Decoding the Gap: A Retrospective Analysis of Women’s Experiences in

Software Engineering

Lucia Happe, Kai Marquardt, Ricarda Trumpf and Ingo Wagner

Karlsruhe Institute of Technology, Germany

Keywords:

Computer Science Education, Stereotypes in Computing, Diversity, Inclusive Education, Early Computing

Exposure, Career Aspirations in Technology, Educational Policy, Societal Perceptions.

Abstract:

The persistent gender gap in software engineering (SE) poses a signiﬁcant challenge in a world where digital

innovation is crucial to societal progress. This paper explores the underlying factors contributing to the low

participation of women in SE education and careers. Through a retrospective questionnaire study, we sought

to capture the experiences and perceptions that deter women from pursuing SE despite initial interest. Our

ﬁndings indicate that stereotypes, misconceptions about the ﬁeld, and a lack of early positive exposure inﬂu-

ence women’s decisions regarding SE. The study reveals that barriers such as the perceived incompatibility of

SE with personal interests, the daunting image of the SE work environment, and the absence of female role

models are critical deterrents. Furthermore, we discuss how early and continuous engagement with computing

can reshape perceptions and foster a more inclusive environment. The paper concludes with actionable rec-

ommendations, emphasizing that efforts to close the gender gap in SE should not only aim for demographic

balance but also harness the full potential of diversity for driving innovation. Ultimately, the study underscores

the need for systemic changes in education and policy to create a more equitable and dynamic SE landscape.

1 INTRODUCTION

The digital age has ushered in an era where innova-

tion is no longer the sole domain of large corpora-

tions or research labs. This democratization of tech-

nology, characterized by accessible digital tools and

lower barriers to entry, has transformed the landscape

of creation and dissemination. Yet, as we navigate

towards a sustainable and equitable future, a glaring

disparity persists in the composition of those who par-

take in technological advancement – particularly in

the ﬁeld of software engineering (SE). This disparity

is most pronounced in the gender divide, with women

signiﬁcantly underrepresented in SE education and

careers. The integration of diverse talents, especially

those of women who constitute half of the global de-

mographic, is not merely a matter of social justice but

a strategic imperative for innovation. A diverse work-

force guarantees products that cater to a broader spec-

trum of society, amplifying the collective earning po-

tential and, consequently, strengthening the economic

fabric (Albusays et al., 2021; Rodr

ıguez-P

erez et al.,

2021; Lorenzo et al., 2018).

Despite marginal advancements in female par-

ticipation in SE, the pace is lacklustre. The pre-

vailing barriers are not insurmountable; rather, they

are steeped in perceptions and myths that have long

coloured the narrative of SE. As identiﬁed by leaders

in the ﬁeld like Maria Klawe (Fidelman, 2012), mis-

conceptions around the allure, approachability, and

professional milieu of SE contribute signiﬁcantly to

the reluctance among women to enter the ﬁeld. How-

ever, evidence suggests that early exposure to com-

puting can reshape these narratives. Google’s 2014

study (Google, 2014) highlights the stark contrast in

attitudes towards computing between girls who re-

ceived academic exposure and those who did not.

The former group associated computing with ’future’,

’fun’, and ’interesting’, whereas the latter leaned

towards ’boring’ and ’difﬁcult’. This underscores

the impact that educational frameworks and curricula

have on the perception of SE among young women.

Despite these insights, a signiﬁcant number of

women who exhibit enthusiasm for SE are deterred

by the frustrations encountered in their learning jour-

ney (Happe et al., 2021; Happe and Buhnova, 2022;

Marquardt et al., 2023; Marquardt and Happe, 2023).

Addressing these barriers calls for interventions that

extend beyond tokenistic efforts for gender balance

(Gorbacheva et al., 2019). It necessitates a genuine

Happe, L., Marquardt, K., Trumpf, R. and Wagner, I.

Decoding the Gap: A Retrospective Analysis of Women’s Experiences in Software Engineering.

DOI: 10.5220/0012732100003693

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 16th International Conference on Computer Supported Education (CSEDU 2024) - Volume 1, pages 227-236

ISBN: 978-989-758-697-2; ISSN: 2184-5026

227

recognition of the value that women bring to SE and a

concerted effort to cultivate an environment that nur-

tures their interests and skills.

This paper probes deeper into the factors that dis-

suade women from pursuing SE. Through a retro-

spective analysis of survey data, we uncover the spe-

ciﬁc challenges and frustrations that lead to attrition

among women in SE. We shed light on the often in-

visible hurdles and aim to pave the way for actionable

solutions that will encourage and sustain women’s

involvement in SE. In our preceding study (Happe

and Buhnova, 2022) (focusing on P1 and P2 only),

we pinpointed various frustrations through quotations

and the deployment of personas within our dataset.

This paper builds upon that foundation by conducting

a comprehensive and systematic analysis of the word

lexicon employed in participants’ expressions.

2 METHODOLOGY

2.1 Study Design

The primary goal of our study was to explore the ex-

periences of women in computing, particularly focus-

ing on those who have disengaged from the ﬁeld as

well as those who have remained. This approach dif-

fers signiﬁcantly from previous studies such as those

by Joshi et al. (Joshi et al., 2013) and Armstrong et al.

(Armstrong et al., 2018). We aimed to understand the

pivotal moments inﬂuencing their career paths, iden-

tify major obstacles and drivers in their journey, and

gather recommendations for enhancing girls’ partici-

pation in computing.

