I may not have found the love of my life on OkCupid, but I did fall in love… with their data analysis!
Their latest super-fun post by Christan Rudder, sent along by Rob Walker, Sara P. and an Anonymous Reader, looks at the lies people tell in their profiles. They do this not by catching any given individual in a lie, but by comparing data on their users to data on the general U.S. population. (It’s unclear what percentage of OkCupid users are American and they don’t specify if they are only looking at U.S. users, so I can’t verify that this is a fair comparison but… if they do restrict the analysis to Americans then…) Since they have 1.51 million active users, we should expect that any distributions should more or less overlap.
But they don’t…
1. Men lie about their height, reporting, on average, that they are about two inches taller than they are. In the figure below, the solid purple line represents the U.S. population, the dashed line represents the reported height of OkCupid users:
2. Women lie about their height too. Here’s the same figure for women (but with a dark purple implied best fit line; you can just ignore it):
3. People exaggerate their income, on average inflating it by about 20 percent (for this data, they controlled for regional differences in income). The figure below, however, shows that the amount of exaggeration is related to age. Both men (blue) and women (red) increasingly inflate their income up until around age 40. After that, they just keep inflating it at about the same rate.
A woman may earn 76 cents on the dollar for the same work as a man, but she can fabricate, like, 85 cents no problem.
Oh and, yeah, there’s a reason why the men are lying (no word from Rudder on the women). Income is highly correlated with how many messages a man gets (red = fewer messages; green = more):
4. It also turns out that not all of the “recent pics” are actually recent. This is especially true for pictures rated “hot.” Rudder says that “hot” photos are more than twice as likely as “average” photos to be over three years old (12% and 5% respectively).
And the older a person is, the more likely they are to upload an older photo: