What has made guys say "right on"? March 3, 2011 6:34 PM   Subscribe

Has anyone ever looked at which posts/comments have been most favorited by men, and which by women?

Just wondering if anyone with programming skills has ever been interested to look at that, or if it's ever been discussed here. It would obviously be impossible to know for certain. But it would at least be possible to know which posts/comments have been most favorited by people who've indicated male/dude/bro/guy or lady/woman/chick/etc in their profiles. I think it would be totally fascinating to see a top 20 list of each (there would probably be a lot of overlap there) and then a top 20 list of "highest % of male favorites when total amount of favorites is over x" or something like that.
posted by Ashley801 to MetaFilter-Related at 6:34 PM (142 comments total) 3 users marked this as a favorite

We don't have any reliable split of MeFites by gender and probably this isn't a road we're likely to go down.
posted by jessamyn (staff) at 6:35 PM on March 3, 2011 [8 favorites]


Yeah, I definitely think it would be something an interested mefite would just look at in their spare time, not something that would ever be displayed in an official capacity.
posted by Ashley801 at 6:40 PM on March 3, 2011


There are men here?
posted by Meatbomb at 6:40 PM on March 3, 2011 [1 favorite]


Yeah, for starters, you'd have to define "men" and "women" or even "male" and "female" and before you even start, someone's going to be oppressed.

Not me personally, I'm too busy being the great oppressor and perpetuating the gender binary by stocking my would-you-please-just-get-born-already baby's drawers with traditionally masculine blue onesies to match his testicles. But someone.
posted by sonika at 6:50 PM on March 3, 2011 [15 favorites]


Well, obviously men have favorited more posts on the blue. The question is complicated, however, by the fact that there isn't a pink. If there were, I'm sure more women would have favorited posts there.
posted by koeselitz at 6:53 PM on March 3, 2011 [7 favorites]


Then, we run into other questions, for example "Is a non-op transexual female who only does cisgendered females a lesbian?" That is when things really get messy.
posted by Ardiril at 6:57 PM on March 3, 2011


My posts tend to be very popular . . . with the ladies.
posted by ND¢ at 6:58 PM on March 3, 2011 [19 favorites]


That's what she said!
posted by blue_beetle at 6:59 PM on March 3, 2011 [1 favorite]


You can omit the "messy" options by just limiting yourself to people who've simply answered M or F. Like they do for most studies of this kind.
posted by DU at 7:01 PM on March 3, 2011


(Wishing I hadn't made a stupid ironical-sexist joke, hoping there aren't too many more)
posted by koeselitz at 7:01 PM on March 3, 2011


That is when things really get messy.

Bring a towel.
posted by Brandon Blatcher at 7:03 PM on March 3, 2011


(Wishing I hadn't made a stupid ironical-sexist joke, hoping there aren't too many more)

The only way to win is not to play.
posted by jessamyn (staff) at 7:03 PM on March 3, 2011 [37 favorites]


Yeah, for starters, you'd have to define "men" and "women" or even "male" and "female"

Perhaps that can be done by examining what comments people favorite.
posted by StickyCarpet at 7:03 PM on March 3, 2011 [4 favorites]


Yeah, for starters, you'd have to define "men" and "women" or even "male" and "female" and before you even start, someone's going to be oppressed.

Not sure if you are being facetious or not so this might be a weirdly serious reply. But I think because everyone gets to self-define in their profiles as male or as female, or as something else, or leave it blank, nobody would be oppressed if the set of males and females were defined based on who indicated in their profiles that they are part of those sets.
posted by Ashley801 at 7:08 PM on March 3, 2011 [1 favorite]


I just assume gender from people's usernames.
posted by Blazecock Pileon at 7:11 PM on March 3, 2011


And myself, because I just don't feel like putting anything there. I still think it would be extremely interesting though, even if it all it showed was the split between people who identified enough with the gender binary/felt like putting something in their profile, vs the split between all male users and all female users.
posted by Ashley801 at 7:14 PM on March 3, 2011


everyone gets to self-define in their profiles

We lied.
posted by Ardiril at 7:16 PM on March 3, 2011 [1 favorite]


Penis!
posted by special-k at 7:19 PM on March 3, 2011


My penis is so big, it won't take Spielberg's calls.
posted by Cool Papa Bell at 7:25 PM on March 3, 2011 [2 favorites]


jessamyn: “The only way to win is not to play.”

Yeah, it's odd how one can think that one is making an innocuous little joke, and yet looking back at it, it really is a useless and even destructive thing, no matter how "neutral." It becomes obvious: it's pretty much always a bad idea to make ironically sexist jokes on the internet.

hal_c_on: “wow. this thread degenerated into garbage faster than gametes after chernobyl.”

Well, to be fair, it's only two bad comments out of the lot so far. One can hope that others are more intelligent than I was.
posted by koeselitz at 7:27 PM on March 3, 2011 [1 favorite]


I just assume gender from people's usernames

No way that works.
posted by cashman at 7:28 PM on March 3, 2011


After that 4th Indiana Jones movie, no part of me would take Speilberg's calls, unless he wanted someone to tell him when the story had reached new levels of crappiness.
posted by Brandon Blatcher at 7:30 PM on March 3, 2011


Not sure if you are being facetious or not

I thought it was clear from the second half of my comment that yeah, there was a lot of tongue in my cheek.
posted by sonika at 7:30 PM on March 3, 2011


There's a lot of tongue in my cheek right now, too, and none of it's mine.

It's getting harder to type.
posted by Pater Aletheias at 7:33 PM on March 3, 2011 [11 favorites]


Why did I know this thread was instantly going to be about gender oppression
posted by tehloki at 7:41 PM on March 3, 2011 [1 favorite]


Because your big manly brain is hard and taunt with knowledge and wisdom.

Or lucky guess.
posted by Brandon Blatcher at 7:57 PM on March 3, 2011


The real question for me is whether this comment was favorited more by men or women.
posted by danb at 8:08 PM on March 3, 2011 [3 favorites]


If we have sent a man to the moon and have created open gender boxes on profile pages, why can't we design a super-smart python script to read everyone's profile, interpret their gender with its humor and sarcasm detection engine, and then sort out the posts, all for the purpose of saying "Girls post like this, while bros post like this."
posted by mccarty.tim at 8:08 PM on March 3, 2011 [1 favorite]


That is EXACTLY the comment I was thinking about when I made this post, danb, and that is exactly what I was wondering about it! You know, every now and then, someone says something here that makes a whole lot of people go, "Shit, that person really nailed something." I think I tend to notice those posts that nail things for a lot of women, but I don't know what they are for guys. I'm totally totally fascinated about what they might be.
posted by Ashley801 at 8:17 PM on March 3, 2011


Don't know, and I don't wanna know. I can imagine anything like it would be immediately used in already-contentious metatalk threads to bludgeon fellow mefites with regards to their gender biases. It would primarily function as a nasty enabler of bad generalisations about genders and mefites both; the equivalent of trawling through someone's user history writ large.
posted by smoke at 8:21 PM on March 3, 2011 [6 favorites]


I think I tend to notice those posts that nail things for a lot of women, but I don't know what they are for guys. I'm totally totally fascinated about what they might be.