2.2 Data Collection

Our primary instrument for data collection was a

questionnaire, crafted to delve into women’s engage-

ment with computing. In this study, we focus on

six open-ended questions of the questionnaire, which

encouraged respondents to reﬂect on their journey

in computing. These questions covered perceptions

of computer scientists, motivational factors, barriers

faced, and suggestions for improving computing ed-

ucation for girls. Participant demographics such as

age, gender, and major interests were also collected.

The open-ended questions asked were:

• Q1. What do you think about computer scientists?

Who are they? What do they do? Is there some

experience you have and would like to share with

us?

• Q2. What makes you feel most enthusiastic about

and interested in computing? What do you use a

computer for? What do you enjoy most?

• Q3. What was or would be the biggest driver on

your way to computing?

• Q4. What was or would be the biggest obstacle

on your way to computing?

• Q5. Are there any key points you think or feel we

need to know to make computing education better

for you or other girls?

• Q6. If you had any computing skills necessary,

what would you do with it (e.g., get a new job,

start a new company, become a teacher, imple-

ment your idea, etc.)?

2.3 Distribution and Sample

The questionnaire was disseminated through channels

catering to individuals with an interest in computing,

with a particular focus on organizations like Czechi-

tas (www.czechitas.cz) that offer late-education op-

portunities for adult women. Our global outreach

predominantly utilized Facebook groups dedicated to

late computing education for women. This strategy

was designed to attract respondents who had a strong

potential for pursuing computer science in their ear-

lier years, setting our study apart from previous re-

search.

From the initial pool of 151 responses, we ana-

lyzed 140 valid submissions after excluding incom-

plete entries and those not matching our target demo-

graphic criteria, such as responses from men. Our

respondents were categorized into three age groups:

18-26 (18%), 27-34 (41%), and over 34 (33%). The

engagement level was high, with 90% of participants

completing all open-ended questions, often providing

thoughtful and detailed responses.

2.4 Analysis Method

The analysis of open-ended responses was conducted

systematically through the following steps:

1. Data Cleaning. We began by removing ﬁller and

overly frequent words, such as redundant terms

from the questions themselves or common ﬁller

language.

2. Word Cloud Generation. Utilizing Mon-

keyLearn’s word cloud tool (https://monkeylearn.

com/word-cloud), we identiﬁed prevalent words

in the responses, which guided our initial catego-

rization.

3. Categorization and Labeling. We scanned

the responses for prominent words, their syn-

CSEDU 2024 - 16th International Conference on Computer Supported Education

228

onyms, and expressions conveying similar mean-

ings. These terms were then grouped into cat-

egories and assigned labels. Notably, responses

often received multiple labels due to the nuanced

nature of the data.

4. Frequency Analysis. Using the ”COUNTIF”

function, we quantiﬁed the frequency of each la-

bel. This was complemented by a custom ”ﬁn-

dIDs” function to trace the occurrences back to

individual responses. The categorized data, trans-

formed from qualitative responses, was then sub-

jected to quantitative analysis.

2.5 Persona Segmentation and

Typiﬁcation

Our survey garnered a total of 140 valid responses,

which were categorized into three distinct Personas

(Happe and Buhnova, 2022) based on their career

paths and relationship with computing (e.g. studied

CS as a primary degree or secondary, ﬁrst choice or

later in life). This diverse participant pool spanned

across three age groups and was nearly evenly dis-

tributed across the Czech Republic, Germany, and

other regions.

Persona 1 (P1) (42% of Respondents). This group

represents women who have pursued a career in com-

puting from the outset. They typically have a back-

ground in computer science or a related ﬁeld and have

maintained a continuous trajectory in SE. This per-

sona encapsulates those who found their calling in

computing early on and followed it through their ed-

ucational and professional journey. Typical Quote.

”I think computer scientists are cool people, gen-

erally socially awkward in my experience but well-

meaning... Also, I think computer scientists have a

hard time maintaining a work-life balance... comput-

ing problems require a lot of attention and sometimes

everything else in life gets sidetracked.”

Persona 2 (P2) (17% of Respondents). Constitut-

ing women who transitioned to computing later in

their lives, this persona reﬂects a non-linear path to

SE. These respondents may have initially pursued ed-

ucation and careers in ﬁelds unrelated to computing

but eventually found their way into SE. This group

is particularly noteworthy as it highlights the appeal

of SE to individuals with diverse initial career paths

and the permeability of the ﬁeld to professionals from

various backgrounds. Typical Quote. ”I am not a

person that enjoys computing on its own, I need some

higher goal. I like to think of it as means of fulﬁlling

my other goals in different ﬁelds. That’s what I’d like

to see more of – showing that IT is not just IT, more

commonly it is connected to some other ﬁeld and you

can work with anything being in IT.”

Persona 3 (P3) (40% of Respondents). This per-

sona includes women who have never considered en-

tering the computing ﬁeld. Their perspectives are

crucial for understanding the barriers and misconcep-

tions about SE that deter women from considering

a career in this area. This group provides insights

into external factors, societal perceptions, or personal

preferences that inﬂuence the decision against pursu-

ing a career in computing. Typical Quote. ”Hm, I

wouldn’t say there was one (obstacle). In retrospect,

it was a smooth autodidactic slide (away from CS) fu-

elled by personal interest.”

2.6 Comparative Analysis Methodology

This comparative analysis aims to discern the expe-

riences that inﬂuenced the career decisions of P1, P2

and P3. We speciﬁcally investigate what factors drove

P2 to eventually pursue computing despite initial de-

cisions against it, or P3 completely against it from

the beginning, and contrast these with the experiences

of P1 who consistently pursued a path in comput-

ing. The analysis seeks to uncover the frustrations and

challenges that steered P2 and P3 away from com-

puting initially and the driving forces that drew them

back to the ﬁeld.

