Examining gender asymmetry in Wikipedia’s victimization coverage

On Wikipedia, men do things. Women have things done to them.

That was one of the uncomfortable hypotheses that I heard at WikiWomenCamp, a gathering of women Wikimedians in New Delhi in 2023. The claim was provocative but hard to pin down. How do you measure this aspect across millions of articles?

After thinking about this for a while, I got an idea. One could analyse the gender imbalance in articles related to victimization, such as murder, disappearance and kidnapping.

The Murder Test

English Wikipedia has a peculiar class of articles. Search for “Murder of” and you’ll find over 2,000 entries on English Wikipedia: Murder of Hae Min Lee. Murder of JonBenét Ramsey. Murder of Junko Furuta. Each one is about a person who is notable on Wikipedia for a single, terrible reason: they were killed.

These are not articles about famous people who happened to be murdered. Gauri Lankesh was a prominent journalist; her assassination is covered in her biography. But Vandana Das, a young doctor stabbed by a patient in India? Her Wikipedia page exists because of how she died. The article’s title is the crime.

This distinction matters. These “Murder of…” articles are, in effect, a record of which otherwise ordinary people Wikipedia’s editorial community found notable — not for anything they did in life, but solely for how they died. And when you sort them by gender, the numbers tell an interesting story about the gender imbalance in victimization coverage.

What the numbers say

To understand this phenomenon, I scraped the titles of every English Wikipedia article titled “Murder of…”, “Disappearance of…”, “Kidnapping of…”, “Assassination of…”, and similar prefixes. That gave me 3,387 articles. After filtering out non-person entries (yes, “Murder of Crows” is a film, and “Abuse of power” is a concept, not a crime), I was left with 3,048 articles about individual victims.

Then came the hard part: figuring out who was male and who was female. Most of these victims have no Wikidata entry with gender information because they’re ordinary people, not public figures. So I, with the help of some vibe coding with Claude AI, built a three-step pipeline. First, I pulled each article’s text from the Wikipedia API and counted pronouns: does the article say “she” and “her”, or “he” and “him”? That classified about a thousand articles with high confidence. For the rest, we ran the victim’s first name through a gender-detection library that covers names from dozens of cultures. Some of the stubborn remainders, names with garbled Unicode, or culturally ambiguous, were classified manually.

The result, across 2,683 classifiable articles:

Among “Murder of…” articles, 54.6% of victims are female.

That might not sound dramatic until you compare it to the real world. According to the United Nations Office on Drugs and Crime, roughly 80% of homicide victims globally are male. Only about 20% are female.

Wikipedia’s murder articles nearly invert that ratio.

2.7 times over

Put another way: if Wikipedia’s “Murder of…” articles reflected real-world homicide statistics, you’d expect about 370 articles about female victims and 1,470 about male victims. Instead, there are 1,007 about women and 837 about men. Women are overrepresented by a factor of roughly 2.7.

And it isn’t just murder. The skew runs through other types of victimisation articles, such as those on kidnapping and disappearance on English Wikipedia:

Look at the “Disappearance of…” section in the graph: nearly two out of every three articles about a missing person are about a woman. “Kidnapping of…” leans female too. These categories don’t have neat global baselines to compare against the way murder does, but the pattern is consistent: wherever Wikipedia has an article about someone who was harmed, it is more likely to be telling it about a woman.

The one exception makes perfect sense. “Assassination of…” articles are 93.7% male. Assassinations overwhelmingly target heads of state, politicians, and military leaders — roles that have been dominated by men for most of recorded history. This category isn’t really about victimization in the same sense; it’s about political power, which has its own well-documented gender gap.

The skew also isn’t a quirk of any particular era. Among murder articles that could be dated, women outnumber men in every decade from the 1940s to the 2010s. The 1980s are the most extreme, nearly two female-victim articles for every male one, likely reflecting the serial-killer media frenzy of that era, which disproportionately centred on female victims. A detailed breakdown of the data, methods and results can be found here.

So what’s going on?

There are several plausible explanations, and they probably all contribute.

The media pipeline. Wikipedia doesn’t decide on its own what’s notable. Its editors rely on published sources. And media coverage of murders has a well-documented skew. Researchers have a name for it: “missing white woman syndrome,” the observation that cases involving young, attractive, often white female victims receive vastly disproportionate media attention. This creates more citable sources, which makes it easier to meet Wikipedia’s notability standards, which produces more articles.

The true-crime boom. The explosion of true-crime podcasts, Netflix documentaries, and Reddit communities over the past two decades has overwhelmingly centered on female victims. These cultural phenomena make certain cases known, and known cases get Wikipedia articles. The feedback loop between popular media also accounts for the proliferation of crime articles with female victims.

Who writes Wikipedia? Wikipedia’s editor base is estimated to be approximately 80% male. It’s possible that predominantly male editors find cases involving female victims more compelling or more tragic, and are therefore more likely to create and maintain those articles. Meanwhile, the murder of a man in a bar fight, however tragic, may not trigger the same editorial impulse.

Deletion politics. I didn’t test this, but it’s worth investigating: are articles about male murder victims more likely to be flagged for deletion? Wikipedia’s “Articles for Deletion” process involves community consensus, and if editors implicitly apply different notability thresholds to male and female victims, the resulting article set would reflect that bias.

The bigger picture

This finding snaps into place alongside something researchers have known for over a decade. Studies of Wikipedia’s biographical articles have consistently found that women are underrepresented, accounting for roughly 20% of all biographies. But within those articles, women’s entries are more likely to mention their relationships, and their family status.

Now add this: women are overrepresented in an entire category of articles defined by victimhood. If you’re a woman, you’re underrepresented among Wikipedia’s biographies of notable individuals. But you’re overrepresented among its murder victims, kidnapping victims, and missing persons.

Caveats, Because this is complicated

A few important qualifications. Our classification pipeline has an estimated 3-5% error rate, mostly on culturally ambiguous names. About 12% of articles were not classified at all. This small study only looked at English Wikipedia — other language editions may differ. The comparison with UNODC data is broad; global homicide rates vary enormously by region, and Wikipedia’s coverage skews toward English-speaking countries. And the study didn’t control for the type of homicide, which matters: domestic violence victims (disproportionately female) may generate more media interest than gang-related killings (disproportionately male).

Here, I also want to be precise about the claim. This analysis shows that Wikipedia disproportionately documents female murder victims, not that it intentionally victimizes women. The bias likely originates upstream, in the media ecosystem that determines which murders become well-known enough to warrant an encyclopedia article.

What next?

There are several obvious follow-ups. First, do articles about female victims tend to be longer and more detailed than those about male victims? If the victimization framing hypothesis is right, we might expect not just more articles about women but more vivid, more emotionally detailed articles. Second, an analysis of Wikipedia’s Articles for Deletion logs could reveal whether male-victim articles are deleted at higher rates. Third, extending this analysis to other language editions would show whether the pattern is specific to English Wikipedia.

Finally, there’s the question of what, if anything, should be done about it. Deliberately creating articles about male murder victims to “balance” the ratio would be absurd. But awareness of the pattern might prompt editors to think twice about which stories they choose to tell, and which they leave untold.

---

The author is a volunteer Wikimedian and researcher whose research on Wikimedia primarily focuses on its gender gap.

The full dataset, classification code, and methodology are available at github.com/nethahussain/wikipedia-victimization-gender-analysis.

This research was motivated by discussions at WikiWomenCamp. Data sources: English Wikipedia (via Quarry) and UNODC Global Study on Homicide.