Perceptions of headscarf survey

January 29, 2008 5:55 PM

A recent poll (PDF) asked for reactions to the same model dressed in two different ways: in a plain shirt with her hair down, and in a blue head scarf of the style of some Islamic women. Perhaps understandably, the survey respondents felt the scarfed image was more traditional and more religious. But some of the other perceptions are less obviously predictable. (via crooked timber)
posted by Rumple (45 comments total)

I thought this was going to be about a potential incumbent for the LDS presidency. Imagine my surprise.
posted by boo_radley at 6:06 PM on January 29, 2008

The greatest tragedy is that almost all respondents overestimated her age!
posted by grobstein at 6:13 PM on January 29, 2008

Also: how come no one answered, "I don't know"?
posted by grobstein at 6:14 PM on January 29, 2008

Hmm...I honestly tried, but I can't find a scrap of information about the methodology used in this study. No mention of the number of participants, their location, background...nothing. For all I can tell, they got a bunch of their friends to take a survey in exchange for beer.

I think the technical term for this type of study is "garbage."
posted by mullingitover at 6:14 PM on January 29, 2008 [5 favorites]

So wait... the fact that more people think the scarved woman is beautiful means that the unscarved woman's uncovered hair is *not* emitting more rape-inducing sex rays that drive men to violent lust?

This is infidel propaganda.
posted by Krrrlson at 6:14 PM on January 29, 2008 [5 favorites]

mullingitover: As far as numbers of participants, it says n=306 and n=312 for the two pictures.
posted by Midnight Rambler at 6:17 PM on January 29, 2008

grobstein: There probably was no such option.
posted by ErWenn at 6:18 PM on January 29, 2008

Midnight Rambler writes "mullingitover: As far as numbers of participants, it says n=306 and n=312 for the two pictures."

Ah, excellent. So where did these people come from? Whomever self-selected for the survey ad on Free Republic? Daily Kos? 4chan? My Little Pony Arena?
posted by mullingitover at 6:20 PM on January 29, 2008

Heh. I was dubious about the statistical significance of the differences until the last page.
posted by danb at 6:24 PM on January 29, 2008

re: methodology, it looks like you might be able to join their survey panel, though it isn't clear what happens next. Their press release gives a miniscule amount of information, noting their survey was "nationwide".

Their home page suggests that for their upcoming superbowl ad survey, you might get paid for taking part.
posted by Rumple at 6:38 PM on January 29, 2008

That woman is clearly a master of disguise!
posted by aubilenon at 6:38 PM on January 29, 2008 [2 favorites]

I'm skeptical that there's no information about the sample. It could be the same group of people answering the questions over and over again. What age group are the respondents? What races are they? What are there ethnic backgrounds? What is their education level? How did you find them? Why are they studying this? Where did they do the survey?

Some of the results are kind of interesting, but without anything to compare it against, we can't really say anything about them aside from "huh. cool." And even then, you really can't be sure as to why.
posted by Gular at 6:42 PM on January 29, 2008

The parent company has a little more information on their methods in general, but I can't find more than that.

If this was a perfect random sample I am guessing the +/- for each question is going to be about 4%.
posted by Rumple at 6:48 PM on January 29, 2008

Since I know there are plenty of statgeeks here, is there not a name for boxing your subjects into a corner? It seems pretty obvious that, were "I don't know" an option, that would get a large percentage of the answers for each question. Is there a word for not including obvious options in a survey and thus damaging your results by forcing people to choose unreasonable positions?
posted by odinsdream at 6:50 PM on January 29, 2008

Newsflash: People judge book by cover.
posted by wfrgms at 6:58 PM on January 29, 2008

Anyone could have predicted these boring results.
posted by kozad at 7:02 PM on January 29, 2008

A reputable poll should always have some language like A nationally representative random sample of XXXX adults was interviewed by telephone between Date A and Date B, Year. The margin of sampling error for the survey is plus or minus Y percentage points; for results based on subgroups, the sampling error is higher in an visible place on the survey, press release, and/or web page.
posted by rtha at 7:05 PM on January 29, 2008

Looking through the results, I can't find anything that looks like a significant difference that didn't seem obviously predictable to me, except for the ~10% difference on the "beautiful" response. Is anyone seeing anything surprising?

uncanny hengeman: Please keep the locker-room BS to yourself. We really don't want to know what you would or wouldn't "hit".
posted by ssg at 7:08 PM on January 29, 2008 [1 favorite]

Since I know there are plenty of statgeeks here, is there not a name for boxing your subjects into a corner?

