Introduction

The impressive dropout rate from computer science courses among women is hardly a secret. Given the centrality of technology in the contemporary economy and society, educators, policymakers, sociologists, and other professionals have been concerned about unequal gender participation in engineering and science careers. Gender issues are particularly significant in computing science education, with some important ones concerning the reasons for why so few women choose to study computer science, and which factors lead them to dropping out and declining a computer career (^{MISA, 2011}).

The literature points out many issues faced by women in computer science. Gender stereotype, humiliation, objectification, and prejudice are daily challenges for women who struggle for a space in technology, starting even before entering university and usually extending throughout their careers. Gender bias begins early, before college, when family members do not support the choice of a technological course; when teachers and classmates underestimate the technical capacity of female students and diminish their merits. Combining these facts to a lack of female references working in computer science, girls are discouraged to enter and stay in the area (^{BERG; SHARPE; AITKIN, 2018}; ^{GOMES et al., 2018}).

According to ^{Strayhorn (2018)}, the sense of belonging differs according to students’ social identities, such as race, gender, and sexual orientation, and the conditions they encounter on campus. Both social segregation and academic unpreparedness are evident among women from working and middle class families. In contrast, women from upper-class families show a continuation of family traditions. Class plays a significant role in construing marks that define “belonging” at elite universities (^{OSTROVE, 2003}).

Some studies suggest that one main reason why women dropout of computer science university courses is the feeling of not belong in the field or major (^{BARKER; COHOON; THOMPSON, 2010}; ^{COHOON, 2002}; ^{MARGOLIS, 2002}). Since a sense of belonging is a well-known predictor of success in college, especially within STEM, discussions on diversity have identified it as a critical component of diversity and inclusion efforts in the tech industry (^{STRAYHORN, 2018}). According to the literature, undergraduate students who experience a greater sense of belonging feel more connected to their environment and are more likely to persist to graduation. Female students experience computing environments differently due to sexism, which potentially leads them to feel unwelcome and lacking a sense of belonging to computing spaces (^{SAX et al., 2018}).

^{Barker, Hovey, and Thompson (2014)} emphasize the importance of classroom experiences for student permanence. Regarding the sense of belonging, they argue that classrooms are the main venue where students are exposed. Knowledge, presentation, and expectations about the types of people who do (or do not) belong in a undergraduate program, while being a minority in the classroom and facing various gender stereotypes, results in a loss of confidence for women, which can then lead to switching majors (^{BARKER; HOVEY; THOMPSON, 2014}).

The gender gap is an issue that concerns both educators and professionals in the field. Computing is an area that presents a high overall dropout rate, and the absence of women contributes to the steep decline in university enrollment (^{MISA, 2011}). Computer science programs are looking for ways to improve permanence and satisfaction for all students, partly due to low enrollment and high dropout rates. Diversity in the field is also important, as it increases creativity and provides a better understanding of customer’s needs (^{CARLSON, 2006}).

A major tool used to aggregate the experiences gained from a range of different studies to answer specific research questions is the systematic literature review (^{KITCHENHAM; CHARTERS, 2007}). Despite challenges faced by women in computer science majors and their consequences, such as lacking a sense of belonging, no systematic reviews were found between 2015 and 2020 that addressed this topic. We opted for limiting our search to this period because gender issues in computer science are relatively recent and the decline in the number of women enrolled in the field is increasing (^{MISA, 2011}). Given the plethora of issues faced by women in computer science, which will be reported in the next section, a study that establishes a shared sense in the literature is of paramount importance. More specifically, this systematic review identifies, delimits, and categorizes the problems experienced by female undergraduate computer science students to verify how such issues relate to the lack of a sense of belonging and, additionally, what is being done to solve them.