This is actually exactly why I have gone out of my way never to specify any of my demographic information. I realize I am probably weird like this but I'd rather let people assume whatever gender/background/shoe size that takes their fancy. I figure this way a reader is as likely to say "Wow, she nailed that!" as "I agree with this fine gentleman." Note that options for disagreement are not mentioned, because hooray for me.

Or, whatever.
posted by FAMOUS MONSTER at 8:28 PM on March 3, 2011 [2 favorites]


Oh yeah, and the biases. I am big on trying to make sure my actual statements don't take a backseat to the person saying it, if only because sometimes I am in contentious threads and shit's complicated enough as it is.
posted by FAMOUS MONSTER at 8:30 PM on March 3, 2011


Oh, it's about THAT comment again.
posted by Ardiril at 8:33 PM on March 3, 2011


jessamyn: "The only way to win is not to play."

Wargames pedantry: It's "The only winning move is not to play."
posted by Chrysostom at 8:33 PM on March 3, 2011 [1 favorite]


and then a top 20 list of "highest % of male favorites when total amount of favorites is over x" or something like that.

I'd also be interested in seeing this information. I imagine I'd be more surprised by the male-dominant comments than the female-dominant ones. In other words, I have an easier time thinking of what kinds of comments would be mostly favorited by women than which would be mostly favorited by men.

I once made a comment (under a different username) where almost all 25 people who favorited it were women. I was proud of it.
posted by John Cohen at 8:43 PM on March 3, 2011


I think this would be interesting too, although it might be interesting in an "oh god, I wish we hadn't known this" kind of way.

I think the free-form nature of the gender field is actually a fairly tractable problem. It wouldn't be all that hard to pull out all the unique values from that field and then map them by hand to M, F, or null (to indicate missing information).

But! The gender field isn't contained in the Infodump, and the omission is intentional. Which makes me think that a process of obtaining it in analyzable form any another way, like scraping all the profile web pages, is at least frown-worthy.
posted by FishBike at 8:47 PM on March 3, 2011



I just assume gender from people's usernames.


That works perfectly well with mine.
posted by louche mustachio at 9:05 PM on March 3, 2011 [1 favorite]


I would expect the results to be both morbidly interesting and consistently depressing, and so am glad there's no such script already installed.
posted by darth_tedious at 9:10 PM on March 3, 2011 [1 favorite]


The real question for me is whether this comment was favorited more by men or women.

That comment was favorited more by people who really enjoy "Everybody Loves Raymond."
posted by drjimmy11 at 9:13 PM on March 3, 2011 [3 favorites]


I think the free-form nature of the gender field is actually a fairly tractable problem.

Absolutely. You'd be stuck with a subset of the site population and there'd be the possibility of some error effect from folks intentionally misidentifying for one reason or another, but it'd be pretty doable and wouldn't be particularly labor intensive if you skip the hand-coding of less common labels and just went with the common unambiguous ones.

But! The gender field isn't contained in the Infodump, and the omission is intentional.

Correct, there's no profile page information in the Infodump and that's likely to continue being the case over the long haul.

That said, I did put together a dump of anonymized gender field labels a while back (from here), which give at least an idea of what the distribution of available unambiguous gender claims are in profiles. If someone would like to take a serious shot at doing some anonymized analysis of the sort of gender-id-driven activity discrepancies folks have talked about in here, I'd be willing to provide them with better data to start from.
posted by cortex (staff) at 10:18 PM on March 3, 2011


I wonder if more males or females answered the gender question. I didn't, but I had to look to see and I have no idea why I left it blank. Maybe I was intimidated by the whole free form of it. Maybe I panicked and couldn't just write "M".
posted by JohnnyGunn at 10:29 PM on March 3, 2011


Was gender always there in the sign up form? It seems like the vast majority avoided answering.
posted by BrotherCaine at 10:59 PM on March 3, 2011


variations on the "yes, please!" joke: ~39.

Remarkably restrained, MeFi. Bravo!
posted by zamboni at 11:05 PM on March 3, 2011


I've done a bit of study about gender online. So far I've found that my own research falls in line with Susan Herring's work in that language use and site behavior is more strongly correlated with the 'gendered genre' of the speech environment rather than the self-ascribed gender of its participants. By gendered genre, I mean that previous work (Herring and others) has shown that blogs tend to fall into 1 of at least 2 categories - the 'diary' blog or the 'filter' blog, with corresponding writing styles. Those categories and their styles have gender ascriptions and people tend to conform to those ascriptions in their participation behaviors, regardles of their own actual gender (or sex).

All that said, it would be pretty fascinating to see if there were a correlation between more societally recognised masculine/feminine self-descriptions of gender ("I'm a dude" vs. "male") and more socially recognised gendered site behaviors ("male-answer syndrome", "I think" vs. "I feel" statements) and *then* see if other behaviors correlate as well, and if there's further support for those correlations suggesting that there may be gendered social ascription underway for certain behaviors.
posted by iamkimiam at 11:18 PM on March 3, 2011 [10 favorites]


Age and geography, on the other hand...
posted by iamkimiam at 11:19 PM on March 3, 2011


I think this would be interesting too, although it might be interesting in an "oh god, I wish we hadn't known this" kind of way.

Perhaps, but it seems that most of the very heavily favorited posts/comments on this site are pretty awesome in one way or another. So I think both the male-favorite-dominated and female-favorite-dominated things would be mostly great/thought-provoking/funny, regardless. Now, 4Chan, I'd be scared to see their results.
posted by Ashley801 at 11:33 PM on March 3, 2011


This whole thread could be surmised in Rules of the Internet, sections 16 and 49, with additional resources in sections 11, 14, 18, 20, 25, 34, 35, and 39.
posted by Mister Fabulous at 11:40 PM on March 3, 2011


...baby's drawers with traditionally masculine blue onesies to match his testicles.

IANAMD but if his testicles are blue I'm pretty sure you need to take him to the doctor.
posted by vapidave at 12:28 AM on March 4, 2011 [3 favorites]


I tend to think of myself as an emasculated Action Man.
posted by arcticseal at 12:35 AM on March 4, 2011


That makes sense, iamkimiam, since gender is a social construct. I dig your research!
posted by aniola at 12:59 AM on March 4, 2011


The only way to win is not to play.

Or to be Charlie Sheen.
posted by Infinite Jest at 1:28 AM on March 4, 2011 [1 favorite]


traditionally masculine blue

Traditionally, pink was the boys colour since it was like MANLY RED, THE COLOUR OF BLOOD but toned down a bit for the kids. Blue was the girly colour because it was delicate and dainty. Around the 1940s it switched round for some reason.
posted by EndsOfInvention at 2:17 AM on March 4, 2011 [2 favorites]


I've got a MeFiMail for arcticseal. I'm happy. Hope you're happy, too.
posted by Eideteker at 4:06 AM on March 4, 2011 [1 favorite]


I think I tend to notice those posts that nail things for a lot of women, but I don't know what they are for guys. I'm totally totally fascinated about what they might be.