3 RESULTS

This section presents the ﬁndings from our retrospec-

tive questionnaire study, aimed at understanding the

reasons behind the attrition of women in software en-

gineering (SE). The results are organized into several

thematic subsections, each addressing a speciﬁc as-

pect of our investigation.

3.1 Q1: Perceptions of Computer

Scientists

The key perceptions of computer scientists as iden-

tiﬁed by our respondents are described as ’problem-

solving’, ’diversity’, and ’social skills’, with ’prob-

lem’, ’women’, and ’work’ being the most prominent

words depicted. Table 1 reﬂects the frequency and

percentage of speciﬁc terms mentioned by respon-

dents when discussing their views on computer sci-

entists.

Decoding the Gap: A Retrospective Analysis of Women’s Experiences in Software Engineering

229

Table 1: Frequency and percentage of speciﬁc perceptions

of computer scientists.

Perception Category All P1 P2 P3

Problem solver 12 (9% ) 5 (8%) 2 (8%) 5 (9%)

Normal, nothing special 12 (9%) 6 (10%) 2 (8%) 4 (7%)

Introvert, low social skills 8 (6% ) 6 (10%) 1 (4%) 1 (2%)

Logical, analytical 7 (5%) 3 (5%) 1 (4%) 3 (5%)

Contributing, altruistic 7 (5%) 2 (3%) 2 (8%) 3 (5%)

Toxic, non-ethical 6 (4%) 2 (3%) 2 (8%) 2 (4%)

Intelligent 6 (4%) 4 (7%) 2 (8%) 0 (0%)

The data indicate a complex and multifaceted

view of the computer science profession among the

participants. While problem-solving skills are highly

recognized, there is also an acknowledgement of the

diversity within the ﬁeld. Social challenges and gen-

der disparities are noted, reﬂecting the need for cul-

tural and educational shifts in the computing environ-

ment. The comparative analysis of the responses to

the perceptions of computer scientists across the three

personas (P1, P2, P3) reveals the following distribu-

tion of responses within various categories:

• P1, who studied and stayed in computing, re-

sponses suggest a more nuanced understanding of

the ﬁeld, with higher mentions of both positive

(problem solver, logical) and negative (introvert,

toxic) traits.

• P2, who transitioned to computing later in life,

seems to have a slightly more idealistic view,

emphasizing altruistic and intellectual aspects, as

well ascontribution to society.

• P3, who never considered entering computing,

tends to have a lower incidence in several cate-

gories and may indicate less familiarity or engage-

ment with the stereotypes or realities of the pro-

fession, as well as it seems they may not strongly

associate intelligence with the ﬁeld.

An interesting observation is as well that all personas

perceive computer scientists as problem solvers at

similar rates (8% and 9%, respectively), and the view

that they are normal is held by 9% overall, suggests

an emerging view of these professionals as ordinary

individuals, challenging the stereotype of them being

markedly distinct or eccentric from the general pop-

ulation. Overall, these observations point to a com-

plex and evolving public image of computer scien-

tists, one that balances the recognition of their spe-

cialized skills and intellectual capacities with an ap-

preciation of their ordinariness and approachability.

This duality suggests a shift towards a more nuanced

and humanized understanding of professionals in the

tech industry, which could have implications for how

the ﬁeld is perceived and engaged with by the broader

society.

Table 2: Frequency and percentage of aspects associated

with enthusiasm for computing.

Perception Category All P1 P2 P3

Social interaction 29 (21% ) 8 (14%) 4 (17%) 17 (30%)

Programming 24 (17%) 9 (15%) 4 (17%) 11 (19%)

Creativity, creating, tinkering 12 (16% ) 12 (20%) 5 (21%) 5 (9%)

Makes work easier 21 (15%) 5 (8%) 4 (17%) 12 (21%)

Finding answers, research 18 (13%) 6 (10%) 4 (17%) 8 (14%)

Data analyses, simulation 16 (11%) 5 (8%) 3 (13%) 8 (14%)

Learning, exploring 15 (11%) 5 (8%) 4 (17%) 6 (11%)

Solving problems 9 (6% ) 4 (7%) 0 (0%) 5 (9%)

Endless possibilities 9 (6%) 5 (8%) 1 (4%) 3 (5%)

Supporting/helping others 9 (6%) 5 (8%) 2 (8%) 2 (4%)

Gaming 8 (6%) 4 (7%) 2 (8%) 2 (4%)

Entertainment (videos, music) 8 (6%) 4 (7%) 2 (8%) 2 (4%)

3.2 Q2: Enthusiasm and Interests in

Computing

Respondents expressed a diverse range of enthusi-

asms and interests within the realm of computing, as

depicted in the word cloud in Figure 1.

Figure 1: Word cloud visualizing the aspects of computing

that participants ﬁnd most engaging.

Respondents expressed a diverse range of enthu-

siasms and interests within the realm of computing.

Communication, coding, and creativity emerged as

the most engaging aspects, highlighting the multi-

faceted nature of computing that captivates interests.

Table 2 details the frequency and percentage of the

mentioned aspects, providing a quantitative measure

of the speciﬁc areas that participants associate with

their enthusiasm for computing.