Negative experiences in computer science

Computer scientists are stereotyped as socially inept, brilliant, and obsessively focused on technology (^{CHERYAN; MASTER; MELTZOFF, 2015}). A study carried out in two secondary schools in Scotland in 2018 revealed patterns of stereotypes and social expectations children have regarding women who pursue careers in computer science. Data was collected by means of rich picture (collaborative drawing), a mechanism for exploring multiple opinions. Before the activity began, children were asked to express, via drawings, why there were few women in computing and what women who followed this career were like. The findings indicated quite negative physical and behavioral stereotypes (^{BERG; SHARPE; AITKIN, 2018}).

Gender-based discrimination against women is more noticeable in the job market, given the wage gap and low representation in high-level positions (^{BRAHNAM; KARANIKAS; WEAVER, 2011}). According to Guia de orientações sobre assédio moral e discriminação no ambiente de trabalho (^{TJDFT, [2019}]), (Orientation Guide on Psychological Harassment and Discrimination in the Workplace) of the Brazilian Court of Justice of the Federal District and Territories, discrimination can be understood as any distinction or exclusion based on gender, sexual orientation, religion, political or philosophical belief, race, or disability. Discrimination seeks to nullify or restrict the exercise, under similar conditions, of human rights and fundamental freedoms. It clarifies concepts such as mansplaining, gaslighting, manterrupting and bropriating. Mansplaining occurs when a man undermines a woman’s knowledge and devotes his time to explain something obvious to her, as if she couldn’t understand it on her own. Gaslighting is emotional violence that occurs via manipulation, leading the woman and those around her to question her sanity or think she is incapable. Manterrupting define situations when women are unable to complete their ideas due to unnecessary interruptions by men. Bropriating happens when a man appropriates a woman’s idea and takes credit instead. Discrimination can lead to harassment, that is, any abusive, extreme, purposeful and frequent action that generates negative consequences for women’s self-esteem, besides harming their dignity (^{BRASIL, 2009}).

Several situations contribute to the gender gap in some fields. For example, women may be aware of the discrimination they may experience in a given field, making entry and progress in STEM careers more difficult. ^{Cheryan et al. (2019)} categorized other issues faced by women beyond perceived discrimination, such as gender devaluation, concerns about the possibility of stereotyping, lack of self-efficacy, low sense of belonging, romantic concerns, incongruence of goals, and fear of backlash.

Besides the physical and personality stereotypes women face when studying computing, common belief says that boys are more competent in the field. Thus, male students emphasize have their abilities emphasized (^{MIURA, 1987}), even though, in practice, boys and girls generally have the same capacity (^{HARGITTAI; SHAFER, 2006}). Besides believing more in their computer skills, boys also seek more support and encouragement from parents and friends (^{BUSCH, 1995}). In fact, the stronger the identification with a gender role, the stronger the belief in self-efficacy, which is necessary for motivation (^{HUFFMAN; WHETTEN; HUFFMAN, 2013}).

Women begin facing such issues at home, even before choosing a profession, due to cultural stereotypes. Their time in college is particularly challenging. Female computer science students can face decisive challenges in staying in the program, including lack of prior programming experience and a sense of isolation due to non-diverse classrooms. Consequently, female students may interrupt progress towards a degree and decrease their self-confidence in their ability to succeed in college and computing careers (^{DEKHANE; NAPIER; NAGEL, 2016}).

Female students’ low sense of belonging is a reaction to more depressing feelings of conformity to the stereotypes associated with computer scientists. This causes women’s lack of interest and motivation in STEM fields (^{MASTER; CHERYAN; MELTZOFF, 2016}). Studies show that the sense of belonging is an important predictor for college success (^{STRAYHORN, 2018}) and is dependent on the environment that students frequent, which shows the important role of educators in promoting this feeling in students (^{SAX et al., 2018}). A stereotypically classroom with gender disparities can result in a lower sense of belonging (^{CHERYAN; MELTZOFF; KIM, 2011}).

Several programs worldwide sought to mitigate the absence of women in computing, such as the work conducted by Lund University, which sought to improve our understanding of campaigns for greater equality by seeking to comprehend the meaning, or meanings, of the #MeToo movement. The study examined how gender relations are expressed by women within IT. The campaign itself provided data showing that women are exposed to sexual assault and harassment within IT-oriented spaces (^{LINDGREN; ENEGREN, 2018}).

