The impact of gender roles and previous exposure on major choice, perceived competence, and belonging: a qualitative study of students in computer science and bioinformatics classes

ABSTRACT

Background and Context

While biology has strong female representation, computer science is the least gender equitable of the STEM fields. A better understanding of the barriers that keep women out of computational fields will help overcome those barriers to create a more diverse workforce.

Objective

We investigated the complexities that influence students’ major choice and their sense of belonging. We were particularly interested in students adjacent to computer science and differences by gender.

Method

We conducted semi-structured interviews of nineteen students from computer science or bioinformatics courses. We used inductive thematic analysis that included iterative readings of interview transcripts, line-by-line coding, and final theme selection.

Findings

Most students described pressures stemming from traditional gender roles as they chose their major, but specific pressures differed by gender. Men were more likely to report early exposure to their major field, and women noted feeling behind without those early experiences. This hurt the women’s sense of belonging, as did having few female peers.

Implications

Increasing early exposure to coding may increase women's representation and sense of belonging in computational fields. Women from contexts with emphasis on traditional gender roles may be drawn to computational fields if they see opportunities for flexible positions.

KEYWORDS:

• Women
• female
• career development
• quantitative biology
• social cognitive career theory
• conservative religious populations

Introduction

Representation of women in computer science and bioinformatics

The computer sciences continue to be the least gender equitable of the science, technology, engineering, and mathematics (STEM) fields according to the National Science Foundation, with only ~20% of computer science bachelor’s degrees earned by women in 2018 (National Science Foundation, 2021). Furthermore, rather than improving, the representation of women in the computer sciences has been declining during the 21^st century (Fry et al., 2021; Sax et al., 2017). The United States recognizes this as a barrier to reaching our innovative capacity and competing on a global scale (Beede et al., 2011). Underrepresentation of women in computer science occurs outside of the United States as well and is thought to be caused by complex issues at multiple levels, from individual experiences and access issues up to broad cultural norms (Michell et al., 2018; Stoilescu & McDougall, 2011).

Bioinformatics, as a younger field described as a hybrid between computer science and the more gender equitable field of biology, has the potential to draw more women into computational fields. This is seen even in its origin, as one of the first founders of bioinformatics, Margaret Belle Dayhoff, was a woman (Masic, 2016). In terms of authorship, the field of bioinformatics is indeed intermediate in terms of gender gaps, with fewer female authors than biology publications but more female authors than in computer science (Bonham et al., 2017). Less research has been done on gender dynamics in bioinformatics education specifically, but more attention to this hybrid field has the potential to increase our understanding of reasons why women are attracted to specific STEM majors.

Previous research on major choice and gender

Numerous studies have investigated the many factors that could influence students’ choice of college major and why some majors are more gender equitable than others. First, students must be aware that a major exists in order for them to choose that major, and one study even found that writing a paper about a specific major increased the likelihood they would pick that major (Fricke et al., 2018). While differences in awareness can help explain some types of ethnic segregation among majors, differences in awareness have not been able to explain gender segregation in STEM majors (Baker & Orona, 2020). Second, students’ interests draw them to specific majors. This may help explain women’s underrepresentation in some fields, as studies have found women to express less interest in STEM fields such as engineering compared to health and medicine (Sadler et al., 2012; Weber, 2012). Interest alone cannot explain gender segregation, however, as men and women often choose different majors even when they report similar preferences (Quadlin, 2020). With equivalent preferences, potential for a high-earning career may attract men to specific fields more than women (Quadlin, 2020), and women may be drawn to careers that focus on achieving communal goals (Diekman et al., 2010). Thus, interests are complicated and can be highly influenced by students’ experiences, successes, and cultural influences.

For example, students’ academic achievement and self-efficacy play a role in the majors in which they show interest and ultimately choose. Tellhed et al. (2017) found that differences in self-efficacy and confidence with STEM fields by gender explained women’s lower interest in STEM. Gender differences in math self-efficacy have been consistently related to gender differences in STEM field selection, but this has become a weaker explanation for gender segregation in majors in more recent years (Sax et al., 2015). Furthermore, numerous quantitative studies have found that gender is still a strong predictor of major choice even after accounting for variables such as Scholarship Aptitude Test scores, math/science achievement, student competency beliefs, and interest in math and science (Dickson, 2010; Evans et al., 2020; Tan et al., 2021; Vu et al., 2019). Ceci and Williams (2010) argue that small differences in spatial and mathematical abilities cannot explain the underrepresentation of women in math-intensive fields and are only secondary factors. Rather, women’s preferences and choices, both freely made and constrained by biology and society, are likely the primary drivers of gender differences in STEM representation (Ceci & Williams, 2010; Ceci et al., 2009).

Many of the gender differences in interest and competency beliefs may stem from social stereotypes about STEM fields. One study found that women feel less competent in computer science after communicating with stereotypical computer science students (Cheryan et al., 2011). In another study, female college students were more likely to describe prototypical computer scientists using traits consistent with stereotypes and rate themselves as different than those stereotypes than were male students. This difference from stereotypes mediated the effect of women reporting less interest in computer science (Ehrlinger et al., 2018). Parents may also have implicit biases about their children’s efficacy and interest due to cultural stereotypes. For example, Tenenbaum and Leaper (2003) found that parents were more likely to think that science would be interesting to their sons compared to daughters, and they also believed science would be more difficult for their daughters. Furthermore, those parental beliefs were predictive of their children’s actual self-efficacy and interest toward science. That parental support, especially from the mother, has been found to be important for women’s pursuit of STEM degrees (Beishline, 2008; Rinn et al., 2013).