I'm not a statgeek, so I might be talking out of my ass, but I don't quite see why it should matter if the subjects were "boxed in."

If two statistically similar groups of people are forced to make the same unreasonable judgment, there's no reason both groups shouldn't make it the same way, right? It seems that the point of the study was showing that the two unreasonable judgments were actually substantially dissimilar from the perspective of the two judging samples, but the only difference presented was obviously the presence of a headscarf.

Giving an "I don't know" option would've spoiled things, because the researchers weren't trying to gather people's considered opinions.
posted by Mr. President Dr. Steve Elvis America at 7:16 PM on January 29, 2008

Actually, ssg, that's not entirely true. If uncanny hengeman was going to hit me, in either the traditional fashion or in the wink-wink way, I'd definitely like time to prepare beforehand.
posted by maxwelton at 7:20 PM on January 29, 2008

Hey can we make with the more joking please? Bar graphs are boring.
posted by ninjew at 7:59 PM on January 29, 2008

So wait... the fact that more people think the scarved woman is beautiful means that the unscarved woman's uncovered hair is *not* emitting more rape-inducing sex rays that drive men to violent lust?

Nah, it's just that the uncovered woman has a crap hairdo. She looks like the kind of person who'd crochet bible verses as a hobby.

A short, spiky, lopsided, layered look would suit her so much better.
posted by UbuRoivas at 8:03 PM on January 29, 2008

I'd be horrible at this type of survey because I'd be begging for an "I don't know" option and would probably not choose anything.

How do they find people willing to answer such questions? Would that be a big factor in the answers?
posted by ODiV at 8:04 PM on January 29, 2008

I found several of the results surprising, and all informing.
posted by jb at 8:09 PM on January 29, 2008

The only response that seemed "less obviously predictable" was the overwhelming preference for the picture with the headscarf as being "Beautiful". That was very surprising, actually, but then led me to wonder again about the sample population.
posted by voltairemodern at 8:11 PM on January 29, 2008

[a few comments removed, "I'd hit it" clock is reset, feel free to complain in metatalk]
posted by jessamyn at 8:29 PM on January 29, 2008 [1 favorite]

The only response that seemed "less obviously predictable" was the overwhelming preference for the picture with the headscarf as being "Beautiful". That was very surprising, actually, but then led me to wonder again about the sample population.

I'm still going with the hairstyle as the deciding factor. The way it just droops down to the points of her shoulders, it probably makes her look at least 5-10 kilos heavier than the scarved version, and it does nothing to accentuate her face. Rather, it draws attention away from her face, without adding any net benefit, in beauty terms. Worn differently, it might be another matter, eg tied up into a ponytail, or with a different choice of top - not a dark colour, but something bright that breaks up that great solid mass of black. Or at least with some better accessorisation, like some silver earrings, or some kind of pendant necklace. Anything to break up the lines a bit.
posted by UbuRoivas at 8:31 PM on January 29, 2008

maxwelton: If uncanny hengeman was going to hit you, he could MeMail you. It would only be polite to keep that out of the thread.

uncanny henreman: I'm not daft; I see your joke, but I don't find it funny. If you weren't aware, we recently discussed this very issue. Sexist talk and jokes make some people uncomfortable and "I'd hit it" jokes were singled out as a particular problem. Please keep such jokes to yourself out of respect for other members of the site.
posted by ssg at 8:33 PM on January 29, 2008

WTF is interesting about those results? Entirely predictable, and probably as fair an effort (ie likely to be right) as any gross generalisation about an individual (which is exactly what these studies are about).
posted by wilful at 8:35 PM on January 29, 2008

Truthfully, the dominant answers reflect pretty much what I would have answered myself. A headscarf strikes me as a pretty good indication that the woman is Muslim. Not necessarily, but more likely than not. I'd also guess that it's more likely than not that she's wearing the headscarf out of respect for a religious/cultural tradition -- and I say respect because she looks relatively happy, secure, and self-confident. Is she middle class? I'd say the chances are good because, as far as I can tell, she's not dressed to the letter of Islamic fundamentalist law -- and I tend to associate (rightly or wrongly) most religious fundamentalisms with poverty (and those who profit from promoting said fundamentalisms).

