💻 computer science

It's Not About Whom You Train: An Analysis of Corporate Education in Software Engineering

This study of 282 Brazilian software engineers reveals that personal sociodemographic profiles do not significantly influence perceptions of corporate training quality, whereas professional trajectory factors and the voluntariness of participation are the primary determinants of perceived effectiveness.

Original authors: Rodrigo Siqueira, Danilo Monteiro Ribeiro

Published 2026-04-09

📖 6 min read🧠 Deep dive

CC0 1.0

Original authors: Rodrigo Siqueira, Danilo Monteiro Ribeiro

Original paper dedicated to the public domain under CC0 1.0 (http://creativecommons.org/publicdomain/zero/1.0/). ✨ 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 a massive software company as a giant, high-tech gym. They spend a fortune hiring personal trainers (instructors), buying the latest equipment (courses), and building state-of-the-art facilities (training programs) to help their employees get stronger, faster, and smarter.

But here's the big question the researchers asked: Does it matter who walks through the gym doors?

Does a 25-year-old woman perceive the workout differently than a 45-year-old man? Does someone with a PhD see more value than someone with a bachelor's degree? Does the size of the company they work for change how they feel about the training?

The researchers, Rodrigo and Danilo, decided to investigate this by asking 282 Brazilian software engineers, "How good was your last training session?" They then cross-referenced those answers with everything they knew about the people: their age, gender, experience, job title, and whether they had to take the class or chose to take it.

Here is what they found, explained simply:

1. The "Who" Doesn't Matter (The Gym Membership Myth)

You might think that a woman would feel differently about the training than a man, or that a junior employee would have a totally different experience than a senior one.

The Finding: Surprisingly, it didn't matter.

Gender: Men and women rated the training almost exactly the same.
Education: A Master's degree holder felt the same as someone with just a Bachelor's.
Company Size: People at giant tech giants felt the same as those at medium-sized firms.

The Analogy: Imagine a restaurant serving a steak. The researchers found that whether you are a vegetarian, a meat-lover, a chef, or a food critic, if you are sitting at the table eating the steak, you all agree on how it tastes. The "personal profile" of the person doesn't change the flavor of the meal.

The Twist: The fact that women rated it the same as men is actually a clue about a bigger problem. Since there are far fewer women in the industry (only 23% of the group), it suggests that the "barrier" isn't inside the training room. The barrier is getting into the gym in the first place. Once women make it past the door and get the training, they enjoy it just as much as the men. The problem is access, not the experience itself.

2. The "When" Matters A Lot (The Career Rollercoaster)

While who you are didn't matter, where you are in your career made a huge difference. The researchers found a strange, non-linear pattern based on years of experience.

The Finding:

0–2 Years (The Newbies): They loved the training. It was fresh, exciting, and everything was new.
3–6 Years (The "Dead Zone"): This is the weird part. People with 3 to 6 years of experience were the least engaged. They felt the training was boring or useless.
7–8 Years (The Recovery): Suddenly, interest bounced back up.
8+ Years (The Veterans): Interest dropped again.

The Analogy: Think of a video game.

Level 1-2 (Newbies): You are learning the controls. Every tutorial is a "Aha!" moment.
Level 3-6 (The Dead Zone): You know the basics. The training is teaching you things you already know, but it's not teaching you the "boss fight" strategies yet. It feels like a waste of time. You are in the "middle game" where the tutorial is too easy, but the advanced stuff hasn't started.
Level 7-8: You are finally ready for the advanced levels, so the training becomes useful again.
Level 9+ (Veterans): The standard training is too basic for you again. You need a custom, expert-level strategy guide, not a generic manual.

The Lesson: Companies shouldn't treat all employees the same. They need to stop giving "Level 1" training to "Level 5" players.

3. The "Mandatory" Trap (The Forced Gym Session)

The single biggest factor that changed how people felt about the training was whether they were forced to go.

The Finding:

If you chose to go: You thought the training was great, relevant, and you learned a lot.
If you were forced to go: You thought it was a waste of time, boring, and didn't help you.

The Analogy: Imagine you love pizza.

If you choose to go to a pizza party, you have a blast.
If your boss forces you to go to a pizza party, you start thinking, "Why am I here? This is annoying. I'd rather be working."
The pizza (the training) is the same, but your willingness to eat it changes the whole experience.

4. The Missing Piece: Soft Skills for Techies

The researchers noticed a gap in what different job roles were getting.

Managers got lots of training on "Soft Skills" (communication, leadership, teamwork).
Architects and Senior Developers (the technical wizards) got almost no soft skills training.

The Analogy: It's like a sports team where the Coach gets a seminar on "How to motivate the team," but the Star Player (who has to talk to the referee, the other team, and the fans) gets no training on communication. The Star Player is left trying to negotiate a contract or explain a technical decision without the right tools.

The Bottom Line

The title of the paper says it all: "It's Not About Whom You Train."

If you are a company trying to improve your training:

Stop worrying about demographics: Don't assume a woman or a person with a specific degree needs a different kind of class.
Focus on the Career Stage: Don't give a "New Hire" class to someone with 5 years of experience. They are in the "Dead Zone" and need something harder.
Make it Optional: If you force people to learn, they won't learn. Let them choose.
Teach the Techies Soft Skills: Your best coders need to learn how to talk and lead, not just code.

In short, the quality of the training depends less on who is sitting in the chair and more on when in their career they are sitting there, and whether they wanted to be there in the first place.