Students’ sense of social belonging also impacts their major choice and can also explain gender differences (Tellhed et al., 2017). Rainey et al. (2018) found that white men find the most belonging in STEM majors and minority females find the least belonging, and this belonging is a predictor of retention. Interpersonal relationships, perceived competence, personal interest, science identity, and perception of the social climate of the institution all contribute to a student’s sense of belonging (Johnson, 2012; Rainey et al., 2018). Cohoon (2001) argued that university department characteristics have a large impact on retention of women in computer science, and women can thrive under supportive social conditions.

Unique considerations for religiously conservative populations

As decisions about college majors and future careers are complex, as are the ways in which gender impacts these decisions, qualitative studies are helpful to provide deeper understanding of how students from specific contexts weigh various experiences, interests, and priorities. Students’ religious background could be an important cultural lens through which students see their gender and its associated role in society. Cultural norms can play a large role in students’ decision making, and conformity to cultural norms about femininity (I.e. being relationship-oriented, caring for children, being domestic) has been shown to negatively predict choosing a STEM major (Beutel et al., 2018).

We conducted this research at a religious institution in Utah associated with the Church of Jesus Christ of Latter-day Saints, and most students at this institution are active members of the sponsoring church. The literature suggests that this population adheres more strictly to traditional gender roles with a male provider and female nurturer in the home compared to other populations in the United States, both historically and currently (Halford, 2021). For example, a study conducted by Jensen and Jensen (Jensen & Jensen, 1993) found that high religiosity was predictive of valuing the traditional female role in the home, and members of the Church of Jesus Christ of Latter-day Saints were more likely to value traditional roles than Catholics or Protestants.

The literature also suggests that this focus on traditional gender roles has an impact on students’ choices of a field, major, or career. In some populations, female students are less likely to choose a male-dominated STEM discipline if they attend a religious school or have a mother who holds more traditional gender ideologies (Rich & Golan, 1992; Steele & Barling, 1996). Crawford (1978) also found that females with more rigid views about gender roles were less likely to choose more male-dominated professions. One study found that men, but not women, in the Church of Jesus Christ of Latter-day Saints perceive that family cohesion decreases when women work outside the home, and family cohesion is a strong priority in this cultural context (Pedersen & Minnotte, 2008). Individuals can be impacted by surrounding culture even when they are not personally religious. For example, Moore and Vanneman (2003) found that Utah is one of the most conservative states regarding gender attitudes and that these attitudes were even held among those that did not personally share more conservative religious beliefs.

Study population and research questions

While countless studies have investigated the potential causes of women’s underrepresentation in specific college majors, this phenomenon is still not completely understood. This is likely because this complex decision is influenced by each student’s unique personal experiences framed in their broader sociocultural context, which can vary for different groups. Thus, we set out to qualitatively study a small sample of religious students at a conservative, private university to understand their experiences more deeply. We were specifically interested in students both in and adjacent to computational majors such as computer science or bioinformatics, so we recruited students from bioinformatics or computer science courses. We were also particularly interested in differences in the experiences of men and women in this specific population. Thus, we focused on the following research questions:

1. What factors influence students’ initial and final undergraduate major choices in a religiously conservative population? Are certain factors more influential for men or women in this population?

2. What makes these men and women feel like they belong in their classes and their major?

Methods

Ethics statement

This study was approved by the Brigham Young University Institutional Review Board (Protocol E2020-069). All participants gave consent prior to being interviewed and were told they could withdraw consent at any time. Each participant was also compensated for their time with a $20 gift card or $20 on their student card.

Research team

The principal investigator is a white woman, PhD holder, assistant professor of biology, and researcher who studies women’s experiences in STEM. The interviewer is a white woman who was an undergraduate at the time of the study (majoring in environmental science), and she is currently a PhD student. We chose an undergraduate interviewer in hopes that students would be more comfortable and truthful in describing their classroom experiences with a peer. The interviewer was trained in interviewing procedures using Kvale and Brinkmann’s Learning the Craft of Qualitative Research Interviewing (2009). The interviewer then conducted seven pilot interviews (not included in the study). These pilot interviews were recorded and then reviewed with the principal investigator. The two researchers would discuss weaknesses in the interview, and the interviewer would make changes before the next pilot interview. After seven pilot interviews, we determined the interviewer was ready to collect data.

The rest of the research team was composed of a group of undergraduate students: two white women (a public health major and a biology education major) and one white man (a bioinformatics major). These researchers took part in qualitative data analysis and manuscript preparation. The entire research team are all members of the Church of Jesus Christ of Latter-day Saints, the sponsoring church of the university in which this study was conducted. Thus, the research team share a common religious and cultural context with the study participants.

Participants

We recruited students from bioinformatics and computer science classes at a large private university. As the university is sponsored by the Church of Jesus Christ of Latter-day Saints, this student population is generally highly religious. We emailed instructors of bioinformatics or computer science classes required for the bioinformatics major. By recruiting from computer science and bioinformatics courses rather than declared computer science and bioinformatics majors, we aimed to recruit students who were adjacent to these fields but potentially had a variety of declared majors. Since the courses we pulled from were not general education classes, the students in the courses had to have at least some interest in computer science or bioinformatics and likely had at least considered these majors.

Because performance in major classes and experiences in the classroom are likely to impact whether students choose a specific major, we wanted to get a variety of students with different experiences. Thus, we asked the instructors to send us the contact information for students who fit in one of four categories: (1) actively involved in class and performed well, (2) actively involved in class but performed poorly, (3) quiet in class but performed well, and (4) quiet in class and performed poorly. From an initial list of 43 women and 25 men from all categories, ten women (23%) and nine men (36%) responded that they would like to participate. Due to privacy considerations, the instructors did not say which students on the list came from which categories. Based on the interview content, we believe most of our participants were high performers except for a few men and women who talked about performing poorly. This suggests that low performers were less likely to agree to be interviewed despite our attempts at a diverse pool. We did seem to get a good mix of high- versus low-participators. Some important characteristics of each of our interviewees are shown in Table 1. We will use pseudonyms throughout the study to protect the identities of participants. The gender noted in the table is based on biological sex listed in official academic records. While we did not specifically ask students their gender, the way participants spoke of their gender (e.g. when talking about their similarities or differences to their peers, when speaking of family roles) matched what is shown in Table 1.

Table 1. Characteristics of interviewees.

Pseudo-nym	Age	Gender	Years in College	Previous majors (if applicable)	Major when interviewed	Stated Drive to Switch Majors
Shannon	25	F	5	Music, Neuroscience	Molecular Biology	Interests
Sandy	23	F	2	Physiology, Dietetics	Molecular Biology	Interests
Jeanette	22	F	2.5	none	Genetics and Biotechnology
Olivia	22	F	2.5	Physiology, Molecular Biology	Bioinformatics	Wanted more of a challenge
Leila	18	F	1	none	Computer Science
Anna	21	F	3	Psychology	Spanish Translation (Computer Science Minor)	Interests
Melanie	23	F	3.5	none	Molecular Biology
Isabella	22	F	3	Mathematics	Computer Science	Interests
Macey	20	F	3	Computer Science	Editing and Publishing	Interests
Julia	21	F	1.5	none	Computer Science (Bioinformatics Emphasis)
Dave	22	M	2	Computer Science	Pre-Business	Felt lacked competency
Paul	23	M	3	none	Computer Science
Derek	21	M	2	Pre-Mechanical Engineering	Bioinformatics	Interests
Jack	24	M	4	Pre-Business	Bioinformatics	Interests
Seth	24	M	4	Mathematics	Bioinformatics	Interests
Walter	20	M	1	none	Electrical Engineering
Stanley	21	M	1	none	Computer Engineering
Graham	19	M	1	none	Computer Science
Kevin	21	M	3	none	Animation Major

Interviews

We did not inform participants of our research questions or the purpose of our study until after the interviews. None of the participants knew the interviewer prior to the interview. One student vaguely recognized the interviewer from an adjacent lab space. Participants may have learned basic information about the interviewer (e.g. age, major) during the interview. The interviewer had an initial list of interview questions (see Supplemental Materials) but was allowed to ask the participants to expand on their answers or go in a different order as the interview progressed. Thus, the interviews were semi-structured, and not all participants were asked the exact same questions as each interview was tailored to understand each individual student. Initial questions were inspired by prior qualitative studies aimed at understanding undergraduate student’s belonging, performance, and participation (Kraft, 1991; Salter & Persaud, 2003). The interviewer aimed to have the students describe their process of choosing a major, any changes to that decision, and why they chose their ultimate major over other majors they were considering. This study analyzed only the portions of these interviews that related to students’ selection of their major, any changes they made to their major, and their sense of belonging in that major.

Each interview took between 40–50 minutes either in person (eight students) or over Zoom (11 students), with the interviewer always in a private room. After each interview, the interviewer took brief field notes of their general impressions and the interview atmosphere. Interviews were audio recorded then transcribed for analysis. Participants were not asked to comment on the transcriptions after they were complete.

Initial stages of analysis included investigating whether new ideas were continually being raised in the interviews or if we had reached data saturation. While there is no hard rule as to what constitutes data saturation, we felt that common themes were being raised consistently by both the men and women. Thus, we decided not to recruit more participants. In the discussion, we acknowledge the limitations due to this small sample size and our inability to generalize our findings broadly. However, we determined that these 19 participants were sufficient to give us a deep qualitative look at some of students’ experiences with majors related to bioinformatics and computer science.

Analysis

Framework – Because we wondered if our study population would differ from previously considered populations in terms of major choice, due to our conservative, religious context, we chose a primarily inductive qualitative approach when analyzing the data. Thus, we used inductive thematic analysis without a priori hypotheses or pre-existing theoretical frameworks in mind as much as possible (Braun & Clarke, 2006). As the entire research team shared the social context of the participants, we acknowledge that we certainly had pre-existing ideas of ways our social context might affect the major choices of men and women, but we strived for a data-driven analysis. Our thematic analysis method was contextualist in terms of epistemology, meaning that we drew from both essentialism and social constructionism. Thus, we attempted to both describe participants’ experiences, realities, and the meaning they drew from them, as well as investigate how those experiences are impacted and shaped by the broader social context (Braun & Clarke, 2006). Our analysis was generally semantic in nature as our initial codes (detailed below) were more descriptive, and we moved into interpretation as we investigated patterns and themes emerging from those codes.

Process – We generally followed the main steps outlined for common thematic analyses (Braun & Clarke, 2006) as follows (see, Figure 1). At the beginning of the research process, the research team listened to and read transcripts of all the interviews multiple times, jotting down initial impressions. Then, as a research team, we discussed interesting ideas across the entire dataset and created a codebook to guide a more systematic analysis. These codes were sometimes nested (e.g. time included daughter codes of childhood, high school, college, and future; people included daughter codes of family members, peers, and mentors) and were generally semantic in nature. Next, teams of two researchers (authors AAB and CHH for the women’s interviews and authors AAB and CS for the men’s interviews) systematically coded all interviews line by line using NVivo software (version 12 for Windows). Researcher pairs would then discuss their coding, coming to consensus on differences. This was somewhat of a cyclical process, as some codes were modified or added during this phase of analysis as researchers found new ideas not captured by our initial codebook. After this systematic coding was complete, we collated data relevant to each code and cross-queried different codes to discover larger themes. Again, this was a cyclical process as the research team would meet, discuss ideas of themes, go back to the interviews and look at quotes in groups to validate themes. We then moved to more clearly defining and naming the themes, again sometimes causing us to go back to the data, question our own biases, and refine the themes. As the interviews provided very rich data, it was impossible to discuss all interesting ideas in this manuscript. Thus, we focused on themes that were prevalent (majority of participants of at least one gender) and showed general differences between men and women. We also note exceptions in the text to demonstrate that our participants are not a monolith. Next, we selected example quotes as evidence of our themes straight from the data, but these quotations are not exhaustive representations. Finally, we summarized the themes in writing and compared our themes to those found elsewhere in the literature to position our findings within the broader context.

Figure 1. Qualitative data analysis process. Arrows that point back to the same step indicate that this step was done multiple times as a cyclical process.

Results

Traditional gender roles can put pressure on both genders for different reasons as they choose a major

First, we saw a common theme from both men and women that considering family roles was an important factor that influenced their major choice, and both genders commonly talked about their choice fitting in with their family roles when planning for future careers. Several of the men said that they wanted to make sure that whatever their career was, that they had enough time and money for their families. Interestingly, that balance seemed to be more important than the raw amount of money earned in a career for the students in our sample. Usually, students only mentioned the salary in terms of whether it was “enough” money. A good example of this balance between providing for a family and having time for family was Paul when he said,

The only things I was taking into consideration was, number one, is it a career that can– that I can take care of my family on? And number two, is it a career that’s going to be detrimental to me being a father in my family? And so, I don’t feel like that limits too much what your career can be as long as you have something that’s going to pay the bills and it’s not requiring you to be gone an inordinate amount of time and things like that. And so that was– those were really the only main considerations as far as lifestyle goes. (Paul)

Similarly, almost every woman described their future family as a high priority when making decisions about their studies. Most of the women wanted a career that had flexibility so that they could still spend most of their time with their future children and that paid enough so that, if necessary, they could be the main providers. For example, Olivia said,

So, in deciding on bioinformatics and switching I was really thinking about just like a future lifestyle, future family, and I’d really like to be a mom, take care of kids. It’s something that (inaudible) still in the biology field that it’s easier to do part time and still make enough money to support a family if necessary. (Olivia)

While a desire to put family first was common among both men and women, the specific pressures they felt regarding family roles and future careers differed. We noted many comments that implied interviewees felt pressured from external sources to initially choose or change their educational or career path. First, we saw that men commonly felt pressure to choose a more practical career that provided enough money to financially support a family, even if they had strong interest in something else. Dave showed this when he said,

I’m a musician right. So, I mean, but I never considered going into music because it’s just I mean, no offense to anyone who thinks otherwise, but I just feel like it’s not a very good way to provide for a family. It’s not a very steady career. (Dave)

Women, however, did not describe feeling this same pressure to find a more practical career, which gave them more of a freedom to choose a major that interested them more. We found this sentiment in Isabelle’s interview when she said,

It’s nice being like a girl with like this lifestyle, like the Church of Jesus Christ of Latter-day Saints. Or like motherhood is really my end goal. So, if I’m going to pursue a degree I don’t have to worry as much about things like financial and like other obligations, and so I have to enjoy what I’m studying which is an awesome thing. (Isabella)

Even though the women did not often feel pressured to find an extremely practical career like the men, this did not mean that the women did not feel any pressure when choosing a major. We found that most the women felt pressure to choose a major that allowed them the flexibility to be a stay-at-home mom in the future. Many women spoke of wanting a career that could be part-time or based at home. We see this in Melanie’s decision to go to PA school over med school. Although she liked both, she ultimately chose PA school because it allowed her flexibility in work style so that she could someday have a family.

That was actually part of the reason I liked P.A. School over med school as well. It’s just that, as you know, especially as a female yeah, and I’m now married. And you know, you have to keep in mind the idea that at some point there’s going to be kids, and it’s never gonna be convenient as far as a career goes it’s never gonna be convenient to have kids, but you have to find something where you can work with both of those … And again, that’s another thing that med school kind of rules out. It’s like if you, if I want to go into med school from what I’ve heard, it’s gotta be all in. And I don’t want to spend the next eight years waiting to have kids you know. So that’s probably the biggest lifestyle choice or lifestyle factor that influenced my choice. (Melanie)

Another interesting theme that we found in the women’s interviews was that many comments women made pertaining to future career involved some form of the phrase “if I need to”. It was used in the interviews to describe wanting financial and job security if their plan A, implying motherhood, somehow did not work out as they had planned. As an example, Olivia said,

Well just specifically is to hopefully be able to be a mom. Work part time potentially if it’s necessary and make enough money to support a family if that was ever needed. That’s not the ideal plan but I’d like to do that if necessary. (Olivia)

Here we can see that Olivia aspires to be a mother. She wants to be able to make enough money to support a family, but we can clearly see that she adds two phrases at the end, “if that was ever needed” and “if necessary”, that give the impression that she would only do it if family life did not work out as planned.

While the different pressures faced by men and women due to traditional gender roles described above were certainly very common in our interviews, we also want to acknowledge that not all men and women described this same perspective. For example, some of the interviewees talked about gender roles in a different way, emphasizing the idea of men and women sharing equal roles. Shannon demonstrated this when she said:

I wanted to yeah have like an equal partnership like to feel like I was an equal with my like future spouse. And so, I made a conscious choice to pursue something that would lead to a career and would allow me if necessary to provide for myself like the rest my life. You know if that was what the circumstance was. (Shannon)

Interestingly, Shannon still used the phrase “if necessary” when talking about her future career even though she was the woman who demonstrated the least reliance on traditional gender roles in the home. Similarly, Kevin was the man in our interviews who seemed to rebel against the societal pressure to pick the more practical job and went with animation due to his interests.