A quick scan through comments I've recently favorited found these little nuggets where this particular guy was like "YEAH RIGHT ON" and the comments got a lot of favorites:

* * * * * * *
posted by Brandon Blatcher at 4:09 AM on March 4, 2011


I can't see what interest this would be, other than reinforcing the existing gender binaries in our society. Women favorite like this, men favorite like that. The differences would be largely be self-reinforcing; much in the same way that girls don't necessarily prefer dolls, but have dolls thrust upon them. All it would reveal are the things we train our young girls and boys to prefer, the same things we've known all along.

So really, this is all just for your curiosity and not for any greater purpose. Tsk tsk, for shame.
posted by Eideteker at 4:14 AM on March 4, 2011


Your kind would tsk tsk like that.
posted by Brandon Blatcher at 4:57 AM on March 4, 2011 [1 favorite]


I can't see what interest this would be, other than reinforcing the existing gender binaries in our society.

Just for example, it might show that there's actually no statistically significant difference in favoriting based on gender. Then we could say: "Take THAT, stereotype reinforcement committee!"

That said, I did put together a dump of anonymized gender field labels a while back [...]

So it looks like there are about 9,300 users who filled this in. About 2,000 of those used something totally unique, and that results in a long list that would be a pain (but not totally impractical) to go through and code. But the list of entries used 2 or more times is pretty short and would cover 7,300 of the 9,300 users who filled in the gender field, roughly. That seems quite practical indeed.

If someone would like to take a serious shot at doing some anonymized analysis of the sort of gender-id-driven activity discrepancies folks have talked about in here, I'd be willing to provide them with better data to start from.

What did you have in mind? I think we're talking about analyzing favoriting activity by gender, so I'm having trouble figuring out how to anonymize the gender information and yet still use it for the analysis.

Or is this maybe an arms length sort of thing where somebody (e.g. me) would provide the mapping of gender field entries to M/F/? and you could provide an aggregated list of favorites for each gender without any user ids in the data?
posted by FishBike at 5:49 AM on March 4, 2011


If someone does do this, I don't anticipate anyone interpreting the results with any particular agenda and causing controversy.
posted by empyrean at 5:49 AM on March 4, 2011


I'm with the OP -- data doesn't have to be perfect to be useful and interesting, it can just be a guide. I'm having a hard time finding good stats on male/female usage patterns on the internet and personally I'm interested in any data I can get. I'm not designing bombs with it or anything.

Such an analysis would strip out for example - me - b/c I don't indicate gender in my profile (I'm female and I'm lazy about filling out profiles and often don't feel inspired) but I would still be interested in the data about the people who do indicate clearly either way.

And including those who either did not indicate gender or who indicated some form of 'other' -- well, their response patterns (and this bucket would include mine) is interesting also.
posted by A Terrible Llama at 6:18 AM on March 4, 2011


FAMOUS MONSTER: I figure this way a reader is as likely to say "Wow, she nailed that!" as "I agree with this fine gentleman."

Whatever, I still just give you favourites because I like your name.
posted by ArmyOfKittens at 6:22 AM on March 4, 2011 [2 favorites]


Blazecock Pileon: "I just assume gender from people's usernames"

That makes you a flaming rooster?
posted by Splunge at 6:45 AM on March 4, 2011 [1 favorite]


What did you have in mind? I think we're talking about analyzing favoriting activity by gender, so I'm having trouble figuring out how to anonymize the gender information and yet still use it for the analysis.

I meant more like "be responsible about anonymizing results". As in, I'd be comfortable handing a trustworthy datawanker a list of userid + genderfield pairs so that they could do some analysis on correlations or lack thereof, on the understanding that the provided list was not for public consumption and the results would just presented as aggregate data without any userid information included.

One of the big challenges with this line of investigation is all the conflating factors that could come into it: if there is some gender valence, how significant is it compared with other sorts of social valence like e.g. contact-correlated affinities or other less explicit social connections? If Alice knows Barbara via meetups or shared subject interest and favorites Barbara's comments proportionally often, is it because they're both women or because they know each other or is it because Barbara writes well or in a way or on subjects that Alice happens to appreciate?

So it'd be interesting to see if there are gender-correlated systemic effects in play at all, but that'd need to be looked at in context to be sure it's not actually a surface symptom of some other gender-unrelated social effect in play, etc.

Anyway, I think it'd be interesting. I'm skeptical of any broad, significant gender-based effects though I wouldn't be surprised at all to see significant effect in very subject-specific contexts and I'm a bit curious if there would be some revealed systemic difference between activity on the various subsites.
posted by cortex (staff) at 6:55 AM on March 4, 2011


Q Are we not men?
A We are Mefites.
posted by The Whelk at 7:36 AM on March 4, 2011 [4 favorites]


Has anyone made the obligatory "Men be favortin' like this, women be favoritin' like that" joke yet?
posted by electroboy at 7:41 AM on March 4, 2011


due to evolutionary biology as hunters men are "programed" to favorite as many times as possible while women "gather" only a few select favorites.
posted by The Whelk at 7:47 AM on March 4, 2011 [6 favorites]


Traditionally, pink was the boys colour since it was like MANLY RED, THE COLOUR OF BLOOD but toned down a bit for the kids.

Yeah, I know. It's weird how that shift happened. Also: blue is associated with the Virgin Mary, which put it right in the "girl" camp for eons. Not that that makes any sense either. It's a COLOR!

But seriously: try buying baby clothes these days. You've got your options of blue, pink, or.... lime green. Why did someone decide that LIME GREEN was an acceptable gender neutral color to put on a poor defenseless baby who is going to cry anyway? Hell, *I'D* cry if you dressed me in a lime green outfit. Why not purple - the sum of both red AND blue?
posted by sonika at 7:50 AM on March 4, 2011


(aside, yeah Red was supposed to be protective of evil spirits and pink was a dilute red. Blue got pinked cause it was complimentary and yeah ..Virgin Mary connotations. Like all bad things, I blame the Victorians. Its also why men can't go around with shoulder length hair and bright yellow checkered vests and show off fine calves in white stockings anymore. So. Not. Fair.)
posted by The Whelk at 7:56 AM on March 4, 2011 [4 favorites]


Green - not blue - is actually the complimentary color to red, but perhaps the pre-Victorians recognized it looks like shit on babies. The Victorians, taking the FUN out of, well, everything since 1890.
posted by sonika at 8:10 AM on March 4, 2011


sonika: purple is Girl. Pink is Girl. Ladybugs and dragonflies? Girl. I have a 4 year old daughter, for whom I purchased a TON of clothes because everything in the stores was so stinkin cute I could barely stand not to leave with it. Then I gave birth to a boy 4 months ago, and. . . wow. There's less available, and it's boring! Your choices are light blue, dark blue, brown, black, grey, and a peculiar orange. Unless you mortgage your home and buy Zutano stuff. If you hate sports and camo, as I do, you're cutting out about half of your options. I'll take the lime green just for variety!
posted by KathrynT at 8:16 AM on March 4, 2011


I'm one of those outliers who boldly goes forth and wrecks linguistic gender profiling attempts - computers always, always think I'm a man, even when I'm wearing a fluffy pink sweater and checkered skirt and listening to Romantic period classical music while eating "health" food. I'm thus pretty sure I don't favorite like a girl.

