Using your browser url history to estimate gender

http://www.mikeonads.com/2008/07/13/using-your-browser-url-history-estimate-gender/

by drp • 16 years ago

35 21 16 years ago

In our house you could nail it a lot less scientifically. Just scan the domain names and add 1 point for male with a /porn/ match and 1 female point for /baby/. I dare say the Google seaerches would be an immediate giveaway too!

borisk• 16 years ago

JS doesn't have direct access to browser history, it can only find if an exact URL is visited or not. Even that will soon go.

sp332• 16 years ago

I think it already doesn't work in the new Firefox 4.0 beta, or the Chrome dev builds.

d0m• 16 years ago

I'm about 60% male but hey, my girlfriend use my computer sometime :D

languagehacker• 16 years ago

Yeah, not only are they probably using two-year-old data, but if the confidence given is 97% for it being the wrong gender for me, I doubt that it's actually an effective tool.

There are all kinds of fuzzy behavioral ways to analyze gender. I remember when the Gender Genie (http://bookblog.net/gender/genie.php) made a similar splash. The issue is that it ignores the finer-grained sociological contexts in how certain people behave.

I can't find their name, but there is an African tribe where, as a courtship ritual, men dress up and wear makeup to attract women. To other people, this would sound like gender roles are reversed, but in reality, it's just their cultural definition of gender roles at play.

Mainstream culture subsumes a variety of subcultures and idioms, meaning that using statistical methods to determine the characteristics of specific behaviors within a coarse-grained social group may overgeneralize. For a more nuanced approach, it would be a good idea to use ensembles, or add subculture membership as a feature to a linear regression classifier.

jedbrown• 16 years ago

Apparently scientific writing is astoundingly gender-neutral, Gender Genie reports the following for my last paper:

    Female Score: 3450
    Male Score: 3259

hugh3• 16 years ago

if the confidence given is 97% for it being the wrong gender for me, I doubt that it's actually an effective tool.

Not necessarily. Of all those it predicts with a 97% confidence interval, you'd expect it to be wrong 3% of the time. Without any information other than that it was wrong about you, there's not much evidence about how accurate it is.

Men dress up to attract women in just about every culture. The fact that they also wear makeup shouldn't be surprising.

My, what a lot of big words. Of course, what we're looking at is somewhat culturally dependent; it just so happens that in our culture there's a heavily male-biased interest in, frinstance, looking at car websites (these seem to be the most male-dominated sites on the list it gives me). What's really interesting, though, is to what extent these patterns persist across different cultures... are men always more interested in transport than women are?

smallblacksun• 16 years ago

Men (in American culture and similar ones) aren't interested in cars because they are transport. They are interested in cars because they make noise and go fast.

jfarmer2• 16 years ago

The probability of me being female is 100%. Not sure what to make of that!

derefr2• 16 years ago

It means, literally, that all the websites you visit are visited more often by women than men. The implication breaks down past there—there are other reasons for visiting female-dominated sites than being female. It could be (if they're mostly content sites/blogs) that you enjoy female writing style/viewpoint more than male writing style/viewpoint. It could be (if they're forums/social sites) that you're single and seeking a female partner. Many other explanations exist. This is a correlation, not a causation.

jfarmer2• 16 years ago

Definitely not true. Just looking at the list of sites they scan, I've visited plenty that have a male/female ratio greater than 1.

KevBurnsJr• 16 years ago

Likelihood of you being FEMALE is 100% - Likelihood of you being MALE is 0%

This is news to me.

derefr2• 16 years ago

Doesn't work in Safari. (Which is a good thing, I suppose.)

geuis• 16 years ago

I suspect what this 'tool' really tells us is that its author needs a new source of data as to what sites men and women visit.

zyb09• 16 years ago

God.. CSS Visited Hack again. WHY does nobody come up with something? I know it's a fundamental issue, but hell.. this is a massive privacy exploit. Please Google? W3C? Someone?

axod• 16 years ago

It doesn't work on Chrome dev builds for a start. I believe other browsers are doing similar.

samdk• 16 years ago

Firefox 4 has a fix for it: http://hacks.mozilla.org/2010/03/privacy-related-changes-com...

orangecat• 16 years ago

In Firefox, you can set layout.css.visited_links_enabled to false in about:config.

endergen• 16 years ago

Did Bart Simpson pick that name. Ring, Ring. "Hello, Is Mike On Ads there?"

My Gonads, Bart Simpson

pingou• 16 years ago

I'm wondering why google.co.uk has a Male-Female Ratio of 1.35.

Using your browser url history to estimate gender

Related Stories