Men talked about getting exposed to coding or other aspects of STEM earlier and more frequently than women

The next major theme that we noted was the idea of previous exposure leading to their specific major choice. Every man we interviewed referred to some type of previous exposure in high school or earlier that led to them being interested in their chosen undergraduate major. These previous experiences were generally more than just knowing someone in the field; rather, the men often talked about influential personal experience with the field. This seemed to be especially important for those considering computer science and bioinformatics majors. For example, some of the men talked about having coding or computer experience from childhood or high school:

I was really good at web design, graphic design, and film making. And I had done a little bit of coding, like really basic level, and I had enjoyed it in high school. (Dave)

I’ve spent a lot of time around computers and technology. (Walter)

I worked a programming job in high school over the summer once. And I just sat at a desk doing my little intern stuff for eight hours a day for nine weeks, and I didn’t like that. (Derek)

Derek also talked about a high school concurrent enrollment anatomy course that drove him to the life sciences. Even though his initial programming experience was negative, he later combined his interest in the life sciences with his coding experience to ultimately chose bioinformatics as a major.

Paul did not explicitly talk about personal experience with coding, but he had a role model in his father which he felt helped him know what to expect:

My dad is a software developer, and so I’ve been around it my whole life and it’s always something that was kind of interesting to me … I think in some ways, because I had exposure to it early, I was coming into college knowing what to expect. And so, the initial decision … was probably a little more firm than a lot of freshmen coming into college because I’d already experienced, I’d seen it. I knew what it is about. I’ve been able to talk to a lot of people. (Paul)

Similarly, Jack’s previous exposure in high school piqued an interest that led to his initial major in Pre-Business. He said,

I guess like in high school and stuff, I looked at stocks a lot, and that’s kind of like– it got me into like, you know, business and stuff. And I really like looking at a charge from like chart patterns. I mean there is like this whole field of like– of like chart analysis where you can like look at stuff in a way like buying and selling pressures and stuff like that. (Jack)

Seth talked about how he was influenced in two different directions by previous experiences before college. The more impactful experience took over and led to his desire to attend medical school. His current major, bioinformatics, aligns better with his medical school plans. He recalled,

So my dad worked on Wall Street for a while. He was a statistician. And so that’s what I wanted to do: use math, do quantitative finance. But then in high school … my grandpa passed away from cancer of a brain tumor, and my brother had a brain tumor that stopped growing. And seeing the same disease with two completely different outcomes really intrigued me. And so, I think first it was curiosity, trying to understand more why, you know, how cancer works, why it works that way, and also to understand better why treatments weren’t available in my brother’s case. (Seth)

Finally, Kevin talked about participating in a specific high school program that allowed him to get his associates in high school. Math and science were prioritized, and he talks about being good at it but not enjoying it. So in Kevin’s case, he had previous exposure but went with his interest in animation when choosing his major.

On the other hand, only three women talked about previous exposure, and these were always brief examples of knowing someone who majored or worked in that same field rather than having personal experiences. The woman who had the most influential prior exposure was Macey, who said,

I started off with computer science, and a big reason I chose computer science was because, you know, my dad was a software engineer, my brother was a software engineer. We all have like really similar personalities, so I just kind of felt like, okay I’m going to be a software engineer. (Macey)

The other two women who brought up previous exposure briefly commented that their parents had that same major or career. Another woman, Julia, actually specifically talked about how neither she nor her female friends took the coding classes in high school:

So, I was talking to her [about her plan to do animation], ‘Like the computer science? That’s like computers. That’s a horrible idea. Nobody likes computers.’ Like none of my friend group had ever taken the one programming class at our high school. (Julia)

However, learning more about computer science from her friend led Julia to change her mind and choose to major in computer science with an emphasis in bioinformatics:

She pulled up this article, like ‘10 Cool Things You Can Do with Computer Science,’ and I was like looking in there and was like, whoa that actually is really cool. That’s some cool like science stuff but without having to dissect anything. And like not having to know chemistry super well. Like I kind of just began to realize what an impact it could have. And also, that it was relatively like logic-based which is something that I like. I like logic. (Julia)

Since we did not specifically ask about previous exposure, it is possible that women had more prior experiences than we heard about, but they did not come to mind as meaningful for their major choice. Regardless, when asked about their process of choosing a major, most women did not place the same emphasis on previous experiences and instead talked about their initial decision being driven by other factors (e.g. interest and flexibility).

Women often spoke of lacking competency and feeling behind, yet spoke of a growth mindset and persevering through difficulties in STEM

Our next major theme appeared to connect with our previous theme of previous exposure, as most of the women felt that they were already behind in their major since many of the men already had experience. This was especially true for majors that included a lot of computer programming. About half of the women commented that they felt inexperienced in their field because they did not have as much prior experience as the men in their male-heavy classes. For example, Olivia said,

There are a lot of people with a lot of computer programming experience and have been doing it since they were like 10 and they were very good at it, and I started two semesters ago. So, I feel a little behind in that aspect of the major. Like I’m just still trying to catch up with everybody else. (Olivia)

As we read about our participants’ obstacles (such as feeling behind), we looked for examples of a growth mindset versus a fixed mindset. In line with Dweck’s work (Dweck, 2010), we classified comments as demonstrating a growth mindset when they discussed their obstacles as learning opportunities or demonstrated an attitude that they could learn to do anything regardless of its difficulty. On the other hand, we classified comments as demonstrating a fixed mindset if participants saw their own abilities as something that was predetermined or as an immutable character trait; in other words, they talked about their talents and abilities as if they were either good at something or not good at something, and they could not change that.

We also observed that throughout these interviews, women mostly displayed a growth mindset in terms of their competency. Seven of the women made comments referencing that although they had weaknesses, that they were able to overcome those weaknesses through hard work.

