top of page


Social Media Experimentation: OKStupid

Apparently, the revelation that Facebook users were unwitting participants in online behavioral experiments was far from an isolated incident perpetrated by a empathy-lacking, big-brotherly, multi-billion corporation. On the contrary, with the latest revelations about OKCupid's shenanigans, an internet dating service,  it is becoming apparent that users of online application & services should surf at their own peril.


The OKCupid experimentations in its most egregious involved falsely matching couples, to check the power of suggestion and in its more benign run some profiles with pictures and no text, and vice versa, to measure the comparative effectiveness of visual vs linguistic cues.


No surprise of course that people were more respondent to pictures rather than written extolments . To quote Christian Rudder, one of the founders of OKCupid, "So, your picture is worth that fabled thousand words, but your actual words are worth almost nothing".  Words are by definition wholly fabricated and their relationship with truth or reality is tentative at best, especially when the purpose is to impress as in this case, whereas a picture (assuming no extreme manipulation has occurred) is a relatively more accurate indication of reality (emphasis on "relatively").


But in a broader sense, and certainly from an evolutionary point of view, vision has always taken precedent over our other senses and especially the linguistic component that is far a more recent development.  Or to put is more crassly, most of us would rather risk going on a date with an attractive (to us) person not knowing if he/she is dumb, uninteresting or boring, rather than going out with a stimulating personality of unknown appearance. We are talking about a hooking-up website after all and not the peer network of the Sorbonne.


And let's not forget the modus operandi of web-surfing is slowly and insidiously corroding our attention spans, making us skim, scroll and browse rather than read, study or pore over.  Glued perpetually to our notebook, tablet or phone screens, we are conditioned to respond more fervently  to visual rather than linguistic stimulations as the former can be more quickly processed ad consumed in order to move on to the next page, tab, link etc.  In other words, if date-seeking people were keen on reading resumes that much they would have stuck to their local rag's classified section.


The other type of OKCupid's stealth testing, telling users that they matched with people that they actually did not, and vice versa, and which showed that to some extend the suggestions do alter the users' willingness to engage with one another, is also commonsensical . If the users of such services did not heed the computational-algorithmical suggestions, they wouldn't have enrolled in the first place. Not that that the false matches resulted in a bonanza of blissful unions mind you, we are talking about nuances of single percentage points here.


It would be interesting at this point to speculate the types of mismatched couples that were encouraged to engage with one another, and in some cases gone out with, and the eventuality of their hitting it off.


>> Vegan freelancing graphic designer in her 40s out on a dinner date with 35-year-old C-level executive in retail industry, watching him devour a bloody steak in abject horror while spooning through her poached cabbage & roasted parsnip crumble.

(Relationship Forecast: Not good except if there's heavy drinking involved in which case there's a high possibility of making out in the back alley.)


>> Wall Street investment banker on a Sunday afternoon stroll in Battersea park with feisty "Occupy Wall Street" brunette activist, the former aghast at the latter's scathing disparagement of his industry, the latter with the banker's hair styling.

(Relationship Forecast: Hate at first sight.)


>> Devout evangelical suburbanite on an after-church brunch with testy agnostic liberal transgender woman, praying for Sweet Jesus to deliver him from her date as soon as heavenly possible.

(Relationship Forecast: Abysmal.)


>> WWF volunteer & part-time stray animal rescuer on a river cruise date with Oil Industry lobbyist career-orientated gal, tempted by her curves, repulsed by her principles, or more precisely her complete lack of any principles not involving her beauty regime.

(Relationship Forecast: Not favorable except if, on the off chance, she has a soft spot for Pomeranians puppies.)


>> Keynesian minded Federal Reserve economist meets Austrian School minded blogger for early drinks in uptown bar, soon both livid about one another's approach to financial regulation, local taxes and Obama's legacy. (Relationship Forecast: Potential intercourse for anger relief.)


>> Edward Snowden skyping Megyn Kelly

(Relationship Forecast: It could work, if they both exploit each other for persona gain.)


>> Ann Coulter on a blind date with Bill Maher.

(Relationship Forecast: Tempestuous yet entertaining.)


As far as the ethical objections about conducting this sort of online testing, they seem slightly superfluous after willing having handed our personal data to the various social media and other internet applications we use. So fine, we are occasionally tested on, but it is all innocuous after all, we are not being brainwashed, like for example when watching Fox News. Plus, as OKCupid's Mr.Rudder writes, it is the only way for all these online services & platforms to test, calibrate and improve what they are doing. Brick and mortar stores have been doing that to their customers for decades now.


In any case, internet users should at least be vigilant of such stealth experimentation when online. Amazon could be subliminally using its recommendation & suggestions listings to test how long it takes for your Miley Cyrus distaste to sizzle into deep adoration, or your Dan Brown aversion to be flipped into mild affection. Instagram could be serving you exclusively monochromatically themed photos to test the psychological impact of color on mood or monothematic collections of food styling  to study the malleability of your eating habits. Google could also be manipulating the auto-complete in your searches to temp you or piss you off, e.g. googling "pro-russian separatist" auto-completed as "pro russian seductress", or "rosary beads" to "rosario dawson playboy sperad".

bottom of page