were also included in the report from the registrar’s
office. Average grades (0 – 4 grade point scale) were adjusted for ACT score. GPA gaps are
calculated as average adjusted female GPA – average adjusted
When all life sciences classes are averaged, female students
are less likely to participate than males but perform equally (Fig.
1). Further analyses found that that this difference was not derived
from instructors’ actions in the classroom (i.e. calling on male
student’s hands more often; data not shown).
Although instructors’ actions in the classroom did not
affect participation and GPA gaps, instructor gender did (Fig. 2).
We found that female students are more likely to participate in a
female instructor’s classroom when compared to a male
instructor’s classroom. We also found that female students
significantly underperform compared to their male counterparts in
a male instructor’s class, but there is no significant difference in
male and female performance when there is a female instructor.
Female students perform significantly better when their instructor
is female rather than male.
Although it is discouraging to find gaps
in participation and performance, interventions
can be developed in order to close these gaps.
Our data suggests that female instructors are
beneficial in closing these gaps, and hopefully
there will be an increase in female instructors in
the future. Female students may perform better
in a female instructor’s classroom because they
benefit from having a strong female rolemodel.
Male instructors might consider choosing
female teaching assistants or inviting female
guest speakers (at the beginning of or
throughout the semester) so female students feel
a greater sense of belonging. Development of
interventions to close these gender gaps is the
goal of our future research.