TechMeme's got huge gender biases, and this helps contribute to the echo chamber effect that Fred's talking about. In September 2010 I wrote about a case where even though 65% of the articles on a subject were written by women, TechMeme's version had on 16.5% by women. http://www.talesfromthe.net/jon/?p=1618#comment-88992
Techmeme simply surfaces the hottest content from tech blogs that's getting linked/tweeted right now. It is descriptive, not editorial. If you notice that only 16% of women are being shown covering a story, then your takeaway should be "gosh, only 16% of women wrote interesting posts that are getting linked to on the web" rather than "gosh, techmeme is descriminating against women."
Techmeme just calls it like it sees it. It's awesome in that way.
Techmeme's definition of "interesting" is "selected by an algorithm that starts with hubs that Gabe says are interesting". If 65% of the stories are written by women, and only 16% of Techmeme's stories are, then Gabe's algorithm finds women a lot less interesting than men.
If you agree, then yeah, Techmeme's awesome.
If you don't agree, then you need to be conscious that Techmeme's presenting you a skewed view of the world. Unless you're actively working to balance it with sources skewed in the other direction then it's going to mean that your overall information sources aren't giving you an accurate picture of reality.
If you looked at the top 5 Google results across all key words and men wrote more of the posts than women, would you accuse Google of bias? Or would you consider that maybe its evidence there are more men writing on the web than women. I don't see a difference here. Gabe's said that his system adds new sources programmatically as they get linked to. Why on earth would he block out female voices?
It depends on how much more. Women write roughly as much as men in the blogosphere, so if the overall results were (say) 75% male it would indicate a bias. The times I've looked at Techmeme and memeorandum they're usually 80-90% male.
I don't know why Gabe blocks out female voices but from the Atlantic article linked to in my thread he doesn't seem to read any women regularly so my guess is that he either (a) doesn't know he's doing it or (b) doesn't care.
Copy pasted for convenience. Will delete later if it's distracting.
_____
The coverage of the Arrington kerfuffle provides some great examples of a form of structural bias against women that most people overlook. At this point about 65% of the posts I’ve linked to above are by women (31/47, I think). Here’s TechMeme’s summary of the debate:
five guys and Shira
16.6% women.
In a comment on Gender, race, age and power in online discussions, chapter n back in 2008, I talked about representation issues on Techmeme:
the majority of these blogs are generally what Susan C. Herring et. al. refer to in Women and children last: the discursive construction of Weblogs as “filter blogs”, commenting on external events, as opposed to “personal journals” or “knowledge blogs”; TechMeme’s technology is a natural for filter blogs, so it’s unsurprising that this leads to underrepresentation of women and youth.
It’s nice to have such a clear example!
At the risk of being pedantic,there are a couple of important lessons here:
- The “neutral” algorithms of sites like Techmeme, memeorandum, and mediagazer are in practice heavily biased against women, and so present yet another challenge for women in general — and women in technology, politics, and media in particular
- If you’re getting a large chunk of your news via Techmeme, you’re getting a very male view of reality
Wait, 65% of the stories you linked to were written by women, while only 45%-50% of stories at large were written by women? Have you tried to examine the roots of your gender bias?
If I'm reading joker's summary of your argument correctly, you're saying that techmeme disproportionately covers blogs about tech news instead of blogs people write about their own lives or about generalized, uncontroversial knowledge; which makes it biased against women?
The biasing in this case seems almost an epiphenomenon; TC tries to focus on certain views of certain subjects and succeeds; whether those views are disproportionately written by men has little to do with anything.
You are not reading it correctly. I'm saying that there were 47 stories on this topic, of which 31 were by women. Techmeme featured 6 stories on this topic, of which only 1 was by women. Whether or not a view was written by a man strongly is strongly correlated with whether or not Techmeme features it.
I didn't see you addressing the actual point I raised, which was that Techmeme looks for certain topics, covered from a certain stance. Just like Encyclopaedia Britannica demands articles inhabiting the Classic Stance[1], Techmeme demands its own stance.
But saying this is a gender problem is like saying that because doctors are more likely male than nurses are, and get paid more than nurses, we should have fewer doctors and more nurses.
The data shows that Techmeme's choices of topics and perspectives result in an overwhelming majority of stories by guys even in situations when more women than men are writing about it. Why don't you see that as a gender bias?
I don't know the data on Britannica authors. If there are more guys then women, then there's gender bias, and Classic Stance may or may not be a contributor. But in any case, I don't see what that has to do with Techmeme situation. Ditto with whether or not there's gender bias in the medical profession.
I understood it exactly as khafra did. I suppose we are confused about this line (which khafra was referring to):
TechMeme’s technology is a natural for filter blogs, so it’s unsurprising that this leads to underrepresentation of women and youth.
This quote makes it sound like women and children write more personal and knowledge stuff, which TechMeme naturally overlooks. Further, it straight up says that this tendency is "unsurprising".
So we are wondering what this has to do with gender bias or why it's surprising that a technology would miss things that it is designed to miss.
But you said he was not reading the summary correctly, so obviously something is not clear here.