But: I love data as much as I love my guilt-inducing Fig Newtons, and would be delighted to see the results of such a study, as inapplicable to my realty as I'm sure they'd be.
posted by SMPA at 8:38 AM on March 4, 2011


There's less available, and it's boring! Your choices are light blue, dark blue, brown, black, grey, and a peculiar orange.

I know! And decide you don't want clothes with footballs on them and you've cut out half your options! We've done pretty well amassing a sartorially respectable newborn wardrobe for the babby, but GEEZ. Thinking ahead to Easter, there are 4,000 little wee Easter dresses for girls and for boys? I guess I get to put him in a sweater vest. Great.

Oh well, I'm also going to stick bunny ears on his head. Because why have a baby and NOT stick bunny ears on its head?
posted by sonika at 8:52 AM on March 4, 2011 [1 favorite]

My posts tend to be very popular . . . with the ladies.
Not just the ladies. IF YOU KNOW WHAT I MEAN.

No, seriously, what do I mean
posted by scrump at 9:04 AM on March 4, 2011 [2 favorites]


I think you mean that he is bitextual.
posted by cortex (staff) at 9:06 AM on March 4, 2011 [7 favorites]


sonika: " But seriously: try buying baby clothes these days. You've got your options of blue, pink, or.... lime green. Why did someone decide that LIME GREEN was an acceptable gender neutral color to put on a poor defenseless baby who is going to cry anyway? Hell, *I'D* cry if you dressed me in a lime green outfit. Why not purple - the sum of both red AND blue?"

Suggest you shop at stores other than Childrens Place. Look at Old Navy, Macy's (sometimes) or Target. They tend to have a wider variety of color choices in baby and toddler sizes. Not *much*. But more.

Honestly, it doesn't matter in the long run. Despite herculean efforts by us to make sure our munchkins were not indoctrinated into gender/color themes, my daughter's favorite color to wear is now either pink or anything that has flowers or hearts on it. (She likes pajamas with monkeys on them.) My son's favorite colors are either dark green, red or any shade of blue, depending on his mood.

At some point, they start making choices and you just roll with it. :)
posted by zarq at 9:08 AM on March 4, 2011


Favorites Pony is neither man nor woman, but Favorites Pony awards lots of favorites.
posted by Favorites Pony at 9:28 AM on March 4, 2011 [1 favorite]


Suggest you shop at stores other than Childrens Place. Look at Old Navy, Macy's (sometimes) or Target.

I have been looking exclusively at Macy's and Target ;) I'm just talking the newborn clothes, perhaps they branch out a bit later. I shall eventually find out.

And yes, some day he will choose his own wardrobe. Until then, I just want the little dude to look awesome. This is one of the perils of having gone to art school and being very, very fussy about things LOOKING AWESOME. So much baby swag is completely non-awesome and a lot of that breaks down around gender lines. Things for girls are supposed to be "cute" and often ARE. Things for boys? Eh, just stick a football on it.

Thinking about it: this is kind of true for women's and men's fashion in general. Things for women are generally prettier and there's a greater selection. Dudes? Well, you can pick whatever team jersey you want!

Maybe my babby will be a designer who will fight for fashion equality. I can only hope.
posted by sonika at 9:30 AM on March 4, 2011


sonika: "I have been looking exclusively at Macy's and Target ;) I'm just talking the newborn clothes, perhaps they branch out a bit later. I shall eventually find out.

Heh. It figures. :)

It does get easier as they get older, especially if you're not planning on putting him in only t-shirts. My son has a bunch of polo shirts and a few button-downs. He has funky stuff like an awesome Hawaiian shirt or two.

For everyday clothes, now that it's winter he mostly wears long-sleeve tees/polos, sweaters and either jeans or some sort of dockers-style pants, in different colors/shades.

Thinking about it: this is kind of true for women's and men's fashion in general. Things for women are generally prettier and there's a greater selection. Dudes? Well, you can pick whatever team jersey you want!

Even growing up I never wore team jerseys or sports-wear unless I was playing sports. There's a huge range of clothing options available for boys and men. You'll see when you have to start shopping for him when he's older. :)

Maybe my babby will be a designer who will fight for fashion equality. I can only hope."

:D Please teach him early that with horizontal stripes, nobody wins. :D
posted by zarq at 9:57 AM on March 4, 2011


Sonika, why not put him in "girl's" clothes if those are more interesting? It's just cloth.
posted by two or three cars parked under the stars at 11:02 AM on March 4, 2011 [2 favorites]


As in, I'd be comfortable handing a trustworthy datawanker a list of userid + genderfield pairs so that they could do some analysis on correlations or lack thereof, on the understanding that the provided list was not for public consumption and the results would just presented as aggregate data without any userid information included.

I'm so on the fence about this right now. I'm wondering how the general MeFi community feels about doing this at all, and then specifically whether or not the mods and other site users think I'm a trustworthy datawanker? Because I would love to give it a try if people are more or less fine with it, but I worry that people are NOT more or less fine with it.
posted by FishBike at 11:38 AM on March 4, 2011 [1 favorite]


I think you're a trustworthy datawanker, but I think any "conclusions" that could be drawn from this are going to get the same old "AHA" league jumping on the same old bandwagons and I find that a little tiresome. That said I don't think it's going to, say, mean a lot more work and other people might find it useful, so hey go ahead.
posted by jessamyn (staff) at 11:40 AM on March 4, 2011


Dude, in the dictionary next to Datawanker comma Trustworthy there is a picture of a fish and also of a bike. If you're interested in giving it a go, let me know.

We can talk about the results you come up with and talk about it on email if you want to sanity-check anything; if someone has strong This Cannot Be Done feelings about this they can speak up in the interim, but there's a big gap between taking a crack at some analysis and doing bad or stupid or overreaching things with that analysis so I don't really feel like we're talking about problematic territory.
posted by cortex (staff) at 11:42 AM on March 4, 2011 [1 favorite]


...and yet looking back at it, it really is a useless and even destructive thing, no matter how "neutral."

What was destroyed, again?
posted by coolguymichael at 12:20 PM on March 4, 2011


Metafilter: I favorite things because of my genitals
posted by found missing at 12:29 PM on March 4, 2011 [1 favorite]


Sonika, why not put him in "girl's" clothes if those are more interesting? It's just cloth.

Eh, I've thought about it, but until he's old enough to decide that he likes girl's clothes, I'm just sticking with the "traditional" route. Though I will absolutely jump at the chance should he decide he needs a princess dress or a tutu - I'd like that to be his decision later on and not something that he discovers his baby pictures of him in pink onesies and gets embarrassed about it.