The analysis suggests that while technical aspects

such as programming and data analysis are signiﬁ-

cant, the social dimension of computing and the abil-

ity to facilitate work and creativity also play a vital

role in sustaining interest in the ﬁeld. The compara-

tive analysis of enthusiasm and interests in computing

across the three personas (P1, P2, P3) for selected cat-

egories reveals the following distribution:

• The category Chat-

ting/Communication/Interacting with People

is represented across all personas, with P3 (30%)

CSEDU 2024 - 16th International Conference on Computer Supported Education

230

showing the highest association, suggesting that

they might view computing as more collaborative

than P1 (14%) and P2 (17%). This indicates a

general interest in the social aspect of computing

across all groups.

• In the category Write

Code/Programming/Coding, P1 and P3 are

quite similar in their number of responses, while

P2 is slightly less represented. This suggests that

coding is a signiﬁcant interest area for those who

stayed in computing (P1) and those who never

considered it (P3).

• The category Creativity/Create Some-

thing/Stuff/Tinkering shows a notable difference,

with P2 having the highest number of responses.

This indicates that individuals who transitioned

to computing later in life (P2) are particularly

drawn to the creative and hands-on aspects of

computing.

The data shows that enthusiasm for computing

spans a broad range of aspects, from the tech-

nical (like programming and data analysis) to the

more human-centric (like social interaction and help-

ing others). Notably, P3 often exhibits higher per-

centages in categories related to practical applications

(like making work easier and social aspects), whereas

P2 show more inclination towards creative and P1 to-

wards problem-solving aspects. This suggests that

different groups may be drawn to computing for var-

ied reasons, and these motivations need to be recog-

nized and nurtured to foster a more inclusive comput-

ing environment.

3.3 Q3: Drivers to Computing

The driving factors that propel individuals towards a

career or interest in computing were derived from the

data collected and are illustrated in the word cloud

in Figure 2. Personal interest, family inﬂuence (espe-

cially by fathers), and ﬁnancial prospects were among

the most cited reasons.

In this section, we investigate driving factors that

propel individuals towards a career or interest in com-

puting were derived from the data collected based on

responses to the questionnaire item Q6: What was or

would be the biggest driver on your way to comput-

ing?. A detailed breakdown of these motivating fac-

tors is presented in Table 3, showing the frequency

and corresponding percentage of responses.

This distribution of drivers underscores the impor-

tance of personal passion and interest in computing

as leading factors, followed closely by familial sup-

port where particularly often ’dad’ was mentioned,

Figure 2: Word cloud representing the drivers to computing

as reported by participants.

Table 3: Frequency and percentage of reported drivers to

computing.

Perception Category All P1 P2 P3

Personal interest 22 (16% ) 11 (19%) 6 (25%) 5 (9%)

Family 13 (9%) 11 (19%) 0 (0%) 2 (4%)

Job opportunities, money 11 (8% ) 3 (5%) 3 (13%) 5 (9%)

Creativity, projects 11 (8%) 6 (10%) 2 (8%) 3 (5%)

and economic incentives. It suggests that while in-

trinsic motivation is paramount, external factors such

as ﬁnancial beneﬁts and social environment also sig-

niﬁcantly inﬂuence one’s pursuit in the ﬁeld of com-

puting. The comparative analysis of responses to the

question about drivers to computing across the three

personas (P1, P2, P3) for selected categories reveals

the following distribution:

• The category Myself/Curiosity/Passion or per-

sonal interest and creativity is a predominant

driver, especially for P1 and those who transition

to computing later (P2), suggesting that intrinsic

motivation is crucial for this group.

• Family Inﬂuence is most prominent in P1, indicat-

ing that early familial support or exposure plays a

signiﬁcant role in shaping a career in computing

and suggesting that familial factors play a signiﬁ-

cant role in deterring the P3 group from consider-

ing a computing career.

• In Financial Prospects, Persona 2 (P2) is more

represented, which could indicate a perception of

better ﬁnancial opportunities in computing and

possibly reﬂect pragmatic considerations in their

decision to transition into computing. The impor-

tance of job opportunities and ﬁnancial consider-

ations is consistent across personas (however it is

less prominent for P1), reﬂecting a universal ap-

peal of the ﬁeld’s practical beneﬁts.

• For Own Projects/Creativity, P1 and P2 show a

notable interest, indicating a draw to computing

for creative and project-based work, especially

Decoding the Gap: A Retrospective Analysis of Women’s Experiences in Software Engineering

231

Table 4: Frequency and percentage of reported obstacles in

computing.

Perception Category All P1 P2 P3

Lack of support, no guidance, discrimination 17 (12% ) 10 (17%) 4 (17%) 3 (5%)

Lack of information/resources, language barriers 16(11%) 7 (12%) 6 (25%) 3 (5%)

Stereotypical surroundings 15 (11% ) 12 (20%) 0 (0%) 3 (5%)

Fear of failing, self-doubt 12(9% ) 6 (10%) 2 (8%) 4 (10%)

Stereotypes about knowledge 12(9%) 4 (7%) 2 (8%) 6 (11%)

Male-dominated environment 10 (7% ) 5 (8%) 2 (8%) 3 (5%)

Time constraints, age-related barriers 9(6%) 2 (3%) 2 (8%) 5 (9%)

Family pressure, societal expectations 8(6%) 5 (9%) 1 (4%) 2 (4%)

among those who transitioned to computing later

in life, and underscoring the importance of inno-

vation and creative freedom in retaining talent in

the ﬁeld.

