Gender imbalances of retraction prevalence among highly cited authors and among all authors

This bibliometric analysis of over 10 million authors reveals that while overall gender differences in retraction rates are minimal, significant field-specific disparities exist, with structural factors such as discipline, career age, country income, and publication volume being stronger predictors of retraction than gender.

The Gist

Imagine the world of science as a massive, bustling library where millions of people write books (research papers) to share new ideas. Sometimes, a few of these books turn out to have serious errors or even fake information, so the library has to pull them off the shelves and put a big "RETRACTED" stamp on them. This is called a retraction.

The researchers behind this study wanted to answer a very specific question: Are men or women more likely to have their books stamped "RETRACTED"?

Here is the story of what they found, broken down simply:

1. The Two Groups They Checked

The team looked at two different crowds in this giant library:

• The Superstars: The top 2% of scientists who are cited (mentioned) by almost everyone else. These are the famous authors with huge reputations.
• The Regulars: Everyone else who has written at least five books.

2. The Big Reveal: It's Not About Gender

If you were to guess, you might think one gender is much more likely to make mistakes than the other. But the study found that overall, men and women make mistakes at almost the exact same rate.

• Among the Superstars, about 3 out of every 100 men and 3 out of every 100 women had a retracted paper.
• Among the Regulars, it was even lower: less than 1 out of every 100 for both groups.

The Analogy: Think of it like a driving test. If you look at all the drivers, men and women crash their cars at roughly the same frequency. Being a man or a woman doesn't make you a "bad driver" by itself.

3. The Twist: It Depends on Where You Drive

While the overall numbers were the same, the story got interesting when they looked at specific fields of study (like driving in different types of weather).

• In Biology, Medicine, and Psychology: Women were actually less likely to have retractions than men. It's like women driving safely in heavy rain while men drove a bit more recklessly.
• In Economics, Engineering, and Tech: The opposite happened. Women had slightly higher retraction rates here.

4. The Real Culprits: Experience, Location, and Volume

The study found that gender wasn't the main reason for retractions. Instead, three other factors were the "villains" driving the numbers:

• Career Age (How long you've been driving): Newer scientists (those who started publishing recently) had higher retraction rates than veterans. It's like new drivers having more fender-benders while old pros have a better track record.
• Country Income: Scientists from countries with lower incomes had different retraction patterns compared to those in wealthy nations.
• Publication Volume: The more books you write, the higher the chance one of them might have an error. If you write 1,000 books, you are statistically more likely to have a bad one than someone who writes 10.

5. The "Younger Cohort" Surprise

Among the famous Superstars, the gap between men and women is actually growing wider for the younger generations.

• For scientists who started their careers recently (after 2011), men had a much higher retraction rate (8.7%) compared to women (4.9%).
• It seems like the "new generation" of male superstars is making more errors than their female counterparts, but this is likely due to the pressure to publish quickly or the specific fields they are entering, not because they are men.

The Bottom Line

The study concludes that gender isn't the main driver of scientific errors. You can't look at a scientist's gender and predict if they will have a retraction.

Instead, the "weather" of their career matters more:

• What field are they in?
• How long have they been working?
• Where are they located?
• How many papers are they publishing?

In short: Science is a complex game. While men and women play by the same rules and make mistakes at similar rates overall, the specific rules of the game change depending on which "stadium" (field) you are playing in and how much experience you have. The focus should be on fixing the structural issues (like pressure to publish too fast) rather than blaming gender.

Technical Summary

Technical Summary: Gender Imbalances in Retraction Prevalence

1. Problem Statement
Scientific retractions, while historically rare, have seen a rising trend in recent years. While previous research has integrated retraction data into bibliometric databases to link retractions with individual impact metrics, a critical gap remains regarding the demographic drivers of these events. Specifically, it is unclear whether gender disparities exist in the likelihood of scientists having their work retracted. This study investigates whether women and men differ in their retraction prevalence, both within the elite cohort of highly cited scientists (top 2%) and across the broader population of active researchers.

2. Methodology
The study employed a descriptive cross-sectional bibliometric analysis utilizing a large-scale Scopus-based authors database.

• Cohorts: The analysis covered two distinct groups:
  • Highly Cited Authors: $n = 217,097$ (defined as top 2% career-long impact).
  • All Other Authors: $n = 10,361,367$ (filtered for $\ge$ 5 publications).
• Gender Assignment: Gender was inferred using NamSor, a probabilistic tool. To ensure data quality, only assignments with a confidence score $>85\%$ were retained, resulting in a classified sample of 8,267,888 scientists.
• Statistical Approach:
  • Descriptive Analysis: Calculated gender-specific retraction rates and the Prevalence Ratio ($R$) of women versus men.
  • Stratification: Data was stratified by publication age, academic field, country income level, and publication volume.
  • Multivariable Modeling: Logistic regression models were fitted to estimate adjusted associations, reporting Odds Ratios (OR) with 95% Confidence Intervals (CI) to control for confounding variables.

3. Key Contributions

• Scale and Granularity: This represents one of the largest analyses of retraction demographics to date, distinguishing between elite scientists and the general research population.
• Contextual Nuance: Rather than providing a monolithic view, the study dissects retraction rates across specific disciplines, career stages, and geopolitical contexts.
• Methodological Rigor: By combining high-confidence gender inference with multivariable regression, the study isolates the effect of gender from other strong predictors of retraction (such as field and publication volume).

4. Key Results

• Overall Prevalence:
  • Highly Cited Authors: 3.3% of men and 2.9% of women had at least one retraction.
  • All Authors: The retraction rate was identical for both genders at 0.7%.
• Field-Specific Disparities: Significant heterogeneity exists across disciplines:
  • Lower Rates for Women: In Biology, Biomedical Research, and Psychology, women's retraction rates were at least one-third lower than men's ($R < 0.7$).
  • Higher Rates for Women: In Economics, Engineering, and Information and Communication Technologies (ICT), women's rates were higher than men's ($R > 1.3$).
• Temporal and Geographic Trends:
  • Career Age: Among highly cited authors, younger cohorts showed widening gaps. For those starting post-2011, 8.7% of men vs. 4.9% of women had retractions.
  • Geography: Pronounced country-level variations were observed (e.g., in Pakistan, 28.7% of men vs. 14.3% of women had retractions).
• Multivariable Analysis:
  • After adjusting for covariates, the association between gender and retraction was minimal to non-existent.
  • All Authors: OR = 0.96 (95% CI: 0.94–0.97).
  • Highly Cited Authors: OR = 0.96 (95% CI: 0.89–1.03).
  • Strong Predictors: Career age, country income level, publication volume, and specific academic disciplines were identified as significantly stronger correlates of retraction than gender.

5. Significance and Conclusion
The study concludes that gender is not a primary driver of scientific retractions when controlling for structural and contextual factors. While raw data suggests minor differences, these are largely explained by the distribution of men and women across different fields, career stages, and national contexts.

The findings shift the focus of retraction analysis away from individual demographic traits (like gender) toward systemic and structural factors. The variation in retraction rates is better explained by the specific risks associated with certain disciplines, the pressure of high publication volume, and the research environments of different countries. Consequently, efforts to understand and mitigate retractions should prioritize these contextual drivers rather than attributing disparities to gender.