💻 computer science

A Critical Discourse Analysis of Gender Representation in Software Engineering Education Videos on YouTube

This study employs critical discourse analysis of 200 YouTube software engineering tutorials to reveal that male dominance in technical roles and the use of masculine linguistic defaults create a hidden curriculum that reinforces gendered norms and acts as a symbolic barrier to women's entry into the field.

Original authors: Isabella Graßl, Alexander Serebrenik, Giuseppe Destefanis

Published 2026-06-19

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Isabella Graßl, Alexander Serebrenik, Giuseppe Destefanis

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the internet as a giant, bustling library where millions of people go to learn how to build software. In this library, there are thousands of video tutorials on YouTube acting as the librarians. You might think these librarians are just handing out instruction manuals on how to write code, but this study suggests they are also handing out something invisible: a "hidden rulebook" about who belongs in the world of software engineering.

The researchers, Isabella Graßl, Alexander Serebrenik, and Giuseppe Destefanis, decided to peek behind the curtain. They didn't just count how many men and women appeared in the videos; they used a method called Critical Discourse Analysis. Think of this as a high-powered microscope that doesn't just look at the words on the page, but examines the tone, the pictures, and the unspoken messages to see what kind of reality the videos are building for the viewer.

Here is what they found, broken down into simple concepts:

1. The "Default" Setting is Male

The study looked at 200 videos (100 in English, 100 in German). They found that the "default setting" for a software engineer in these videos is almost always a man.

The Name Game: When a video introduces a character to explain a concept (like a "Bank Manager" or a "System Admin"), that character is overwhelmingly named "Bob," "Thomas," or "Heini." In the German videos, for every single woman named, there were nearly five men. In English, it was better, but still skewed (almost two men for every one woman).
The Pronoun Trap: In German, the language has a grammatical rule called the "generic masculine." It's like saying "The doctors are here" to mean all doctors, even if some are women. The study found that in German tutorials, this rule is used so heavily that it acts like a fog, making women invisible. Even when the video is about a generic user, the language defaults to "he." In English, creators tried harder to use neutral words like "they," but the bias was still there in other ways.

2. The "Action Hero" vs. The "Sidekick"

The researchers looked at Agency, which is a fancy word for "who gets to do the important stuff."

The High-Agency Roles: These are the roles that make decisions, build the code, and solve the hard problems (like the Architect or the Lead Developer). In the videos, these roles were almost exclusively played by men.
The Low-Agency Roles: These are the passive roles, like the "Customer," the "User," or the "Secretary." When women did appear in the videos, they were almost always stuck in these sidekick roles. They were the ones using the software, not the ones building it.

The Analogy: Imagine a cooking show. If the host is always a man chopping vegetables and seasoning the soup (the "builder"), and the only woman on screen is the one tasting the food or setting the table (the "user"), you start to believe that cooking is a man's job and eating is a woman's job. That is exactly what these software tutorials are doing.

3. The "Hidden Curriculum"

The paper argues that these videos have a Hidden Curriculum. This is like the unspoken rules you learn just by watching a show, rather than the official lesson plan.

The Official Lesson: "Here is how you write a Java loop."
The Hidden Lesson: "The person who writes Java loops is a man. The person who uses the app is a woman."

Even if the video creator doesn't intend to be sexist, the pattern of showing men as the experts and women as the consumers sends a powerful signal to female students: "This space isn't really for you."

4. English vs. German: A Tale of Two Languages

The study found a difference between the two languages, but the problem existed in both:

German: The bias was very loud and obvious. The grammar itself (using "he" for everyone) combined with male names created a wall that was hard to climb.
English: The bias was quieter. Creators used neutral words like "they" or "the user," which is good. However, the visuals and the roles still leaned heavily toward men. Even when they tried to be neutral, the "mental model" of the expert was still a guy.

5. Why This Matters

The researchers aren't saying these videos are "bad" tutorials. They are technically accurate. But they are like a map that only shows half the territory. If a young girl watches these videos and never sees a woman building the software, she might subconsciously decide that software engineering isn't her path.

The paper concludes that to fix this, we don't need to rewrite the code. We just need to rewrite the story. We need to see more women as the "Architects" and "Developers" in the examples, not just the "Customers." It's about making sure the invisible rulebook tells everyone, "You belong here," rather than just a specific group of people.

Technical Summary: A Critical Discourse Analysis of Gender Representation in Software Engineering Education Videos on YouTube

Problem Statement
Educational resources play a critical role in framing students' perceptions of belonging within a discipline. In software engineering (SE), where a persistent gender gap exists, "hidden curricula"—implicit messages regarding identity and belonging embedded in examples, language, and context—may reinforce stereotypes that exclude female learners. While prior research has examined gender bias in textbooks and specific computing tropes, there is a significant gap in understanding how these norms are reproduced in online learning spaces, specifically YouTube tutorials. Furthermore, the impact of linguistic structures, such as the German generic masculine versus English gender-neutral pronouns, on the construction of technical identity remains under-explored in this medium.