I make these sorts of estimations based on appearances all the time. We all do. So long as we never forget that we might be completely wrong in every individual case we ever encounter, I don't see any problem with (or escape from) that.
posted by treepour at 8:43 PM on January 29, 2008

"I'd hit it" clock is reset, feel free to complain in metatalk

Er, no thanks.

Carry on, ladies!
posted by uncanny hengeman at 8:54 PM on January 29, 2008

Oh, I haven't seen that one! Is it as funny as the other movies in the series?
posted by UbuRoivas at 9:26 PM on January 29, 2008

Is anyone seeing anything surprising?

I wonder why so many people thought she was a Catholic. Can somebody "look Catholic?"
posted by Afroblanco at 9:53 PM on January 29, 2008

Can somebody "look Catholic?"

Well, you can certainly look like a member of an ethnic group that tends to be Catholic. Latin and Latino people are prime examples. In this case, the only guess I could make is that the woman looks as though she might be have Irish blood and hence look Catholic, but who knows what people were thinking.
posted by ssg at 10:08 PM on January 29, 2008

Not sure how I'd respond to this survey. The first lady looks.. normal? I don't think I'd make any assumptions (unless coerced by say, a survey, to do so).

However, a woman wearing a hijab definitely has more significance to it than "a simple headscarf". I think perhaps people picked "beautiful" cause they didn't want to sound judgmental - her face is identical either way (Photoshop maybe?).

Beyond that - I don't see that this survey proves much of anything : we, as a culture see something that's generally of religious significance and assume.... the person adheres to that religion? Ooooh... shocker!
posted by revmitcz at 10:43 PM on January 29, 2008

Am I the only one who noticed that the hair-down woman is both more sexually straight and more likely to be a lesbian? Numbers didn't look significant, but come on, people, wtf?
posted by agentofselection at 11:22 PM on January 29, 2008

It's kind of weird that the surveyors apparently didn't know the adjective form of "Islam". (Christian, Jewish, Islam?)
posted by mendel at 5:03 AM on January 30, 2008

Am I the only one who noticed that the hair-down woman is both more sexually straight and more likely to be a lesbian?

No, I caught that too. I figured it had to be due to a small part of the surveyed population that didn't understand what either "Sexually straight" or "Lesbian" meant, or that they were opposites. Strange, to be sure.


It's kind of weird that the surveyors apparently didn't know the adjective form of "Islam". (Christian, Jewish, Islam?)

Yeah...weird and annoying.
posted by voltairemodern at 6:06 AM on January 30, 2008

Even without the dubious sampling, the survey is useless because it has no kind of control or contrast.

It might mean more if there were other factors, like different hair colors, different hair styles, etc.

As it stands, we can't know... we can't even presume... that it is the scarf itself causing the distinction.
posted by vertigo25 at 9:01 AM on January 30, 2008

Actually, I think it would be interesting to compare the survey results to the woman's actual information. However, things like "has a sense of humor" and "a planner" would be difficult to quantify.
posted by Afroblanco at 10:37 AM on January 30, 2008

The only thing this taught me is that, apparently, a majority of people would prefer that a poorly-photoshopped picture of a woman not be their neighbor.
posted by googly at 10:52 AM on January 30, 2008

I'm no statistician, but isn't n = 300 awfully small? (Not that it matters, this study is a piece of crap for other reasons.)
posted by proj08 at 4:31 PM on January 30, 2008

jessamyn: "I'd hit it" clock is reset

... Huh? *digs around* Aha!

jessamyn: I will change my name to Cooter when there are 30 days of "I'd hit it" free MeFi, in both non-irony and irony flavors.

What, am I the only one who didn't get this immediately?

And also, I think "cooter clock" is funnier, though possibly less clear.
posted by Pronoiac at 4:51 PM on January 30, 2008

cooter countdown: March 1
posted by Pronoiac at 7:06 PM on January 30, 2008

related
posted by Rumple at 12:10 PM on February 2, 2008