The interesting thing is that if taller and more well-to-do men are more attractive to their mates, would it stand to reason that the average OkCupid user might actually trend lower than the national average, thus making the exaggerations more pronounced?
Or would it be that the taller and more well-to-do are more confident in their abilities to date and thus trend above average?
...so if I ever set up a profile on OKCupid, I should list myself as four inches taller than my actual height, since everyone who reads my profile will assume I have overestimated my height by two inches.
Presumably this is like tilting a pinball game, and people are willing to exaggerate just enough not to have people call it out the moment they see you.
I wonder if there's a similar effect in populations of men where metric heights are commonly used.
6 ft ~ 183 cm, which isn't a particularly round number.
Maybe that psychological benchmark occurs at 180 cm, which suggests Europeans and Asians aspire to be approx. 5'11.
OKCupid's usually pretty good about data, but they blew it on a couple of these. For example:
Their method of controlling for income only takes into account zip code. There's another obvious source of bias: the fact that these people are on an online dating site probably indicates that they are time-poor, have decent computer skills, and are in professions that value introverts. Many of these jobs are in high-paying fields, eg. software. People who work 9-5 in blue collar jobs often go out to bars to pick up women instead. It's no big surprise that the average OKCupid user makes more money than the average person in general.
Side note: there're a couple interesting things in the raw data for income. One is the big dip in response rates for 23-24 year olds that say they make $100K+. The other is the same dip in response rates for 29 year olds. I wonder if those are people assuming that the other person is lying.
The bisexual angle is also easily explained by looking at the profiles of the people you're aggregating. Many girls say, as the top line of their profile "NO MEN! I'm married, but looking for a bi or lesbian girl to have fun with on the side." They are bisexual, but they found the partner they want for one gender already, and are only looking for someone of the opposite gender.
The height stuff was interesting though, particularly since I listed myself as 5'11" despite thinking that I'm only 5'9". I finally revised my estimation of my own height upwards last year, when I found out I was at least an inch taller than my sister's boyfriend, who she swears is 5'10".
It'd be biased. My guess is that the "Rather not say" distribution has fat tails. It includes people whose income is abnormally low, because they don't want women to know and hence be scared off, and it also includes people whose income is abnormally high, because they don't want to attract gold-diggers.
their height data makes me wonder about their conclusion. if the average height data includes all ages, i think the differential could be more readily explained by the fact that okcupid tends to skew much younger than the average population sample, and young people are simply growing taller now on average due to a higher quality average standard of living.
Your theory doesn't explain the sudden jerk around 6'. Your theory doesn't explain why women self-report as shorter than the overall population. And also, anecdotally, I've known women to complain about men lying about their height on online dating sites. If it is common enough for women to complain, then it should show up in their statistics.
i have no doubt that lying does happen. i also have no doubt that i have no data to back up my theory. but it seems somewhat unlikely that lying about your height is so uniformly prevalent that it affects the curves as uniformly as it does.
also, the article doesn't say that women self-report shorter. they also self-report taller. almost exactly the same as men. which again makes me feel like its just a growth thing.
just throwing out an alternative theory, though. i could be wrong, of course. i'd just rather see the data pulled out better -- self-reported height by age, vs national average by age.
I wonder why the (putatively) tall users exaggerate their height in the same direction as the shorter users. Wouldn't the tall men (or women) have a tendency to present themselves as closer to the mean? Or if not that, to simply report their actual height?
Does additional height confer an advantage even for the already tall?
Or would it be that the taller and more well-to-do are more confident in their abilities to date and thus trend above average?