I’ve discovered that there was like perspiration and inspiration Where like inspiration is you’re naturally smart and perspiration is you work really hard … I’m more on the perspiration level where it’s like I might not have all the inspiration, but I work my tail off for it. And so I’m in computer science, so I guess like I’m smart but it’s more that I just work hard. (Isabella)

Leila also demonstrated a growth mindset when she said,

I would say I’m not really the most technical person. Like programming hasn’t really come to me super easily. But I really enjoy, like, being faced with a really hard topic, and or just like, I don’t know, something hard and being able to work really hard at it and get better. Even if I’m not like super proficient, if I just can get better. (Leila)

Although it was obvious to her that other people had more experience in her field, and she even commented that she contemplated changing majors, she decided that she could see incredible results if she worked hard academically. This resilience in the face of challenges was shared by most of the women that we interviewed. Interestingly, women feeling behind and like they were playing catch up did not seem to lead to them changing majors. Many of them discussed considering switching majors due to difficulties, but not actually switching. Most of the women who had switched majors talked about doing so to better match their interests (see, Table 1).

We observed that the men we interviewed mostly displayed a fixed mindset regarding competency. Five men made comments that suggested they thought that their abilities were set, which was usually a positive as they felt they were naturally competent. Only one man talked about his competency in a negative way as something he could not overcome. Here, Dave describes why he changed his major after taking a computer science course:

I mean, it definitely played a big role in it. Just recognizing my weakness that I wasn’t able to–like that computer science just isn’t my thing. That definitely helped me to shut some doors. (Dave)

Kevin also demonstrated a fixed mindset. In the following quote, he talks about how he is simply not cut out to do certain things but has natural aptitude for others:

Oh, gosh. I’m not sure if I could do that. I’m not sure if I could learn Dutch either. I’m not that talented … I can draw, and I can animate. I’m really good with computers and how that works, that I know I could be an animator and entertain through that. (Kevin)

One man, Graham, differed from the others as he talked about how working through his difficult major is fulfilling for him. His description of his growth mindset was more similar to the women’s descriptions:

I started to like it more and more. Even if the course is really hard and really frustrating, sometimes just the understanding that even though I sometimes might not feel like I am gonna make it, the sense of accomplishment especially is what keeps me going. (Graham)

Perhaps this general difference by gender is a byproduct of the imbalance in prior experience and stereotypes regarding gender in STEM, and computer science in particular. Many women that we interviewed talked about their strong growth mindset as if they had to have one to succeed. On the other hand, we found that most of the men had prior experience and already felt competent, or close to it, in their major by the time they had reached college. Perhaps because of this, the men displayed more of a fixed mindset in their descriptions.

Competency and peer relationships were the most impactful influences on students’ sense of belonging

As part of the interview, each interviewee was asked two questions: What makes you feel like you belong in the major, and what makes you feel like you do not belong in your major? The two biggest themes we noticed were that students felt more belonging if they viewed themselves as competent and/or if they had social connections or shared similarities with their peers:

So my entire life, I’ve kind of been a pretty nerdy guy. My mom was a math major at BYU and I guess I kind of inherited that same side of the brain because I’ve always liked math. So I’ve always wanted to do some type of engineering. I’ve known more about computers than my parents since I was like 3. (Stanley)

I think for me, I managed to get a good group of friends my freshman year where we went through [CS] 142 together and then [CS] 235, so that really helped a lot. But it’s just nice to share–to be in the classroom and share the same interests as people, you know (Graham)

Like luckily, I could talk to people, and I love doing that. So, I make friends, and so like the community of it kind of makes me feel like I belong (Isabella)

I’ve been lucky because I’ve been able to do research in my department and to be a T.A in my department and to have taken classes from people and from professors in my department … things like that I guess help me to feel like, yes, I feel comfortable in this department. (Sandy)

On the other hand, participants who felt they were quite different from their peers or lacked competency did not feel like they belonged:

For me it’s like, OK great. I’m surrounded by these people and none of them are like me. (Julia)

… usually I don’t feel like I belong because I’m like not catching up as quickly as other people seem to be. (Jeanette)

I don’t like generalizing people, but there is kind of a stereotype-ish, you know, as to people in a computer science like area … and I feel like I don’t have as much of that stereotypical kind of the personality you think about when you think computer science or math or engineering or something like that, you know. (Stanley)

There’s a lot of people that have prior experience with computer science. I mean I have a little bit, but there’s some people that have already like made super big projects, done all these things … But especially right at the beginning when I was struggling to learn how to do like simple stuff, when people were like, ‘Oh, this lab was so easy.’ And it’s just like I spent five hours on it yesterday, you know. (Graham)

As noted in previous sections, women in our sample were more likely to feel behind and thus lack perceived competence. This decreased their sense of belonging as expressed here:

I definitely felt intimidated, and I felt like, you know, they’re doing these really impressive things and I’m like really far out from being like that competent … I entered this late in the game and maybe I’m behind in terms of like my research experience or my knowledge. Yeah, just feeling like intimidated by really smart competent people can sometimes make me feel out of place. (Shannon)

Lacking belonging due to feeling in the minority was also a major theme for women, specifically. Nine of the ten women brought up the concept of gender without being asked about it, while none of the males brought up gender at all when discussing their major. Interestingly, the women often talked about not just being in the minority but also how this was connected to lacking competency and feeling behind as we discussed above. For example, Leila described her experience belonging in her major in this way:

When I first came [to college], I really struggled with that. I was really young, obviously, but I was also like one of the only girls in my classes. And sometimes I was a little bit intimidated by other people’s knowledge like I felt like I was really far behind everybody else. (Leila)

Another example of a woman reflecting on a lack of female representation in STEM majors is Anna, who described her experiences feeling judged because of her gender or hearing peers make comments about women in STEM:

Also, I think kind of being female might play a little bit of a role because there is still kind of a minority in STEM and computer science. So yeah, I hate the way that sounds but I kind of get those vibes from some people. Where it’s like ‘oh you’re female, I’m a male, I’m better at this. I got this, you’re new, I’m going to outperform you.’ … I don’t want to jump to stereotypes or large conclusions but … stuff that I heard in class. Like ‘Well what are you doing here? Why don’t you go do a social science?’ or whatever. (Anna)

It is important to remember when reading quotes like this is that we were specifically interviewing women that took at least one computer science class, which are generally male dominated. It was clear that these women felt like a minority in their majors. This is alarming when we remember that relationships with peers help people to feel like they belong, and the females may have a more difficult time relating to others if there are few women in the major. However, the women we interviewed did not cite this difficulty as a reason to avoid the major or leave the major.

Discussion

We took a qualitative approach to deeply understand the experiences of nineteen students in or adjacent to computer science and bioinformatics majors as they chose their major. As we were interested in the gendered experiences of students in this specific conservative religious population, we took an inductive approach without a priori hypotheses. We found four major themes: (1) pressure from traditional gender roles for most students, (2) women lacking the level of previous exposure of their male peers, (3) women feeling behind but persevering with a growth mindset, and (4) decreased sense of belonging linked to a lack of perceived competency and/or lack of similar peers.

Our major themes fit well within social cognitive career theory, SCCT (Lent et al., 1994). According to SCCT (see, Figure 1) from (Lent et al., 2000), person inputs such as predisposition and gender interact with one’s background and context from an early age to influence the types of learning experiences in which one participates. These experiences then impact self-efficacy and outcome expectations, which influence interests, goals, and actions. Social context can also influence goals and actions during the active phases of major choice and career selection. In summary, SCCT explores the reciprocal linkages between person, environment (both objective and perceived), and behavior (Lent et al., 1994, 2000).

Our first theme of pressure stemming from traditional gender roles for both men and women validates SCCT’s emphasis on social context impacting outcome expectations and decisions both distally and proximally. For our participants, important outcome expectations related to the students’ desired work/life balance. Distally, cultural adherence to traditional gender roles and gender stereotyping could impact students from an early age (Jackson, 2007; Kollmayer et al., 2018). Both men and women we interviewed emphasized family-focused goals, primarily men providing for children and women prioritizing childcare, and these goals and aspirations likely formed from an early age. Cultural influence regarding family gender roles was also proximal for our participants, as some were already married, one had a child, and many others spoke of getting married and having children soon.

While research suggests that traditional gender roles have become less constricting over time, with breadwinning and parental roles being more shared between men and women than in the past (Perrone et al., 2009), participants in our population still spoke of feeling pressure to adhere to classic roles. Perhaps more conservative and religious populations such as this one may not see the blurring of traditional gender roles at the same rate as other populations, but gender stereotyping has also been shown to be persistent in the general population even as behaviors change (Haines et al., 2016). Furthermore, our participants did show attitudes congruent with some blurring of gender roles. Many men we interviewed appeared to ascribe to what has been termed the new fatherhood ideal, as men become more and more interested in being involved in their children’s lives (Petts & Knoester, 2018), even if the breadwinner ideal persists. In our study, men spoke of wanting a balance between earning just enough money to provide while also having time to be with family. In addition, while some of our women participants spoke only of having a career if needed, many wanted some type of career, even if it was from home or part-time to prioritize their role as a mother. While all our subjects share similar cultural contexts due to attending the same school and association with the same sponsoring church, not all of them responded to it in the same way, and some even explicitly pushed back against cultural gender norms. This aligns with the consideration for perceived environment in SCCT in addition to objective environment and highlights the complex ways in which individuals are influenced by their culture (Lent et al., 2000).

Our second and third themes of women lacking the previous exposure of our male participants and thus expressing less self-efficacy align with SCCT’s proposal that personal characteristics like gender interact with the environment to determine learning experiences and thus self-efficacy. Our interview results suggest that men and women experienced different learning experiences before college even though cultural contexts were similar, and this could explain the major difference in the way our male and female participants spoke of their self-efficacy. Sheu et al. (2018) describe both direct and indirect learning experiences that influence self-efficacy. Direct learning experiences include personal performance successes and failures, verbal or social persuasion, and physiological or affective states/reactions like anxiety. Indirect learning experiences include vicarious learning by observation of social models. Many men in our study described direct learning experiences, often personal success, in STEM fields before college, but the women rarely did. The most common previous exposure the women spoke of was vicarious, and the role models they described were usually men. Previous research has found that negative affective states, such as anxiety, to be strong predictors of lowered self-efficacy for women especially (Sheu et al., 2018), but our participants did not explicitly talk about this in our interviews.

In our third theme, we saw that even with this self-described lack of competence or self-efficacy, women were overcoming barriers with a growth mindset. Van Aalderen-Smeetts and Van Aalderen-Smeets and van der Molen (2018) proposed three possible models by which a growth mindset could impact STEM major or career choice: mediated through self-efficacy, stereotypical thinking, or through motivation. Based on our qualitative results, self-efficacy and/or motivation appeared to mediate the effect of women’s growth mindset on their major choice. Women with a growth mindset described a greater self-efficacy because of their growth mindset (e.g. “even if I can’t do it now, I will learn it”), and these women also described the motivation to push through setbacks and persist even if they felt behind. The women with a growth mindset still described computer science majors in a stereotypical way (e.g. as geeky men), so we do not really have evidence that their growth mindset decreased their stereotypical thinking. Huang et al. (2019) tested a model that included growth mindset in addition to anxiety and self-efficacy. They found that growth mindset had an indirect effect on interest (through self-efficacy) for middle-school boys but not girls. However, our qualitative results found the opposite, and women’s interest and perseverance seemed more dependent on their growth mindset than the men, perhaps because they had more barriers to overcome.

Finally, we found a fourth major theme that lack of perceived competence and/or lack of similar peers both decreased students’ sense of belonging in their majors. Our results support those of Lytle and Shin (2020) who found that growth mindset beliefs predicted greater self-efficacy, which then predicted greater belonging in STEM. Competence emerged as a major theme for our participants when talking about belonging, and those with a growth mindset talked about the drive to persevere even if that competence was questioned. However, we also found it interesting that women’s descriptions of their competence did not always match objective measures of competency. For example, one female student described feeling behind and struggling to understand, yet she received an A in the course she was referencing. She also mentioned feeling different than her peers in computer science, so perhaps feeling in the minority and that lack of belonging led her to question her competence.

Limitations

Our conclusions are limited due to our study methods. As our focus was qualitative, our small sample size of nineteen students does not allow for quantitative comparison of any kind, and we are unable to conclude whether the themes found in our study are generalizable to larger populations. We also acknowledge that our focus on participants from a religious institution may limit the applicability of our findings more broadly, since most other populations will not share the specific religious context of our sample. However, the majority of the United States population is religious, so we hope our findings are still helpful to other groups.

Furthermore, we purposefully recruited participants from computer science and bioinformatics courses so that we would get a variety of actual majors rather than recruiting from declared computer science or bioinformatics major lists. This has the advantage of increasing our chances that some of our participants chose a different major after taking the course or eventually chose to leave the field, which allows us to learn about more negative experiences with computer science/bioinformatics. However, this also limits our conclusions since the participant pool is quite heterogenous and we are learning about somewhat diverse experiences rather than learning about the shared experiences of a very cohesive group (e.g. students who declared a computer science major and then switched majors after an introductory course).

We also acknowledge possible selection bias in our sample, because low-performing students were less likely to respond to our invitations to participate in the study. Additionally, students who switched out of computational majors early may not have taken the classes from which we pulled our participants, and women without a growth mindset or the motivation to work through difficulty may have switched out of these majors early.

As with all qualitative research, we also acknowledge that our research team is not free from bias, and we bring our own cultural context and personal experiences with us to the data analysis process. However, we strove to continually question our biases using an iterative analysis process that continually required us to go back to the interview text to find evidence for our themes.

Future research

We plan to use the themes found in this qualitative study to design a survey that we can distribute to a large number of students in this conservative religious population. With a larger sample size, we could quantitatively examine the generalizability of our themes to the larger population and investigate predictors of variables such as previous exposure. Longitudinal studies would also allow us to follow young students with different types and levels of early exposure and track their perceived competency, academic performance, and major choices. Quantitative studies would also allow us to more fully investigate the relationship between competency and switching majors. Almost all of the students we interviewed that had switched majors reported that it was due to their interests, but other parts of their interview hint that feelings of belonging and their academic performance may have played a role in their major switch. Finally, more work is needed on the ways in which a growth mindset fits into SCCT and impacts the gender gap in certain STEM fields.

Implications

There have been calls to diversify our STEM workforce in the United States if we are to reach our creative potential and compete on a global scale, particularly by increasing women’s participation in STEM fields that lack such representation like computer science (Beede et al., 2011). Based on our interviews, we suggest two ways to increase the representation of women in computer science and related fields, especially in conservative religious contexts.

First, our results suggest that if women had more previous exposure to coding or related skills when they were younger, they might be more likely to choose a major related to computer science, they would likely feel less behind in computational courses or majors, and that increased perception of competency would increase their sense of belonging in computer science or bioinformatics. We are not the first to make this suggestion. As early as the 1990ʹs, increasing prior experience with computing was suggested as a mechanism by which to achieve gender parity in computer science (Taylor & Mounfield, 1994), and more recent studies suggest that prior experience may be especially predictive of positive attitudes toward computer science for women (Hinckle et al., 2020). Other studies have shown that women are less likely to have prior programming experience than men, and that prior experience predicts better performance in introductory programming courses (Holden & Weeden, 2003; Wilcox & Lionelle, 2018). While these studies noted that the benefit of prior experience disappeared by later courses, students may transfer out before that point, or if they stay, that decreased sense of belonging and feeling of being behind may still have lasting effects. Our qualitative results support previous results and suggest that increasing young girls’ exposure to STEM, and coding skills specifically, is still needed and has the potential to broaden participation in college computer science and related majors.

Second, students in this conservative religious context are still heavily influenced by traditional gender roles and strong family values. Both the men and women we interviewed prioritized current or future family and children when making decisions about their major and future career. For women, this meant valuing future job opportunities that were flexible and allowed for the possibility of working part-time and/or from home. Thus, increasing the availability of these types of flexible positions could drive more women into computational fields. Again, this is not a new idea and others have advocated for more flexible work options in STEM fields to increase gender equity, especially part-time options that are more highly valued and allow for transitions back to full-time work when desired (Cech & Blair-Loy, 2019). Some of the women we interviewed already saw that potential for flexibility in computer science or related fields, and that was one factor that drew them to those majors. However, even where flexible work options already exist, making more girls and women aware of that potential when they are making important decisions about their college major is likely to increase participation by women who prioritize family responsibilities.

Acknowledgments

We would like to acknowledge Anika Hubbard for organizing interview transcripts prior to analysis and Jack Stalnaker and Kevin Gutierrez for sharing thoughts on the interviews after an initial listen.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/08993408.2022.2160144

Additional information

Funding

This work was supported by Brigham Young University through a College Undergraduate Research Award given to undergraduate students Emilee Severe and Anika Hubbard.