This brief overview of the issues faced by women in computer science highlights the importance of a more in-depth understanding of this scenario. Seeking to unveil the main problems affecting women in computer science and how we can tackle them, we present in the next sections the method used and the results of our systematic literature review.

Method

A systematic review is a method of synthesizing evidence focused on one or more particular, well-defined questions, area of knowledge, or a phenomenon of interest (^{KITCHENHAM; CHARTERS, 2007}) that aims to gather, critically evaluate, and interpret all available and relevant data on the topic in a quantitative and/or qualitative manner. It is a rigorous methodology that applies explicit and systematic research methods, in which the procedures developed must be recorded at each moment, allowing replicability of the process.

Systematic review is a reproducible strategy for screening and including studies, and so it can be considered a secondary study that has primary studies—scientific articles that report first-hand findings—as its data source. Moreover, the systematic review also proposes to identify gaps in the literature and suggest future studies, besides help to inform practices and policies by providing unbiased and integrated evidence.

The present systematic review was elaborated using the reliable, rigorous, and auditable method proposed by ^{Kitchenham and Charters (2007)}, which comprises planning and performing the review, and reporting the results obtained.

The steps applied during the development of this study include but are not limited to: identification of the need for a review; research design; preparation of review protocol; extraction, analysis and synthesis of data obtained in database searches; selection of primary studies; reading of selected papers to answer the research questions; and report writing.

Study selection

Several articles report the choice of female high school students not to pursue careers in STEM fields due to gender discrimination, stereotyping, lack of representation, and even repulsion for exact sciences (^{VAARMETS, 2018}). Some projects were created to combat this lack of female representation in computing, seeking to spark the interest of high school girls in pursuing a career in Computer Science (^{CHERYAN; MASTER; MELTZOFF, 2015}).

Many papers pointed out problems faced by all genders in the STEM academic environment, such as stress, depression, anxiety (^{BOEHM et al., 2016}), suicide ideation (^{PRUSTY; PANDA, 2017}), and even the lack of didactic methods (^{RYOO, 2019}).

Others presented the challenges faced by female university students in all fields, such as sexual assault, fear, and the fact that they do not seek help (^{HARRIS; TERRY; ACKERMAN, 2019}). Despite being very interesting, these papers do not address specific issues faced by female computer science undergraduates specifically, and were thus not accepted for data extraction.

The search strategy used returned a total of 818 articles from the six online databases. Fifteen duplicated articles were excluded. Only 24 works were accepted for data extraction. This reduced number of selected articles is due to problems with the indexing databases, which returned several articles outside our scope after the search strategies were entered. We only accepted papers that met the inclusion criteria. Figure 1 presents the flow chart of study selection and data extraction.

Source: Prepared by the authors

Figure 1 – Study selection and data extraction 

Figure 2 presents the articles accepted per year, after study selection and criteria assessment. We can observe that research on the topic under study had a substantial growth from 2015 to 2017, dropping in 2018 and increasing again from 2018 to 2019.

Source: Prepared by the authors.

Figure 2 – Accepted articles per year 

Articles that met the inclusion criteria were cataloged by abstract, author, title, year, and journal. Papers addressing problems faced by female students in STEM were not excluded, since computing belongs to these fields.

The accepted articles were:

1. Balancing the equation: mentoring first-year female STEM students at a regional university (^{REID et al., 2016}).

2. Understanding student retention in computer science education: the role of environment, gains, barriers and usefulness (^{GIANNAKOS et al., 2017}).

3. Do you think you can? The influence of student self-efficacy on the effectiveness of tutorial dialogue for computer science (^{WIGGINS et al., 2017}).