3.4 Q4: Obstacles in Computing

The obstacles that participants identiﬁed as barriers

to entering or continuing in the ﬁeld of computing

include a lack of support, insufﬁcient resources, and

stereotypes within the educational and social environ-

ment. Table 4 provides a quantitative summary of

these barriers, indicating the frequency of each ob-

stacle as mentioned by respondents.

This data highlights the multifaceted nature of the

challenges faced by women in computing. It em-

phasizes the need for a supportive environment, ac-

cessible resources, and positive role models to miti-

gate these obstacles and foster a more inclusive atmo-

sphere in the computing domain.

The comparative analysis of responses to the ques-

tion about obstacles in computing across the three per-

sonas (P1, P2, P3) for selected categories reveals the

following distribution:

• Lack of Support are signiﬁcant obstacles for P1

and P2, with P1 slightly more affected. This sug-

gests that individuals who pursued computing ini-

tially or transitioned later face challenges related

to support systems and P2 prominently as well to

Lack of Information/Resources access.

• Stereotypical Surroundings are particularly high

for Persona 1 (P1) indicating a shared perception

of the computing environment as unwelcoming

or stereotyped, potentially impacting especially

those in computing.

• Worries/Fear of Failing is fairly evenly distributed

across P1 and P3. Highlights the psychological

barriers such as fear of failure and lack of self-

conﬁdence. This could be linked to the challenges

of entering the ﬁeld and the pressures associated

with it.

Table 5: Suggestions for improving computing education

for girls.

Perception Category All P1 P2 P3

Introduce female role models 10 (7%) 7 (12%) 2 (8%) 1 (2%)

Start technical education early 10 (7%) 4 (7%) 2 (8%) 2 (4%)

Avoid differentiating between boys and girls 7 (5%) 5 (8%) 1 (4%) 1 (2%)

Provide more encouragement 17 (12%) 4 (7%) 1 (4%) 2 (4%)

Offer all-female courses 7 (5%) 4 (7%) 0 (0%) 3 (5%)

Stop promoting stereotypes 7 (5%) 4 (7%) 0 (0%) 3 (5%)

Employ good and inspiring tutors 6 (4%) 4 (7%) 0 (0%) 2 (4%)

Teach useful and problem-solving skills 6 (4%) 3 (5%) 2 (8%) 1 (2%)

3.5 Q5: Improving Computing

Education for Girls

Efforts and suggestions to improve computing educa-

tion for girls range from introducing female role mod-

els to creating a non-discriminatory learning environ-

ment. The data in Table 5 summarizes the frequency

of speciﬁc recommendations provided by respondents

to enhance the computing education experience for

girls.

The recommendations underscore the signiﬁcance

of early exposure, role models, and an inclusive envi-

ronment as key factors for encouraging girls to pursue

and thrive in computing education. The comparative

analysis of responses on how to improve computing

education for girls across the three personas (P1, P2,

P3) for selected categories reveals the following dis-

tribution:

• Female Role Models are equally important for all,

with a notably high representation in P1, indicat-

ing the importance of having relatable ﬁgures and

mentors in the ﬁeld to inspire and guide.

• The emphasis on Starting Early/Teaching Chil-

dren is seen fairly evenly distributed across per-

sonas, suggesting that early exposure to comput-

ing is crucial for those who choose or transition

into the ﬁeld.

• No Difference Between Boys and Girls and More

Encouragement are highlighted in P1, indicating

the need for a more inclusive and supportive edu-

cational environment.

• All Female Courses and Avoiding Stereotypes

are signiﬁcant for P1, highlighting the impor-

tance of a non-discriminatory learning environ-

ment. Points to the potential beneﬁts of creating a

comfortable learning environment where girls can

thrive without the pressure of gender dynamics.

• The value of Good Tutors/Teachers is acknowl-

edged in Persona 1 (P1), underscoring the impact

of quality education and mentorship.

Recommendations span from structural changes

like early education and all-female classes to more

nuanced approaches like mentorship and stereotype

CSEDU 2024 - 16th International Conference on Computer Supported Education

232

Table 6: Aspirations with computing skills as indicated by

participants.

Perception Category All P1 P2 P3

Obtain a new or better job 31 (22%) 7 (12%) 3 (13%) 21 (37%)

Start a company 14 (10%) 7 (12%) 3 (13%) 4 (7%)

Engage in IT teaching and clubs 10 (7%) 4 (7%) 2 (8%) 4 (7%)

Work on personal projects with friends 10 (7%) 6 (10%) 0 (0%) 4 (7%)

Develop apps, games, and designs 6 (4%) 1 (2%) 2 (8%) 3 (5%)

Pursue a research career and academic advancement 5 (4%) 4 (7%) 0 (0%) 1 (2%)

dismantling. The variation in suggestions among the

personas indicates different needs and perspectives,

suggesting that a one-size-ﬁts-all approach may not

be effective.

3.6 Q6: Aspirations with Computing

Skills

The aspirations and desires that participants have with

their computing skills are varied and ambitious, they

range from securing a new or better job to starting

their own companies and engaging in research. Ta-

ble 6 details the speciﬁc aspirations and the frequency

with which they were mentioned by survey respon-

dents.

This data illustrates a strong connection between

computing skills and personal growth ambitions,

highlighting the empowering nature of technology in

pursuing diverse and meaningful career paths. The

comparative analysis of responses of aspirations with

computing skills across the three personas (P1, P2,

P3) for selected categories reveals the following dis-

tribution:

• The desire for a New/Better Job is particularly sig-

niﬁcant for P3, indicating a strong link between

computing skills and career advancement oppor-

tunities. The high percentage of P3 respondents

looking for job opportunities suggests that many

see computing as a pathway to career improve-

ment.