(Also: half of his family is Portuguese and while not terribly *conservative* they are more traditional than I am and would be at least totally baffled if I dressed him all up in pink, so in terms of picking battles to fight, that one's not worth it.)
posted by sonika at 12:31 PM on March 4, 2011


I really dislike the implication that I favorite things because of my genitals.

The things that my genitals favorite is a very, very small portion of what I myself favorite. My genitals have extremely high standards.
posted by sonika at 12:34 PM on March 4, 2011 [4 favorites]


My genitals have their own login name at Metafilter, and I've got that name killfiled because I'm so embarrassed by what they post.
posted by found missing at 12:38 PM on March 4, 2011 [2 favorites]


I hope they have the decency to lysol the keyboard after they take their turn on the 'filter.
posted by bonehead at 12:53 PM on March 4, 2011 [1 favorite]


My genitals have their own login name at Metafilter, and I've got that name killfiled because I'm so embarrassed by what they post.

We've gone over this before. The term is sock puppet, OK?
posted by zamboni at 12:58 PM on March 4, 2011 [7 favorites]


Well, I favourite things, so what does that make me?
posted by ob at 1:00 PM on March 4, 2011


Oh, why oh why did I just ask that?
posted by ob at 1:01 PM on March 4, 2011


Sounds like some of you may have a congenital favorites condition.
posted by It's Raining Florence Henderson at 1:03 PM on March 4, 2011


I'd be happy to lend whatever support, info, research etc. that I can towards interpreting the results and trying to disambiguate the conflating factors cortex mentioned above. I am absolutely butt-crazy passionate about this sort of research.

The datacrunching can be done while preserving anonymity and respecting people's boundaries. Is there any reason why userids even have to be involved at any point in the process beyond the initial read of the datadump file?
posted by iamkimiam at 1:09 PM on March 4, 2011


The datacrunching can be done while preserving anonymity and respecting people's boundaries. Is there any reason why userids even have to be involved at any point in the process beyond the initial read of the datadump file?

Well, if we've got one table that links a user ID to a specific comment or favorite, and another table that links a user ID with a specific gender, then we do need to make use of the user ID to link comments or favorites with gender, just for example.

Once summarized, then the user ID can drop out of the picture depending on what is being summarized. And I think part of the trick would be not to present any results that still have to have a user ID (like any kind of top-N users based on this, whereas top-N comments or posts would be fine in this regard).
posted by FishBike at 1:18 PM on March 4, 2011


Oh, right, that makes sense. But the user ids don't mean much to you while you're crunching, and if they're dropped out on the summarize, then I just don't see much to worry about. But if there are users with issues, then hopefully they will come forth with their concerns.
posted by iamkimiam at 1:34 PM on March 4, 2011


until he's old enough to decide that he likes girl's clothes

When my daughter's daycare had a "wear your favorite sports team" day, I kind of blew some of the teachers' minds by putting her in this shirt. For the most part I'm totally with you, but I figure as long as we're buying the clothes and making the dressing decisions, I'll decide which teams she supports.

Sadly, she's now beginning to make some of her own decisions on dressing and seems not to like the black and browns of St. Pauli. She'll come around, though.

More seriously - my sister-in-law studiously avoided pink with her daughter, and the pro-pink rebellion started early and strong, and brought with it Barbies and princesses. We've not taken strong measures to avoid pink, but also haven't pushed it. It's much too early to say, but at this point our daughter seems to like it OK but I wouldn't say it's a candidate for a favorite color.
posted by nickmark at 1:44 PM on March 4, 2011 [1 favorite]


Just hear to give a mighty shout for PINK! Pink for boys! Pink for girls! Pink for profiles!
posted by ThePinkSuperhero at 1:48 PM on March 4, 2011 [4 favorites]


I figure as long as we're buying the clothes and making the dressing decisions, I'll decide which teams she supports.

Word. Soon-to-be-babby already owns his first Red Sox jersey.

(Pedroia, out of respect to his Portuguese heritage.)
posted by sonika at 1:49 PM on March 4, 2011


Lime green is the perfect color for a a vexing imp.
posted by vapidave at 2:13 PM on March 4, 2011 [1 favorite]


The only sports-team-logo clothing either of my kids have is this.
posted by KathrynT at 2:18 PM on March 4, 2011


There's some irony in the fact the original poster has a gender-neutral named and doesn't choose to select male or female or anything on their profile. But they are wanting to judge men and women on what they favourite!
posted by crossoverman at 2:35 PM on March 4, 2011


a gender-neutral named

Ashley? That reads very obviously female to me, seeing as it was the most popular girl's name in the US not too long ago. I know it was originally a boy's name but it's so very rarely used that way anymore.
posted by sonika at 2:39 PM on March 4, 2011


EndsOfInvention: Traditionally, pink was the boys colour since it was like MANLY RED, THE COLOUR OF BLOOD but toned down a bit for the kids. Blue was the girly colour because it was delicate and dainty. Around the 1940s it switched round for some reason.

I'm taking it way back (circa 1800s), and dressing Jr. filthy light thief in white dresses and skirts. We were talking about going with an all-white Baby Toga, but I think that can be reserved for around-the-house.
posted by filthy light thief at 3:00 PM on March 4, 2011


I don't understand why this kind of information would be a bad idea. Or why there would be any judging, negativity, or oppression that would happen on metafilter if this information was released. Can someone explain what harm this could do?
posted by royalsong at 3:18 PM on March 4, 2011


Because it forces people to pick an assigned role just to answer a question. And without accurate information the results are skewed anyway. There are a few "mails" on this site. Are you going to force me to go in and select a sex (or gender)? For someone like me that's probably not that big of a deal, but some people struggle with this aspect of identity, and a lot of the labels don't apply to many people.

It's also a bit creepy. That's just one dude's opinion.
posted by cjorgensen at 3:59 PM on March 4, 2011


Can someone explain what harm this could do?

I too find that surprising, but I think it's because some aggregated data could be used to reinforce tedious perceptions of behaviors associated with genders and that it will also slight those who feel confined by social (and UI) requirements of picking a gender. It gets tiresome.

Not everyone has a team, and not everyone is on the team of the uniform they happen to be wearing -- and I'm tired. Somebody do something about that metaphor, please and please cut me some slack about it because it's all I've got handy and my metaphorical language has been sucking lately.

Anyway - I've used shoe shopping as a reference point lately professionally (context of targeted ads) a lot more than I would if I were a guy, and if the male audience I'm talking to wants to blow me off on the topic because I'm referencing shoe shopping they can kiss my ass. I'm not going to sit there and reference flash cookies that direct me to buy hockey sticks or handguns (JOKING).

I think the underlying worry is that someone could use such data to say it "proves" something, and in particular, something that could be used to dismiss males, females, and those unaligned in a manner that feels true to them.

Data out of context is used to marginalize and misrepresent people all the time--most of the time when you see some heady percentage presented somewhere you don't really know the quality of the research behind it, the other factors, whatever.

So, I don't know -- maybe you take people who receive targeted ads about shoe shopping and draw some sort of correlation between that and being really stupid.

Apologies if I just misrepresented the entire thing, but that's my take.
posted by A Terrible Llama at 4:06 PM on March 4, 2011


Because it forces people to pick an assigned role just to answer a question.

I'm thinking of the passive "research on data as it exists now, for the sake of research" thing here so maybe you're answering a question about something else and I'm just barking up the wrong tree, but: who is being forced to pick anything in order to do anything? If someone previously put an unambiguous gender label in their gender field, that's useful from a research perspective for looking for correlations of past site activity, but that's it. There's no compulsion to do anything with the gender field, no compulsion to take corresponding action after having done so, etc.

I guess my feeling on this stuff is mostly that bad use of data is the problem, not data itself, and research that is clear about its own assumptions and limitations and methodology is mostly an unambiguous good if created and presented in good faith. It's one thing to say "according to this chart, WOMEN ARE FROM VENUS, GET IN THE KITCHEN"; it's another entirely to say "so I looked at data x, y, and z, compared them by this method, and found a slight correlation between a female-identifying gender field and favoriting rate for comments on askme vs. metafilter".

Fundamentally, someone who has pushy or jerkish stances about the world or how it should be partitioned doesn't really need data to be like that, and data doesn't need to be hidden from view just because someone with a crappy attitude might be a butthead about it. A butthead is a butthead, and nothing is stopping them from scraping and collating what is, fundamentally, semi-public data on their own if that's the kind of butthead they're inclined to be.

In the mean time, mefites mostly seem like pretty decent and thoughtful people and I pretty much trust the sorts of folks with a passion for looking at this community analytically for the joy of analysis to be good about doing that in a responsible way that minimizes harm.
posted by cortex (staff) at 4:29 PM on March 4, 2011 [8 favorites]


Because it forces people to pick an assigned role just to answer a question.

Particularly, the special snow flakes.
posted by Ardiril at 4:38 PM on March 4, 2011


Has no one brought up the issue of how favorites are used? Possibilities that I can think of: a mark that "I agree with this comment", a mark that "I find this funny", a mark that "I find this insightful", a reminder to read later for casual enjoyment and unmarked when read, a reminder to mine for information, a reminder to check comments later.

With that, I suggest an antithesis to November '09: add a bucket of "sorting options" for favorites, and include "favorite" as an option to opt-out of the game.

But wait, that's not enough! Each favorite option needs a modifier. Options include: knee-jerk favorite, OMG Awesome!!! favorte, , took me a minute to realize it's brilliance favorite, sarcastic favorite, favorite because I like this poster, and null.

Ooh, one better: VOICE FAVORITING. Because marking text just doesn't carry enough weight to really understand why people make the choices they do. And this is all about understanding people, right? WE NEED MORE DATA.
posted by filthy light thief at 5:06 PM on March 4, 2011


I traded a small electric treadmill I wasn't using for 2nd hand girl's babyclothes. The lady came by with a car-load of baby and toddler clothes, most of it unworn and still with the labels on, and brand-name, yuppie-label stuff we wouldn't normally bother with. Firstborn girl in a =very= large extended family, apparently - they got more clothes as presents than they could use, and now it was just in storage. She kept trying to pay me $50, for the treadmill of for taking the clothes off her hands, I dunno. I told her to keep the money all the same.
posted by Slap*Happy at 5:10 PM on March 4, 2011


I'm just saying, at least INVITE me to the christening or, oh man, I hope you don't like spinning wheels.
posted by The Whelk at 5:25 PM on March 4, 2011 [1 favorite]


Not everyone has a team, and not everyone is on the team of the uniform they happen to be wearing

Actually, in the context of the data we're dealing with, I think that addresses the issue pretty nicely. To me, the questions that this investigation would be likely to answer (to the extent that it would answer them) aren't particularly interesting, but to a few folks they're fascinating. As long as those people are responsible - and I think they're likely to be, based on who's piping up here - I don't have a problem with it.

The only sports-team-logo clothing either of my kids have is this.

An excellent friend gave us a bib with this (just the skater, though) embroidered on...
posted by nickmark at 5:28 PM on March 4, 2011


But hey, it's not my time or my research project, so have fun with it.

And the point really is just to have fun with it, because there are a bunch of us who find this kind of analysis fun.

But I don't want to do something that's going to interfere with other people's enjoyment of participating on this site. If it makes people feel like their activity here is being used to prop up stereotypes, or that this idea is adding fuel to a larger and problematic "divide the world into men and women" sort of dynamic, then that's not good. So I really am interested in the things people are saying here in this thread, to see if there's any space to answer the original question that doesn't cause those kinds of problems.

Some of the potential difficulty can be avoided through carefully reviewing what turns up in an analysis. If we turn up with a bunch of arguably stereotype-supporting comments that have heavily unbalanced male vs. female favorite counts, for instance, that's the kind of thing I'd look at and just say, no, let's not post that.

But if people are significantly upset by the whole idea of looking at the site activity data categorized by gender, regardless of the outcome, then that's tougher to work around. It'd be more of a "let's just not do it, then" thing in that case, at least for me.

Anyway, this isn't something that would be happening in the next day or two by the sound of it, so there's plenty of time to talk it over and figure out what, if anything, we can do to answer the original question without upsetting a bunch of nice people.
posted by FishBike at 5:50 PM on March 4, 2011


> Some of the potential difficulty can be avoided through carefully reviewing what turns up in an analysis. If we turn up with a bunch of arguably stereotype-supporting comments that have heavily unbalanced male vs. female favorite counts, for instance, that's the kind of thing I'd look at and just say, no, let's not post that.

So... perhaps I've completely misunderstood what you've written, but it seems like you're saying, "YES! DATA! Um... unless I don't like the results of said data."

Certainly, do this or don't do this, but it would be nice to have some planned consistency about the outcome.

Am I misinterpreting your intent?
posted by darth_tedious at 6:01 PM on March 4, 2011 [1 favorite]


that's the kind of thing I'd look at and just say, no, let's not post that.

You know, I think that withholding data which has already been collected because you don't like possible interpretations/conversations which you think might possibly arise is as much an act of dishonesty as fabricating data from whole cloth. If you think that you might withhold results, for whatever reason, then I would urge you not to begin the project at all. (Note: I'm not making any statement one way or the other about whether the collection of this data would be in any way valuable or insightful.)
posted by frobozz at 6:23 PM on March 4, 2011 [2 favorites]


Well, when you leave out people who specifically choose not to gender identify in a recognizable way, you leave out a lot of the diversity even amongst those of us who choose to identify as a woman or a man.

Absolutely. That's one of the explicit limitations of any such analysis going in, and recognizing that when talking about how any analysis is done and how it is interpreted is very much an important (and in my opinion fairly interesting) part of the process. The idea of this sort of thing would not be to say Okay, Folks, This Is How It Is. Not for the folks captured in the data of the analysis, let alone the folks who aren't.

I'm just not a fan of reinforcing the gender binary using non-representative data for questionable benefit.

Me either. I can't stand shitty pop science reporting or schism-bait bullshit like Louann Brizendine's horrid books. Using non-representative or mis-represented data to reinforce stereotypes (or sell papers/books/whatever) sucks.