4. STEMinism (^{MYERS; GALLAHER; MCCARRAGHER; 2019}).

5. Computing and STEM in Greece: gender representation of students and teachers during the decade 2002/2012 (^{KORDAKI; BERDOUSIS, 2017}).

6. A model of threatening academic environments predicts women STEM majors’ self-esteem and engagement in STEM (^{CASAD; PETZEL; INGALLS, 2019}).

7. Gender and student course preferences and course performance in computer science departments: a case study (^{IOANNIS; MARIA, 2019}).

8. Closing the gaps and filling the STEM pipeline: a multidisciplinary approach (^{DOERSCHUK et al., 2016}).

9. College admissions viewbooks and the grammar of gender, race, and STEM (^{OSEI-KOFI; TORRES, 2015}).

10. Swimming against the tide in STEM education and gender equality: a problem of recruitment or retention in Malaysia (^{GOY et al., 2018}).

11. Broadening participation not border protection: how universities can support women in computer science (^{MICHELL et al., 2017}).

12. Female undergraduate STEM persistence: a focus on the role of living and learning communities (^{NIX; ROBERTS; HUGHES, 2016}).

13. Development and validation of the Value-Expectancy STEM Assessment Scale for students in higher education (^{APPIANING; VAN ECK, 2018}).

14. African-American women’s experiences in graduate science, technology, engineering, and mathematics education at a predominantly white university: a qualitative investigation (^{ALEXANDER; HERMANN, 2016}).

15. Attrition of women in STEM: examining job/major congruence in the career choices of college graduates (^{XU, 2017}).

16. Perceptions of male and female STEM aptitude: the moderating effect of benevolent and hostile sexism (^{REILLY; RACKLEY; AWAD, 2016}).

17. Investigating gender issues in an undergraduate computing program (^{AMARAL et al., 2017}).

18. The status of women in STEM in higher education: a review of the literature 2007-2017 (^{BLACKBURN, 2017}).

19. Institutions developing excellence in academic leadership (IDEAL): a partnership to advance gender equity, diversity, and inclusion in academic STEM (^{BILIMORIA; SINGER, 2019}).

20. Supports and pushes: insight into the problem of retention of STEM women faculty (^{PASCALE, 2018}).

21. Gender in academic STEM: a focus on men faculty (^{SATTARI; SANDEFUR, 2019}).

22. Starting at the crossroads: intersectional approaches to institutionally supporting underrepresented minority women STEM faculty (^{ARMSTRONG; JOVANOVIC, 2015}).

23. Lesbian, gay, bisexual, transgender, and queer students’ sense of belonging in computing: an intersectional approach (^{STOUT; WRIGHT, 2016}).

24. Accounting for the role of policy in the underrepresentation of women in computer science (^{PATISAS, 2016}).

Results

Discussion

All the issues discussed by the papers analyzed lead to the women’s low sense of belonging in the computer science environment, making them feel isolated and demeaned. Cultural problems such as machismo are responsible for instilling low self-esteem in female students. Low confidence begins to be built in childhood and stems from their experiences with family and friends. Many family members do not believe that women suffer any aggression because they are women; they strongly assert that if a woman is as good as a man, she will be equally respected. Unfortunately, however, women require help and encouragement to achieve an excellent technological level in computing, for even at home they face discrimination and social stigma. Pretentious judgments and statements like these further increase women’s low self-esteem, as if the responsibility for being treated differently or receiving lower wages was all theirs. Studies show that, depending on how girls are raised and their developed cognition, they may develop difficulties in some subjects or believe that they incapable of completing challenges in computing. Soon their beliefs about their abilities are impaired, leading to low self-confidence and low self-efficacy. Women may therefore experience difficulty in some subjects such as programming, and often poor didactics, peer discrimination, and lack of confidence are serious factors that lead them to abandon university and their careers.