• The aspiration to Start their Own Company is

more prevalent in P1 and P2, suggesting an en-

trepreneurial spirit among those who transitioned

to computing later in life. P1’s focus on research

and personal projects indicates a deeper engage-

ment with the ﬁeld, beyond just career advance-

ment. P2’s interest in entrepreneurial and cre-

ative endeavours reﬂects their motivation to utilize

computing in versatile and innovative ways.

• Teaching IT is mostly aspired to by P2, indicat-

ing a commitment to the ﬁeld and a desire to con-

tribute to the next generation of computing pro-

fessionals.

• The lack of responses from P2 in Own

Project/Idea Implementation suggests a more tra-

ditional career path, while P1 and P3 show inter-

est in personal and creative projects, indicating di-

verse aspirations in computing.

3.7 Comparative Analysis on Activity

Levels in CS Classroom

The data from the survey indicates that active par-

ticipation in computer science (CS) classrooms is a

strong predictor of continuing to study CS. Among

the different personas, 44% of Persona 1 (P1), who

studied and remained in the ﬁeld of computing, were

actively participating in class. This percentage is no-

ticeably higher compared to the other groups. In con-

trast, only 31% of Persona 2 (P2), who transitioned

to computing later in life, and 22% of Persona 3 (P3),

who never considered entering computing, were ac-

tively participating. These ﬁgures suggest that higher

engagement and active involvement in CS classrooms

are closely linked to sustained interest and pursuit of

studies in the ﬁeld of computing, highlighting the im-

portance of fostering active learning environments to

encourage continued interest in CS.

The survey reveals a notable correlation between

classroom participation levels and attitudes towards

computer science (CS). Students who were passive

in the CS classroom predominantly exhibited more

negative attitudes toward the subject. This passiv-

ity also aligns with their educational choices: 100%

of these students initially pursued disciplines other

than CS, with only 27% eventually studying CS later.

In contrast, among students who were active partici-

pants in the classroom, the trend differs signiﬁcantly.

While 63% of these active participants initially stud-

ied a different discipline, a much higher percentage,

73%, chose to study CS later. This contrast under-

scores the potential impact of classroom engagement

on students’ perceptions and career trajectories. Ac-

tive engagement not only fosters a more positive at-

titude towards CS but also seems to inﬂuence stu-

dents’ decisions to pivot towards CS studies later in

their academic or professional journey, suggesting the

transformative power of an engaging and inclusive CS

educational environment.

3.8 Comparative Analysis on the Age of

First Computer Use

The data on the age of ﬁrst computer use across the

different personas (see Table 7) reveals patterns that

offer insights into their early experiences with tech-

nology.

Predominantly, the larger cohorts within Persona

1 and Persona 2, encompassing individuals who ei-

ther remained in computing or transitioned to it later,

Decoding the Gap: A Retrospective Analysis of Women’s Experiences in Software Engineering

233

Table 7: Frequencies for How old were you when you ﬁrst

used a computer on your own? for different Personas.

P1 P2 P3

More than 14 years 15 (22%) 4 (17%) 10 (18%)

10 to 14 years 12 (17%) 6 (25%) 19 (33%)

7 to 9 years 19 (28%) 10 (42%) 14 (25%)

Less than 6 years 9 (13%) 2 (8%) 5 (9%)

N 59 24 57

encountered computers at a relatively young age, typ-

ically between 7 to 9 years. Nearly half of the respon-

dents within the P1 group (41%) and P2 group (50%)

had their ﬁrst own experience with computers prior to

the age of 9. Conversely, within Persona 3, comprised

of individuals who never contemplated computing as

an area of interest, merely 34% encountered comput-

ers before the age of 9. Among this subset, the ma-

jority had their ﬁrst own experience with computers

at an older age, with 18% encountering computers af-

ter surpassing 14 years, and 33% falling within the

age bracket of 10 to 14 years for their initial expo-

sure. These ﬁndings suggest a potential correlation

between early exposure to computing and the propen-

sity to sustain interest or pursue a career in this ﬁeld,

while delayed exposure might diminish the likelihood

of embracing computing as a career or ﬁeld of study.

Noteworthy is also the relatively high number of

individuals within Personas 1 and 2 who engaged with

computers during the formative years of 10 to 14, in-

dicating that pre-adolescent and early adolescent ex-

posure remains within a pivotal window capable of

fostering enduring interest in computing.

There are fewer individuals across all personas

who used a computer before the age of 6. This might

be due to generational access to technology, as the

youngest age group (less than 6 years old) for early

exposure was less common historically.

Regarding the reasons for ﬁrst computer usage,

the data suggests that games and creative activities

are the most common entry points across all personas.

However, the use of computers for learning and home-

work also has a notable presence, especially in the

context of schoolwork for Personas 1 and 2.

4 DISCUSSION

This discussion seeks to delve into the nuances of the

survey responses, particularly focusing on differences

in perception and aspirations between current com-

puter science (CS) professionals and those not in the

ﬁeld, the impact of age on perceptions of CS, and the

implications of terminology on student engagement.

4.1 Perception of Computer Science and

Stereotypes

Our ﬁndings reveal a complex landscape of percep-

tions associated with computer scientists. Terms

like ’problem-solving,’ ’diverse,’ and ’contributing to

society’ were frequently mentioned across all Per-

sonas, reﬂecting a positive view of the ﬁeld. How-

ever, stereotypes such as ’introversion,’ ’lack of social

skills,’ and ’male dominance’ persistently emerged,

echoing existing literature on the negative impact

of these stereotypes on women’s engagement in CS

(Cheryan et al., 2015; Master et al., 2016, 2021).