But talking about data for its own sake, with full recognition of and emphasis on the fact that inter-group variation tends to be dwarfed by intra-group variation, is a far cry from that kind of dung-headed axe-grinding or exploitation. Being cautious and thoughtful about the stuff and keeping it in context is the important thing; at a certain point beyond that, the desire to not do analysis that could in some way be interpreted as supporting one or another stereotype or socially-constructed partition means just never doing any analysis at all.

I don't like people who do crappy misrepresentative stuff with data, but I also don't like settling on any sort of entrenched incuriousness.

So... perhaps I've completely misunderstood what you've written, but it seems like you're saying, "YES! DATA! Um... unless I don't like the results of said data."

I think FishBike is trying to say "this sounds like an interesting project but I don't really want to start a fight by presenting results", which I can understand and I know from talking to him a lot previously is motivated by a desire, as he's expressed in here, to not accidentally put people out. I think he's trying to be more cautious on that front than he needs to be—I have a hard time imagining a plausible end result that'd actually be problematic—but I appreciate the hell out of that willingness to put not pissing folks off above the joy of datawankery.

I think by the time we get as far as crunching data we'll have talked through some of this stuff more and figured out what makes sense to look at and what doesn't; right now it's just brainstormy stuff, and in that context trying to account for Good Idea and Bad Idea stuff is part of that.
posted by cortex (staff) at 6:24 PM on March 4, 2011 [1 favorite]


Am I misinterpreting your intent?

No, I think you've got it. I think it'd be interesting to see the top 20 lists of comments that Ashley801 started this thread to ask about. But if I generate that list and look through it, and it's exactly what people who have expressed concerns here were worried it would look like, then it seems like it would be a bad idea to carry on and post it anyway.

You know, I think that withholding data which has already been collected because you don't like possible interpretations/conversations which you think might possibly arise is as much an act of dishonesty as fabricating data from whole cloth. If you think that you might withhold results, for whatever reason, then I would urge you not to begin the project at all.

Well, for me it's more of a matter of not publishing results of an analysis if it's going to upset people or lead to a fight, rather than, say, hiding results that would otherwise defeat some argument I'm trying to make (since I'm not trying to make one at all).

We run all sorts of queries in these datawankery discussions and often do not post the results because they're boring or don't really end up answering the question. And I know there's at least one I ran that returned basically a list of comments by moderators in metatalk announcing user bannings, and I didn't think that was a good thing to draw more attention to.

The difference here is that with previous ones, they're all based on the standard Infodump data, so anybody else who wanted to go look at the same thing was able to do so. Since that might not be the case here, that's a significant difference that's worth talking about.
posted by FishBike at 6:41 PM on March 4, 2011


Hey sonika, you've seen ThinkGeek's baby section, right? Because I was recently shopping on there for a friend's little peanut and OMG, baby Star Trek and Superhero onesies.

(Also a really cute Cthulhu shirt.)

(Also someday I am going to be punished with a non-Geek kid)
posted by PhoBWanKenobi at 6:48 PM on March 4, 2011


But if I generate that list and look through it, and it's exactly what people who have expressed concerns here were worried it would look like, then it seems like it would be a bad idea to carry on and post it anyway.

With all respect--should we relish the sensation of being judged unable to handle the truth? Is the value of factual information now judged based on how much you think it's going to upset people or lead them to fight?

I realize this is perhaps a very small issue of no actual consequence to me and do not begrudge you your actions. But I wonder if our world really needs more gatekeepers with their own constructs of our best interests at heart.
posted by Phyltre at 7:29 PM on March 4, 2011


People do need to be protected from the Terrible Secret of Demographics.

Also, youtube commentors are trying to beanplate that video. Cute.
posted by mccarty.tim at 7:36 PM on March 4, 2011 [1 favorite]


the value of factual information

I really get what you're saying (in the full comment) but I also think it's important to recognize that factual information is often falsely conflated with significant information (separate and distinct from the question of statistical significance), frequently with no good result. I think it's that tendency to conflate (intentional or not) which is giving FishBike pause (forgive me for speaking for you here, and correct me if I'm wrong).

It's partly because of that that I personally don't have a lot of interest in this sort of project, though the larger reason is that this particular sort of datawankery isn't my preferred flavor. Frankly, FishBike's up-front admission that if his view of the results is "wow, that would create a lot of bad feelings and probably more heat than light" then he wouldn't post the results is part of what makes me okay with the idea.
posted by nickmark at 8:14 PM on March 4, 2011


What it comes down to is that I'm responsible for whatever it says in the comment box when I press the "Post Comment" button. That's true whether it's some words that I wrote, or some query output that I pasted in there. This data analysis stuff is supposed to be fun, so if I come up with something that I think will cause the opposite of fun by posting it, then I just won't post it.

Normally that's not a problem. We don't require people to post every useful fact that occurs to them during a discussion, or every query they run on the Infodump. For the latter especially, it's available to everyone and so if I don't feel like posting a particular result, nothing stops somebody else from posting it and being responsible for doing so.

So maybe the key thing here is that there's a community expectation that if anyone is given special access to data that's not easily available to others, for the purposes of doing some specific analysis with it, then part of the deal is you have to post the results regardless of what they are.

I don't think that would be an unreasonable expectation, and if that's really how people feel about it, then we probably shouldn't go any farther with this particular idea, because it's a touchy subject and it seems too risky to commit now to posting something that nobody has seen yet.
posted by FishBike at 9:20 PM on March 4, 2011


Well, for what it's worth, FishBike, as to this issue:

So maybe the key thing here is that there's a community expectation that if anyone is given special access to data that's not easily available to others, for the purposes of doing some specific analysis with it, then part of the deal is you have to post the results regardless of what they are.

I, for one, wouldn't have any kind of expectation like that towards you. Or any kind of expectation period. After all, in total seriousness, anyone has the option to build a reputation for decency and thoughtfulness on the site, and to become a trustworthy datawanker, and then perhaps they can take their own shot. It doesn't appear as if, if you choose not to do something, then everyone else would be barred from it forevermore. In other words, I don't think it would be an abuse of special access to data, because anyone has the opportunity to develop a history such that they might have special access to data at some point too.
posted by Ashley801 at 9:44 PM on March 4, 2011


sonika: Ashley? That reads very obviously female to me, seeing as it was the most popular girl's name in the US not too long ago.

Know what I've always wondered? Why English people don't like my name. It was in the US top 5 for something like 20 years ... every other name on the US top 5 has been at least in the English top 50 in the same year, usually top 10. Ashley's never cracked 100.
posted by Ashley801 at 9:46 PM on March 4, 2011


>>Is the value of factual information now judged based on how much you think it's going to upset people or lead them to fight?

What else would the value of information -- or the value of anything -- be judged on if not its consequences?
posted by J. Wilson at 9:56 PM on March 4, 2011


There are lots of kids named Slappy. I name them before I slap their parents, pro-actively. Then I'm happy.
posted by Slap*Happy at 10:48 PM on March 4, 2011


"I really dislike the implication that I favorite things because of my genitals."

That's okay, I really dislike the implication that I favorite things with my genitals.

