« Shakespeare on Silver Street | HomePage | The Gender Genie says this post was written by a man. »

10/22/2007

The Gender Genie says this post was written by a man.

The Gender Genie purports to predict, with 80% accuracy, the gender of the author of any text inputted into a form.

(Tip of the something-or-other to squib, who linked to this toy earlier this week.)

It supposedly works according to an algorithm developed by a pair of computational linguists. The program searches for the following:

feminine keywords: with, if, not, where, be, when, your, her, we, should, and, me, myself, hers, was.

masculine keywords: around, what, more, are, as, who, below, is, these, the, a, at, it, many, said, above, to.

So, the "Gender Genie" seems to hinge on the idea that women use conditionals and talk about themselves more often than men. There have indeed been studies that show that women -- for cultural, not biological reasons -- tend to make their statements sound less definite (with "if"s, "maybe"s, "likely"s, vocalized pauses, etc) and use personal pronouns more often when speaking. Does this algorithm (perhaps incorrectly) assume that women exhibit the same patterns in their writing?

A columnist with The Guardian was surprised to find that the Gender Genie identified all but one of the paper's female columnists as male, yet it "correctly" identified all of the male columnists.

In the interest of self-reflexivity, the Gender Genie identifies this post as having been written by a man, apparently because (I'm not kidding) I used the word "the" seventeen times.

08:35 Permalink | Comments (6) | Email this | Tags: internet, gender

Comments

And my post [1] is feminine!

[1] http://www.tdaxp.com/archive/2007/10/22/dozier-internet-law-did-it-again.html

Posted by: Dan tdaxp | 10/22/2007

However your post is only 271 words (not that I was pedantic enough to ACTUALLY count every word) and it works best on blocks of text 500 words and over. Give the poor genie a chance!

I would have liked to know how the various keywords or calculations thereof changed between genres (fiction, blog, non fiction) and why...

Posted by: squib | 10/22/2007

Dan,

So does that mean "law firm that totally misunderstands copyright" is a key feminine phrase? ;)

Posted by: PrimroseRoad | 10/22/2007

squib,

It seems to use the same keywords for fiction, so I'm not sure how it differentiates. For some reason, I see this site and can't help thinking of Jack Nicholson's line from that movie with Helen Hunt where he explains how he writes female characters - "I think of a man and I take away reason and accountability."

Posted by: PrimroseRoad | 10/22/2007

lol PR, couldn't agree more!

Posted by: squib | 10/23/2007

Woohoo!

I submitted one of my longest posts [1] after removing the blockquotes, per squib's suggestion, and it's male!

Female Score: 2472
Male Score: 4083

The Gender Genie thinks the author of this passage is: male!

I'm safe!

[1] http://www.tdaxp.com/archive/2005/07/20/dreaming-5th-generation-war.html

Posted by: Dan tdaxp | 10/25/2007

Post a comment