The age of respondents appears to inﬂuence their

perceptions, with older participants often holding

more traditional views of the ﬁeld. This could

be attributed to cultural and generational shifts in

how CS is presented and perceived. Moreover, re-

spondents from different professional backgrounds or

study ﬁelds showed varied perceptions, suggesting

that exposure and experience signiﬁcantly shape one’s

view of CS.

Some respondents’ preference for the term ’De-

veloper’ over ’Scientist’ suggests that terminology

may inﬂuence interest in the ﬁeld. This raises the

question of how different terms are used in Ger-

man, such as ’Informatiker/-in’ for computer scientist

and ’Datenwissenschaftler/-in’ for data scientist, and

whether these terms could be affecting student per-

ceptions and potential engagement in CS. Moreover,

it’s worth noting that ’Informatiker’, often equated

with ’Programmierer’ (programmer), might be ap-

pealing to those who are focused on programming.

However, this equivalence can obscure the diverse op-

portunities available in the ﬁeld, potentially deterring

those who might be more motivated by the applica-

tions of programming than programming itself (Mar-

quardt and Happe, 2023). This highlights the need to

carefully consider how the roles and career paths in

CS are communicated and labelled, as these designa-

tions can either narrow or expand students’ views of

the ﬁeld’s possibilities.

4.2 Differences Across Age Groups

Our study indicates distinct preferences and inter-

actions with computing across different age groups.

Younger individuals’ afﬁnity for social media and

older respondents’ preference for activities like online

banking reﬂect evolving technological landscapes and

usage patterns. This variation underscores the impor-

tance of age-speciﬁc approaches in CS education and

outreach.

CSEDU 2024 - 16th International Conference on Computer Supported Education

234

4.3 Drivers and Career Choices

Interestingly, those who view themselves or their fam-

ily as the primary motivation (’biggest driver’) are of-

ten already involved in CS. In contrast, those driven

by ﬁnancial or job prospects tend to be outside the

CS ﬁeld. This dichotomy raises questions about the

obstacles that deter or facilitate a career in CS and

suggests the need for targeted support and guidance

to help individuals overcome perceived barriers.

4.4 Early Exposure to Computing

A recurring suggestion for improving CS education

was to start early. Previous studies have found, that

girls spend on average less time with computers in

their private time at home than boys (Mumtaz, 2001;

Selwyn et al., 2009). This leads to the question of

whether CS careers are considered by children as

early as kindergarten or elementary school and how

these early career aspirations are formed. This gap

suggests a need for educational materials and initia-

tives that introduce young children to CS in engaging

and accessible ways, akin to resources available for

other professions. It would be interesting to explore if

there are educational materials for young children that

introduce CS careers, similar to those for construction

sites or hospitals.

4.5 Aspirations and Knowledge

Barriers

While many participants aspire to leverage CS skills

for career advancement or entrepreneurship, there is

also a notable interest in teaching IT and participat-

ing in coding clubs. This points to a potential pool of

individuals who, despite having basic CS knowledge,

hesitate to engage due to perceived inadequacies. Ad-

dressing these knowledge barriers could foster a more

inclusive and participatory CS community.

4.6 Methodological Limitations

This study’s reliance on self-reported data and its

cross-sectional design pose limitations. Notably,

there were participants who chose not to respond to

certain open-ended questions. This observation war-

rants an analysis of whether the same respondents

consistently abstained across all questions or if non-

responses were speciﬁc to certain queries. Addition-

ally, it is pertinent to consider whether the motivation

to answer free-text questions decreases towards the

end of the survey and how this might affect the data

collected. Furthermore, the self-reported nature of

the data may introduce biases, and the cross-sectional

design precludes causal inferences. Future research

should aim to longitudinally track participants to bet-

ter understand how these factors interact over time to

inﬂuence women’s decisions to enter, stay in, or leave

the ﬁeld of CS.

5 CONCLUSION

The journey toward gender equity in software engi-

neering (SE) is not a mere corrective trajectory to bal-

ance demographic scales; it is a fundamental require-

ment for a robust, dynamic, and innovative digital fu-

ture. This study has illuminated the multifaceted and

often subtle barriers that deter women from partici-

pating in SE. Through a comprehensive analysis, we

have endeavoured to move beyond the surface-level

statistics to understand the deeper currents that inﬂu-

ence women’s decisions regarding SE education and

careers.

Our ﬁndings reveal that the underrepresentation of

women in SE is not a simple case of different pref-

erences or inherent disinterest. It is a consequence

of a complex tapestry of societal perceptions, educa-

tional experiences, and entrenched stereotypes. These

factors coalesce to form a daunting barrier that many

women ﬁnd insurmountable. The stereotype of com-

puting as an isolating and monotonous ﬁeld, domi-

nated by an unwelcoming male majority, continues to

be one of the signiﬁcant deterrents for women. How-

ever, the narrative can be changed. Our study points

to the powerful role of early and positive exposure to

computing in shaping perceptions. Educational insti-

tutions and policy-makers have a critical role in in-

tegrating computing into curricula in a way that is

engaging, relevant, and accessible to all students, ir-

respective of gender. By showcasing the versatility,

creativity, and collaborative nature of SE, we can be-

gin to dismantle the outdated stereotypes that cloud

the ﬁeld. Furthermore, the aspirations of women in

computing—ranging from entrepreneurial ambitions

to social impact projects—highlight the potential loss

to innovation when their talents are not nurtured. En-

couragingly, there is a reservoir of enthusiasm among

women for SE that can be tapped into with the right

support systems and interventions.