Granted, no one's made it, but I still don't like it.
posted by Eideteker at 11:16 PM on March 4, 2011


I really dislike the implication that I favorite things with my genitals.

It's the only way to get through the Lady Gaga threads.
posted by NoraReed at 11:34 PM on March 4, 2011


Green - not blue - is actually the complimentary color to red, but perhaps the pre-Victorians recognized it looks like shit on babies. The Victorians, taking the FUN out of, well, everything since 1890.

This made me curious about the history of availability of dyes throughout time so I googled it and came up with this neat timeline which is not particularly relevant to the discussion at hand but I LOVE COLOR; COLOR IS INTERESTING.
posted by girih knot at 11:37 PM on March 4, 2011 [3 favorites]


My genitals have their own login name at Metafilter [...].
posted by found missing at 12:38 PM on March 4


Makes sense. Your regular login name would be a pain for them to use.
posted by rjs at 1:35 AM on March 5, 2011 [1 favorite]


Is the value of factual information now judged based on how much you think it's going to upset people or lead them to fight?

Posts are judged on that basis all the time on Metafilter. As a community it's often more important to consider their impact than holding facts on a pedestal — in my opinion.

Of course there's all kinds of data available. I'm interested in politics. Which posts/comments mentioning Israel/Palestine have self-identifying Americans favorited the most? How about non-Americans? That'd be interesting, wouldn't it? There are a lot of questions that break people down into categories that could produce interesting data. Should they be asked? Is it a good idea for the site?
posted by empyrean at 3:20 AM on March 5, 2011 [1 favorite]


What else would the value of information -- or the value of anything -- be judged on if not its consequences?

The real heart of my point is that in this case, there's no comprehensive idea of the consequences. Assuming "worst possible" data, we can likely expect some fighty threads or some poorly-considered discussions. Even then, the idea that statistical analysis should be kept quiet just because it might lead to some people who run with it in the wrong direction seems condescending on its face. Because at the end of the day you have someone who is saying, "I don't think people can handle the data I have collected." My internal response is, "Who do you think you are to be a gatekeeper based on your perceptions of our response and the sum of its effects?" I mean, real data statistical isn't always easy to come by. I'd be hard pressed to even contemplate that what we really need is less.

Sure, judge things on their consequences, but don't judge things before the consequences are even known. And even then--consequences aren't always going to be evident. That's the problem with judging things. Lots of things like social rights movements, or revolutions and so on have lots of negative repercussions up front that take years to resolve into positive effects. Historical analysis is something academics spend their lives on.

Summing up here! No such thing as undesirable facts...predictively judging net-worth consequences for others is hubris...disruptive information is not as a result negative information. That's my philosophical take.
posted by Phyltre at 6:35 AM on March 5, 2011


FishBike, nobody is going to be justifiably angry at you if it turns out on average self-identifying males and females in the US like to read slightly different things.

I hate that ftfy thing, but let's be honest, people get angry for almost no reason here all the time
posted by jessamyn (staff) at 8:33 AM on March 5, 2011 [1 favorite]


Thanks for trying to answer my question, everyone. I have a better understanding. I guess I'm just frustrated at why people get offended over this sort of thing.

In my mind, I have a hard time understanig what exactly is wrong with averages.

We all want to shine like a star, and I'm all for it, but the sky is full of stars and no star is less beautiful because of it.
posted by royalsong at 9:44 AM on March 5, 2011


I LOVE COLOR; COLOR IS INTERESTING.

Have you read Josef Albers? Great stuff. If, um, you're a complete nrrrd who likes looking at color wheels. Not that I know anyone like that or anything...
posted by sonika at 10:08 AM on March 5, 2011


I thought it was already established that, even with a free-form text field for gender, MeFi can reliably determine a user’s sex 75% of the time. 0.75 of a giant data set sounds like a good corpus to start with.

Hence the factual basis of Jessamyn’s immediate dismissal of the idea seems to be void, which then leads one to question why this basic demographic question “isn’t a road we're likely to go down.” At the very least, can’t it be turned into a giant infographic that Google and blind people can’t read? Why is this a question that can’t be researched?
posted by joeclark at 10:39 AM on March 5, 2011


I thought it was already established that, even with a free-form text field for gender, MeFi can reliably determine a user’s sex 75% of the time.

Established by whom, when, and how? What are you referring to?
posted by cortex (staff) at 11:17 AM on March 5, 2011


ND¢: "My posts tend to be very popular . . . with the ladies."

So if I favourite this, that makes me a little bit gay?

Assuming you're a dude.

Because I can't tell from your profile.

But if it turns out you're a woman, I'm going to favourite this so hard.
posted by bwg at 3:56 PM on March 5, 2011


Cortex, joeclark's referring to this comment by some dude named mathowie:
As a data point of this in action, we’ve had an open gender field at MetaFilter for over 9 years now, and some simple analysis showed that we could still predict how a user wanted to be referred to themselves over 3/4 of the time if we wanted to (about 15-20% are completely unique).
Which links to this post by some dude named cortex. I don't see the 75% Matt is talking about but that awake thing and I have been having an argument all day.
posted by mrmorgan at 4:34 PM on March 5, 2011


Do I contradict myself? Very well, then I contradict myself, I am large, I contain multitudes.
Walt Whitman
posted by found missing at 5:35 PM on March 5, 2011


Ah, I read that assertion as very different, that someone had a means of simply and reliably establishing the gender of those who hadn't put an unambiguous, or any, label in the gender field, one that had a 75% success rate. Which would have been a hell of a thing. Your reading makes much more sense.
posted by cortex (staff) at 8:01 PM on March 5, 2011


Thinking about this conversation this morning, I realized that something that seems a lot more interesting to me is the question of whether there's a pattern to a user's joined-on date and the way they use favorites. Seems you could avoid the gender-related and data-quality related issues looking at that, and you might (possibly) learn something interesting about the way people use favorites as well as what they favorite. Just a thought.
posted by nickmark at 2:03 PM on March 7, 2011


Thinking about this conversation this morning, I realized that something that seems a lot more interesting to me is the question of whether there's a pattern to a user's joined-on date and the way they use favorites.

I gather it'll be a little while before the gender field data could even be made available, so since we've already got this datawankery thread going, what did you have in mind specifically?

I'm picturing something like a graph with join date as the X axis and favoriting rate as the Y axis. We'd probably have to look at the favoriting rate over a relatively recent period of time (say, the past 12 months) to try to control for site-wide changes in that over time. For example, the fact that the feature didn't exist when a lot of long-time users signed up.

And we'd probably have to find some way to control for overall user activity levels, too. People who joined up a long time ago are probably more likely to have wandered off than those who signed up just recently.

So maybe not favoriting rate, but favorites:comments ratio? Note: not the usual favorites received per comment made ratio, but the ratio of favorites given to comments posted, for each user.

Is that the sort of thing you're going for?
posted by FishBike at 5:54 AM on March 10, 2011


« Older So that's how you pronounce it!   |   This is why we can't have nice things Newer »

You are not logged in, either login or create an account to post comments