1. Problem Statement

Despite significant investments in corporate training within the software industry, organizations often lack mechanisms to assess the effectiveness of these programs or understand how different professional profiles perceive them. While previous research has identified what makes training effective (e.g., instructional design), there is a gap in understanding who the training is for. Specifically, it remains unclear whether sociodemographic variables (gender, age, education) and professional trajectory variables (experience, role, area of work) significantly influence the perception of training quality and effectiveness. This study addresses the Research Question (RQ): To what extent do sociodemographic and professional variables influence the perception of quality and effectiveness of corporate training in Software Engineering (SE)?

2. Methodology

The study constitutes the third phase of a multiphase research effort, utilizing a quantitative analysis of a dataset collected from 282 Brazilian software engineering professionals.

Data Collection: A survey was conducted online between September and November 2025. Participants were recruited via convenience sampling (LinkedIn, WhatsApp, Telegram) and met specific inclusion criteria (18+, working in SE, participated in corporate training within the last 12 months).
Instrument: The survey comprised:
- 9 Sociodemographic/Professional Variables: Gender, age, education level, state, experience length, professional level, company size, area of work, and nature of participation (mandatory vs. optional).
- 27 Perception Items: Likert-scale items (1–5) organized into four dimensions of the Salas Framework (Needs Analysis, Antecedent Conditions, Instructional Methods, Post-Training Conditions).
Statistical Analysis:
- Approach: A systematic significance analysis crossing 27 perception items with 9 variables, resulting in 243 combinations.
- Tests: Non-parametric tests were used due to ordinal data and lack of guaranteed normality.
  - Kruskal-Wallis: Applied to categorical variables (gender, experience, level, etc.).
  - Spearman Correlation: Applied to the continuous variable (age).
- Post-Hoc: Dunn's test with Bonferroni correction was used for pairwise comparisons where significance was found.
- Effect Size: Calculated using Epsilon-squared ( $\epsilon^2$ ) for Kruskal-Wallis and $r$ for Spearman.
- Significance Level: $\alpha = 0.05$ .

3. Key Contributions

Empirical Evidence on Demographics: Provides data-driven evidence that personal profile variables (gender, education, age) and company size do not significantly determine the perception of training quality and effectiveness in SE.
Non-Linear Experience Pattern: Identifies a specific "low-engagement zone" in the relationship between professional experience and training impact, occurring between 3 and 6 years of experience.
Gender Barrier Hypothesis: Offers a novel interpretation that the absence of gender differences in training perception suggests that gender barriers in SE operate before training (in access and representation) rather than during the learning experience itself.
Strategic Segmentation: Proposes that training programs should be segmented by career stage rather than personal demographics.

4. Key Results

A. Dominance of Training Nature (Mandatoriness)

The most significant finding was that the nature of participation (mandatory vs. optional) was the dominant factor.

24 out of 27 perception items showed statistical significance based on whether training was mandatory.
Mandatory training consistently resulted in lower scores for motivation, relevance, and overall satisfaction compared to optional training.
This confirms previous findings that voluntariness favors autonomous motivation and learning transfer.

B. Absence of Personal Profile Differences

Gender: No significant differences were found between men (76.6%) and women (23.0%) across any of the 27 items.
Education & Company Size: No significant differences were observed based on education level or company size.
Interpretation: The study suggests that once individuals overcome entry barriers to the industry and access training, their perceptions converge, regardless of gender or educational background.

C. Professional Trajectory: The "Low-Engagement Zone"

Professional experience was the second most influential factor, revealing a non-linear descriptive pattern:

0–2 Years: High perception of impact (likely due to onboarding and foundational learning).
3–6 Years: A distinct drop in perceived value. This group (often mid-level) finds generic training superfluous as they have passed the initial learning curve but do not yet face the complex demands of senior roles.
7–8 Years: A transient recovery in perceived value (likely transitioning to senior/leadership roles).
8+ Years: A decline again, suggesting standard training exhausts its utility for specialists.
Soft Skills Gap: Significant differences were found by area of work. Architecture roles reported the lowest focus on soft skills (mean 2.10), while Management roles reported the highest (3.95), indicating a gap in behavioral training for advanced technical roles.

D. Regional and Age Nuances

Region: Differences were found in structural conditions (e.g., training during working hours, relevance to competitiveness) rather than instructional quality.
Age: A weak positive correlation ( $r=0.147$ ) was found between age and the perception of training impacting personal time, suggesting older professionals are more sensitive to off-hours training.

5. Significance and Implications

Shift in Focus: The study challenges the assumption that training must be tailored to personal demographics (e.g., gender-specific programs). Instead, it argues that instructional design and career stage are the critical determinants of effectiveness.
Diversity & Inclusion Insight: The lack of gender disparity in perception (despite a 23% female sample) implies that the primary barrier for women in SE is access to the profession and training opportunities, not the quality of the training itself once accessed.
Curriculum Design: Organizations should avoid "one-size-fits-all" training. Programs should be segmented by career stage, specifically addressing the "3–6 year" gap with advanced, context-specific content rather than generic updates.
Voluntariness: The strong negative impact of mandatory training suggests organizations should foster a culture of voluntary upskilling to maintain high engagement and perceived value.

6. Limitations

Sample Bias: The sample is predominantly male, highly educated, and concentrated in large Brazilian companies, which may limit generalizability to other cultural contexts or smaller organizations.
Self-Reported Data: Reliance on subjective perception rather than objective metrics (e.g., code quality, productivity).
Training Type Confounding: The study did not differentiate between types of training (e.g., onboarding vs. compliance), which may confound the experience-based results.
Multiple Comparisons: With 243 tests, there is a risk of false positives, though Bonferroni correction was applied to post-hoc tests.

Conclusion

The paper concludes that "It's not about whom you train, but at what point in their career they are trained." The perception of corporate education effectiveness in Software Engineering is driven primarily by the voluntariness of participation and the alignment of content with the professional's career stage, rather than their sociodemographic profile.