In conclusion, the drive toward a more gender-

balanced SE domain should not be pursued out of a

sense of obligation to achieve numerical parity but

from a strategic vision of what a diverse workforce

can accomplish. It is a vision that recognizes the

unique contributions of women to SE, appreciates the

richness that diversity brings to problem-solving, and

Decoding the Gap: A Retrospective Analysis of Women’s Experiences in Software Engineering

235

understands that the future of technology is the bright-

est when it beneﬁts from the talents of the entire pop-

ulation. As we forge ahead, it is imperative that we

commit to creating an environment in SE that is wel-

coming, inclusive, and conducive to the ﬂourishing

of all individuals who wish to be a part of the digi-

tal vanguard. Finally, the power of words in shap-

ing futures cannot be understated: a thoughtful ar-

ticulation of the various roles and career trajectories

in CS/SE, the way these roles are termed and pre-

sented can signiﬁcantly inﬂuence students’ percep-

tions, either limiting or broadening their understand-

ing of their opportunities within the ﬁeld.

ACKNOWLEDGMENTS

This work has been in part supported by Vector

Stiftung, Project “M

adchen f

ur Informatik begeis-

tern“ at Karlsruhe Institute of Technology (KIT), by

the COST Action CA19122 – European Network for

Gender Balance in Informatics (EUGAIN), and by the

Federal Ministry of Education and Research (BMBF).

We also want to thank Professor Anne Koziolek for

her continuous support and valuable comments in our

discussions.

REFERENCES

Albusays, K., Bjorn, P., Dabbish, L., Ford, D., Murphy-Hill,

E., Serebrenik, A., and Storey, M.-A. (2021). The

diversity crisis in software development. IEEE Soft-

ware, 38(2):19–25.

Armstrong, D. J., Riemenschneider, C. K., and Giddens,

L. G. (2018). The advancement and persistence of

women in the information technology profession: An

extension of Ahuja’s gendered theory of it career

stages. Information Systems Journal, 28(6):1082–

1124.

Cheryan, S., Master, A., and Meltzoff, A. N. (2015). Cul-

tural stereotypes as gatekeepers: Increasing girls’ in-

terest in computer science and engineering by diversi-

fying stereotypes. Frontiers in psychology, page 49.

Fidelman, M. (2012). Here’s the real reason there are not

more women in technology. Available online at URL

https://www.forbes.com/sites/markﬁdelman/2012/06/

05/heres-the-real-reason-there-are-not-more-women-

in-technology/.

Google (2014). Women who choose computer sci-

ence – what really matters: The critical role

of encouragement and exposure. Available on-

line at URL https://edu.google.com/pdfs/women-

who-choose-what-really.pdf.

Gorbacheva, E., Beekhuyzen, J., vom Brocke, J., and

Becker, J. (2019). Directions for research on gender

imbalance in the it profession. European Journal of

Information Systems, 28(1):43–67.

Happe, L. and Buhnova, B. (2022). Frustrations steering

women away from software engineering. IEEE Soft-

ware, 39(4):63–69.

Happe, L., Buhnova, B., Koziolek, A., and Wagner, I.

(2021). Effective measures to foster girls’ interest

in secondary computer science education. Education

and Information Technologies, 26(3):2811–2829.

Joshi, K., Trauth, E., Kvasny, L., and McPherson, S. (2013).

Exploring the differences among it majors and non-

majors: Modeling the effects of gender role congruity,

individual identity, and it self-efﬁcacy on it career

choices. In Proceedings of Frontiers in Education:

Computer Science and Computer Engineering.

Lorenzo, R., Voigt, N., Tsusaka, M., Krentz, M., and

Abouzahr, K. (2018). How diverse leadership teams

boost innovation. Boston Consulting Group, 23:112–

134.

Marquardt, K. and Happe, L. (2023). Saving bees with com-

puter science: a way to spark enthusiasm and interest

through interdisciplinary online courses. In Proceed-

ings of the 28th ACM Conference on Innovation and

Technology in Computer Science Education.

Marquardt, K., Wagner, I., and Happe, L. (2023). Engaging

girls in computer science: Do single-gender interdis-

ciplinary classes help? In ICSE ’23: Proceedings of

the 45th International Conference on Software Engi-

neering Association for Computing Machinery, New

York, NY, United States. accepted.

Master, A., Cheryan, S., and Meltzoff, A. N. (2016). Com-

puting whether she belongs: Stereotypes undermine

girls’ interest and sense of belonging in computer sci-

ence. Journal of educational psychology, 108(3):424.

Master, A., Meltzoff, A. N., and Cheryan, S. (2021). Gen-

der stereotypes about interests start early and cause

gender disparities in computer science and engineer-

ing. Proceedings of the National Academy of Sciences,

118(48):e2100030118.

Mumtaz, S. (2001). Children’s enjoyment and perception of

computer use in the home and the school. Computers

& Education, 36(4):347–362.

Rodr

ıguez-P

erez, G., Nadri, R., and Nagappan, M. (2021).

Perceived diversity in software engineering: a system-

atic literature review. Empirical Software Engineer-

ing, 26:1–38.

Selwyn, N., Potter, J., and Cranmer, S. (2009). Primary

pupils’ use of information and communication tech-

nologies at school and home. British journal of edu-

cational technology, 40(5):919–932.

CSEDU 2024 - 16th International Conference on Computer Supported Education

236