NEW YORK – A recent article in Cosmopolitan addresses a vexing question which has plagued humankind for centuries: Where the fuck did I put my car keys?

Wait, wait… Sorry, that’s my question, not the one Cosmo addressed. Cosmo’s question is “Why do so many straight women prefer girl-girl porn?”

According to the article, “almost a quarter of straight millennial women prefer girl-on-girl pornography to porn that depicts their actual sex lives.”

As a middle-aged and admittedly pervy straight man, I applaud this news with a thunderous one-handed clap. As a sex-stat skeptic, however, I must ask: Is it true?

In addressing the question, the first thing we must to do as Amateur Social Scientists and Junior Internet Detectives, naturally, is assess the reliability of the source data that provides the foundation of the statistical claim.

“According to data sourced by Pornhub’s research and analysis team, the most popular category among the site’s female viewers is ‘lesbian.’”

Well, then, case definitively closed.

I mean, this is a hard fact provided by the Center for Transparently Self-Promotional Data at the esteemed University of Pornhub, Montreal—which, as I’m sure you’re aware, recently received the coveted “Fully Erect” rating from Times Higher Education.

Even with such a venerated data source, however, I’m sorry to report there is one potential issue here.

The Cosmo article fails to take into account what experts in internet demographic data analysis call der Geinderfugenfactor, which loosely translates as “the percentage of chicks on the internet who are really just dudes pretending to be chicks on the internet.”

The term traces back to Technische Universitaet Muenchen, where researcher Dr. Karl von Schwanzenharder first scrutinized the reliability of self-reported information provided via the internet in his watershed 1999 research paper “Im Ernst: Wie viele Milliardäre mit Lear Jets Kann es tatsächlich sein kann zu einem bestimmten Zeitpunkt?”

While the primary focus of Schwanzenharder’s initial research was the question of whether internet users routinely overstate their accumulated wealth and/or current income, he later expanded the scope of his analysis to include questionable information provided by web users concerning their age, military background, personal connections to major news stories of the day—and their gender.

Schwanzenharder’s second groundbreaking study, “Warum tue alle Frauen, die ich treffen Online Have Penisse?” uncovered a shocking truth: Many so-called “women” on the internet were anything but ladylike.

“Based on empirical data gathered by our research staff, we estimate up to 67 [percent] of people on the internet purporting to be female are, in fact, male—or at the very least, in possession of male genitalia,” Schwanzenharder noted in the English translation of the paper’s abstract.

Granted, Schwanzenharder’s respondent pool was not exactly a “random sample,” as it was composed primarily of individuals who had rebuffed his approaches on Match.com, but it did establish some important benchmarks regarding the trustworthiness of information people self-report online.

Later attempts to derive a more precise measure of Geinderfugenfactor were undertaken by researchers at MIT, Stanford and the Taurida National University at Simferopol’s Center for Determining Which Internet Users are Worth Targeting with Email-Delivered Exploits.

Given the difficulty of gathering accurate data from someone determined to provide you with inaccurate information, it’s no surprise the various research teams who have tried to pin down the Geinderfugenfactor have only been able to come up with an estimated range, as opposed to a single data point average on which we can all depend.

Ultimately, given all the uncertainties and unknowns, the researchers concluded somewhere between one-quarter and five-eighths of all users who are not women pretend to be women on the internet, while between two-thirds and three-fourths of users who are women do not pretend to be men.

So, the next time you read an article claiming that almost a quarter of women prefer one thing or another, and the data underpinning the claim was gathered via the internet, just remember: What you’re really being told is somewhere between 5 percent and 12.5 percent of women who aren’t men prefer to watch women who aren’t lesbians have sex with other women who may or may not be lesbians.

Or something like that.