Don't Be Ugly By Accident
blog.okcupid.com
blog.okcupid.com
For instance, this bit: "the general pattern is that more complex cameras take better pictures". That's certainly true. Think about the person behind the camera, though. If they own a dSLR, they're more likely to be good at composing photographs.
This, too: "a 28 year-old who used a flash is as attractive as a 35 year-old who didn't". Most people don't know how, or even when, to use a flash. Using a flash properly can and does lead to a better photograph, but it takes a lot of practice.
Therefore, if you want to have an attractive picture, have somebody that's good at taking pictures take it.
That's a very valid point, selection bias is the most-missed bias in studies of every stripe. But they're not beginners:
>And we also found similar numbers looking only at people who uploaded all three types of photos.
They didn't show a graph of that data, because this is an article rather than a formal study. Getting similar results in an intersection study is an excellent way to show that selection bias was at least not overwhelming - depending on the sample size, it might not even be significant.
The conclusion they made was pretty well supported, and seems like something you would agree with - the details they give are quick guidelines, but it boils down to "photographic skill matters".
Um.. we’re talking about the skill of the photographer here, not the skill of the model.
An example of another minor quabble, iPhones are expensive. I would expect that owning one correlates with nicer clothes, makeup, and just more/better status symbols in general. It's no news that money is attractive.
But the point stands: signaling wealth is what counts.
I recognize that my P&S camera isn't fast enough to take pictures in low light without the flash, but are there other situations where I should be using a flash?
Generally, using a flash makes my pictures look "cheap" (better than blurry, I suppose). Even when there is strong backlighting and I need to use the flash to balance it out, it still looks wrong.
Much to my wife's consternation, I usually tend to push the no-flash as much as possible (I've missed quick-moving children and other shots because of this). I suppose if I used the flash I would have actually captured the moment, but I don't think I've really seen a picture I've taken with the flash that I've found to be remarkable looking.
Good tip though if I ever find myself in the possession of a more sophisticated camera.
(Note: "perfect" means people will ask you how you did it, and be suitably impressed at your photography skizzles)
Firing flash along the axis of the lens is a poor idea almost always. Also, your pop-up flash has an effective range of ~10 feet (depending on camera), so popping it beyond that range (which I see many people do) is basically pointless. In fact, if there are objects near, what the camera will do is expose for those objects, leaving your main subjects, who are too far away, in shadow.
This is why you'll see those crazy pictures where everyone is real dark and has glowing eyes.
But all in all, the solution is to get a faster lens - there are many point and shoots that have premium, fast lenses. Going for those will allow you to skip the flash for a lot more occasions, making better pictures overall.
If the data is accurate, then the correlation is real. It might be that most people don't know how to use a flash; it might also be that natural light makes people appear more attractive, or that the position of the light source (i.e. not in the same place as the camera) does the same.
Building a reliable causality model out of correlations can be extremely difficult, but that doesn't mean that anything short of a complete causality model is worthless. In this example, dispensing with the flash would still likely improve the ratings of most users' photos (other things equal, of course), even if they might squeeze a bit more out of the result by mastering the nuances of flash photography.
Argh, no! That is correlation. More complex cameras are probably used by more experienced photographers. I would bet a high percentage of those are made by professionals.
As you can see, the general pattern is that better pictures are taken by more complex cameras.
This does not conflict with your hypothesis.
They also have better image sensors which means less degradation in low light situations, less graininess, and with better lenses, less distortion.
My take, anyways...
Amazon has a deal on that combo right now too http://amzn.to/brBxsE
However, you're certainly correct that conventional photographic wisdom is that people look better on longer lenses, for just the reason you cite -- wider angles can distort faces unpleasantly. It would have been interesting to see an analysis of lens length & perceived attractiveness from this dataset.
It depends on the sensor size. A 20mm lens in (35mm) full-frame is a very wide lens. A 20mm lens on a Canon/Nikon DSLR is the same as a 35mm lens on a (35mm) full-frame. A 20mm lens on a point-and-shoot is even "longer".
More clearly: that the pictures taken with point-and-shoot cameras are crappy is not the fault of those cameras, and pictures taken at the same time by the same person with a different camera would likely also be crappy.
Since more expensive optics usually tend to have a larger max aperture ...
and, considering that for most things I take photos for (usually something I'm selling) have a relatively low 'good enough' bar, a good camera is the cheapest way for me to get above that bar.
Strangely enough, this strikes me as hilarious instead of depressing, and I'm actually looking forward to the challenge. I'm also looking forward to the laughter and teasing when I show off the results to my friends.
And, yes, the semi-professional glamour shots can be an issue too, but in my experience they are much rarer. And at least in that case the person has enough self-esteem to try to look good.
http://cdn.okcimg.com/blog/camera/SexAndSmartPhonesByAge.png
However, I do like OKC in that they take tons of data and make actual conclusions about it. Regardless of whether or not it's eventually proven wrong, or (as stated above) there are some obvious discrepancies not taken into consideration, it's still cool to see overall trends in human psychology based on something so simple as how a photograph was taken.
My guess is that it's one or more of:
1. iPhone users tend to have more money as a group and that correlates with more sex.
2. as a group they tend to be more concerned with appearances and sensual experience and sensation seeking.
3. as a group they are more impulsive.
There is some evidence that higher-status individuals of both genders have more sex; and money is a fairly strong proxy for status in our society.
Droid Eris is not a MotoDroid is not a N1 is not a Sidekick is not a HTC Evo ...
Really, there are certain Android and Motorola phones that could have very similar demographics to the Blackberry or iPhone. Furthermore, Comparing the cameras on the phones more recent than the MotoDroid to those that came before is night and day.
This is just a series of classic correlation versus causation mistakes. Even beyond the purely statistical view of the issues here, it is clear to anyone with some basic knowledge of photography that most of their conclusions are flawed.
Hopefully they'll correct some of the wording to remove logical implication or just do a better job next time.
(it's just a joke: pity for the Pre. I kind of liked the innovative feeling they brought with them)
As a photographer this makes me want to shout "Duh."
Has OkCupid done any investigations of relationship pseudoscience/superstition (astrology, biorhythms, etc.) using their data?