The few women found in computing and on-campus reflcts on the minimal number of female professors and students. This scenarion can inhibit or even lead to the isolation and estrangement of the few women who remain. Unfortunately, as shown by many papers, gender segregation is cultural and changing the culture is a complicated endeavor. The fewer women in computing, the fewer girls become interested in computer science. The lack of female representation in computing leads to the abandonment of the field. Many studies have pointed out the lack of interest of high school girls in entering computing, but the lack of female representation in the field is also responsible for the dropout of girls who, at some point in their lives, tried to be part of it. This lack of interest is often not because women dislike or are not curious about technology, but because they rather not face gender stereotyping and isolation, as well as fear of not being salaries the same as men.

The stereotype of women persuing a computer science degree is the perception others have of them. Society sees women who study computing as strange and unfeminine people who lacking a good appearance. Women in STEM also face many judgments about their personality and how they live their lives. All this pressure leads women to develop low self-esteem and fear people’s reaction when declaring their major. Another problematic differentiator is that women and girls are more likely to be victims of crimes such as sexual harassment and sexual assault stemming from machismo. Their chances of hearing sexist jokes on campus or experiencing gender discrimination are high. Also, the burnden of motherhood is often much heavier than that of fatherhood. Female students who have children are judged because, according to society, they should spend all their time taking care of their children and not studying or working, while the same is not required of men. Often, women stop advancing in their professional careers to exclusively take care of their children. Hence, students who are mothers have a great potential to give up their careers for lack of support.

Although they are of paramount importance, the initiatives discussed in the “Answering Research Question 2” section focus solely on increasing women’s confidence and interest in Computer science. While change is still up to women, there needs to be real support, and we must enhance respect and empathy from everyone on campus. The initiatives presented have a small-scale, local impact and are difficult to implement on campuses. It is essential that measures within the classroom and on campus are taken to support female computing students. If women are actively recruited, and if classroom dynamics are deliberately adjusted to encourage their sense of belonging, female students are more likely to enter STEM fields and stay (^{MYERS; GALLAHER; MCCARRAGHER, 2019}). Cognitive gains and supportive environments positively affect permanence, while non-cognitive gains hinder it. Institutional neglect and the indifference of professors and colleagues only make matters worse, leaving the women alone “against the world,” who often lack support from those who should take care of them most—family and friends.

Awareness-raising campaigns, student support training, guidance from parents and guardians, improvement in teaching didactics, and projects that include women and motivate them in computing are examples of positive attitudes to be taken by educational institutions. Besides the aforesaid initiatives, in-class initiatives are essential to mitigate the issues faced by female undergraduate students. This systematic review is relevant for such educational interventions, providing information on the main problems faced by female students in computer science courses. Our results can help in theoretically-based academic, managerial and administrative decision-making processes regarding gender issues.

Conclusion

This systematic literature review on the issues faced by female computer science undergraduates examined the key reasons reported for career abandonment. It sought to highlight the main problems that must be overcome within university campuses and within the classroom, as well as to identify the limitations of the proposed solutions.

We found few specific studies on the problems faced by women in computer science on campus and especially in the classroom. Female student dropout in computing results from a low sense of belonging caused by lack of representation, negative gender stereotypes, low sense of self-efficacy, gender depreciation, and indifference of counterparts. These problems can disproportionally affect Black, Asian, and queer women.

The job market sees team diversity as a significant factor in better serving customers, just as colleges have an interest in retaining students. Consequently, some initiatives have been developed to mitigate those issues. But these initiatives have a local and temporary effect, being difficult to implement on other campuses and not guaranteeing the long-term permanence of women in computing fields. There is a demand for more effective initiatives to encourage women to persevere in undergraduate computing courses.

Pointing out the main problems experienced by female computing students and identifying the limitations of initiatives taken to solve them, is the first step for future research to propose good ways around them and to outline specific solutions within the classroom, that raise awereness of such issues among educators and even classmates. Attention to these issues may spark the interest of researchers in working to make undergraduate female students’ experiences while pursuing a STEM degree more positie, decreasing the chances of dropout. With the knowledge and understanding of the problems discussed by this systematic review, further research should focus on actions to mitigate or prevent such issues.