Methodology
The authors conducted a novel exploratory study utilizing Critical Discourse Analysis (CDA) to examine gender representation in SE education.

Data Collection: A pipeline using the YouTube Data API and youtube-transcript-api collected 1,225 videos across English, German, and Italian.
Sampling: A stratified random sample of 200 videos (100 English, 100 German) was selected for manual analysis. Italian videos were excluded from the manual coding phase due to resource constraints but retained in the dataset for future replication. The sample covers videos from 2010–2025, with a focus on recent content (66% from 2020–2025).
Analytical Framework: The study employed a qualitative content analysis based on Mayring's approach, guided by a framework comprising four dimensions:
1. Contextual Domains: Categorizing examples into Business/Commercial, Technical Only, Social/Entertainment, and Public/Education.
2. Identity Markers: Analyzing named characters, role references (pronouns linked to titles), and generic defaults.
3. Bias Indicators: Identifying patterns such as "Male-Centric Technicality" (all experts are male), "Generic Masculine Default," and "Passive Role Assignment" (females in support roles).
4. Gender Representation Index: A scoring system ranging from -1 (inclusive) to 3 (strongly biased) to assess overall discursive content.
Reliability: Two researchers independently coded a subset of videos, achieving an inter-rater agreement of Cohen's $\kappa = .87$ . The first author then coded the remaining videos, with discussions held for borderline cases.

Key Results
The analysis reveals a pervasive "agency gap" where technical authority is disproportionately assigned to male actors, with significant variations between linguistic contexts.

Dominance of Male Characters: Male actors dominate the tutorials. In the German sample, male entities account for over 80% of named characters. In the English sample, while more diverse, male entities still outnumber female entities in named roles (1.8:1 ratio).
Linguistic Divergence:
- German: Heavily relies on the generic masculine. 78.69% of videos use the generic masculine as the default for unspecified identities. When specific roles are named, the male-to-female ratio for high-agency roles is 10.4:1.
- English: Shows higher surface-level neutrality, with 53 videos using purely neutral language (e.g., singular "they"). However, when gendered pronouns are used, they remain skewed toward male defaults for technical roles.
Agency Gap: A critical finding is the distribution of agency. High-agency roles (decision-makers, architects, developers) are almost exclusively coded as male. Female actors are frequently absent or relegated to low-agency roles (users, customers, administrative support). In German videos, the male-to-female ratio for low-agency roles is 18.4:1, indicating women are rarely used even as passive placeholders.
Contextual Framing: German tutorials are predominantly framed within Business/Commercial contexts (59%), whereas English tutorials show a more balanced split between Business and Technical domains. Both languages underutilize Education/Public and Entertainment domains.
Bias Scores: The average bias score for German tutorials was 0.77 (indicating Minor to Clear Bias), primarily driven by the generic masculine and male-centric technicality. English tutorials averaged 0.16 (Balanced), though 14.67% still exhibited bias indicators, often through the use of "bros" or male pronouns for peers.

Key Contributions

Empirical Evidence: This study provides the first empirical evidence of gender representation patterns in SE education materials on YouTube, moving beyond textbook analysis to dynamic video content.
CDA Application: It successfully applies Critical Discourse Analysis to technical education, demonstrating how linguistic choices (pronouns, naming) and visual representations construct a "hidden curriculum" that associates technical authority with masculinity.
Cross-Linguistic Comparison: The study highlights how grammatical structures (e.g., German generic masculine vs. English neutral "they") interact with cultural norms to produce different manifestations of the same underlying bias: the erasure of female technical agency.
Replication Package: The authors have made the full dataset, stratified sample, and codebook publicly available to support future verification and extension of this work.

Significance and Claims
The authors claim that software engineering education on YouTube may be reproducing gendered norms that serve as symbolic barriers to entry for non-male learners. The study argues that the "hidden curriculum" in these videos signals to female students that they do not belong in technical roles, potentially contributing to the field's gender gap.

The paper posits that these findings are not merely a "soft" humanities issue but a matter of technical and professional accuracy. By identifying the discursive patterns that reinforce a "dude culture," the authors suggest that creators can improve representational accuracy by using diverse naming conventions, neutral pronouns, and balanced role assignments. The study concludes that while no explicit sexism (Strong Bias) was found, the systemic, subtle reproduction of male-centric mental models requires attention to create a learning environment where all learners feel they belong. Future work is suggested to examine the causal effects of these representations on student self-efficacy and to expand the analysis to non-binary identities and other linguistic contexts.