Posted: 13 Feb 2020
DATA STORIES 4
Whiteness: A data story [aka: Love’s dirty veil]
[TRANSCRIPTION – Ep 4] Hi again. So this one, this episode is like a continuation of the previous. Providing the much needed upward thrust is data yet again. It’s the currency we can’t get away from. It is an all pervasive continuum, and remains at the heart of most of today’s contemporary discussion points. So here you have it; for today’s pep talk, we’ll continue the chat about data, but with a twist of Valentine’s Day. With it fast on on the approach, a brooding February 14th, Valentine’s Day is the perfect terrain for checking-in with a look at Lervvvve data. But love isn’t just a one day aﬀair, it’s a perennial challenge.
Here’s the challenge: stay with this. Are you ready?! So, let’s say ‘the data’ you inputted to the system, a dating website, ‘Sound FX’ has already crunched, pushed and pulled that data into shape to predict a compatibility score for you. It has done the same for 1,000,000 other people.
Then ‘ping’ – ‘you have been paired’. Bingo! You have a perfect date. Achieved, because according to the data you and that other people have similarly high compatibility rating. But here is the burner; you can’t see the appearance, the proﬁle photo. What do you do?
One dating website OkCupid pitched such a challenge to its users many years back, 7years ago; and which became known in the media news-cycle as the now infamous Love Is Blind Day, distorted in the media with the headline The OKCupid Experiment, a little bit scientiﬁc that way. The crucial idea behind the experiment being to measure how appearance, the ‘physical attractiveness’ of someone aﬀects our perception of others.
And it’s not the Internet that’s at fault. The black mirrors we use, our phones, the screen we use to communicate via the Internet simply act as gateways. Visual gateways. If something is not to our ‘taste’ then we generally move on. But isn’t it better to meet someone anyway. You’ve been matched perfectly, and then if you meet there will be that burst of instinct and chemistry that determines whether you like a person or not. But it seems our egos and our pathology kicks into action, forcing us to exercise an extreme level of caution.
But caution to the wind, here is this episode’s anecdote, discussing the perils and the upshot of dating:
So the personals, the printed text in the personals section of a physical magazine or newspaper. In it people write extremely brieﬂy about what they’re seeking: In just a quick sentence or two you sum up your request. The emphasis was on brevity, being brief. Being brief equated to a cheaper costing ad. And given also there was no guarantee of anyone even responding to your ad, keeping the cost low was the main aim. You see, to get the ad printed you had to run to a public callbox, dial a premium rate number and place the ad. Paying like £2 for the privilege. 5 Mobile phones call plans were pretty expensive at that time, in the mid-1990s.
That meant, people premium numbers meant double the cost. Also Internet was a one trick pony, for emailing and Googling only, and the most trafﬁcked social network websites were! Well, there was probably only one and that was Geosites. This is 1995 I’m talking about, 25 years ago when I was a just a blossoming 22 years of age. So I shimmied my way to the local public callbox, near home in the Birmingham District of Harborne and I set about leaving a voice message personal ad. In the limited amount of characters and words I stated my vital statistics: height, age, etc but also something about my interests.
And from that a personal ad in the magazine would be transcribed and appear very shortly after. 6 To ﬁlter your request you had to select a sub-category of where your ad will appear, basically your target group, or a gender match/sexual orientation: either W4W, M4M, or W4M or M4W let’s say. The subcategory I chose was M4M, men seeking men. The sort is, I’d just recently come out of the closet, as we say.
After dating girls since puberty, even through University, I was, well, gay. I always felt it, always knew it, but wasn’t emotionally now emotionally ready to go the whole nine yards. So I placed the small ad in total secrecy. One week later or so I was feverishly checking the “small ads” not for my own ad, but checking in the M4M section. These ads it’s worth mentioning are anonymously listed. No names, just a number to call and a code number to request. 7 So attracted to one of the proﬁles. Reading the text only description.
If you like the sound of someone’s proﬁle you then pay 50p sterling a minute (half a quid, a premium cost) to listen to the voice of the person who has left that ad. Already then in the public callbox, I eagerly slotted in the 50 pence coin and dialed the number. Brrrring, Brrrring, ke Ca, uuuuu- a voice requests: ‘to proceed please key in the 4 digital keycode of the person’s proﬁle you would like to hear and press the hashkey’. 8 D-D-De-D-Dee. The prerecorded message starts playing, and it follows along the lines of: “small ad number 9453:” the message begins.
But, geeez, it, was me!!! Yes, I’m such a narcissist. Err… The only ad which attracted my attention was of a description which reﬂected me perfectly. But not only reﬂected me, sadly it was me, my own personal ad. Had I really just responded to my own ad. Oh what painful comedic tragedy is to discover your only compatible partner is yourself. Comedy, pain and tragedy aside, the obvious upshot of this anecdote is that dating is not for everyone. Oh, yeah, and not forgetting that I had just outed myself as being narcissist. Oh dear!
And two glorious takeaways from that tragedy is maybe, 1) a proﬁle photo may have stopped me actually hitting on myself. And the second is more of the verbal Dorian Grey realisation that: 2) we are each looking for a reﬂection of ourselves in others. Can seeing a person’s proﬁle photo, their proﬁle pic, a deal breaker? It’s a cultural phenomenon that two people who feature with high ‘data-driven’ compatibility ratings, ﬁltered by an algorithm which has crunched 3.2 billion words of description – ‘ping, compatibility rating 98%’ – that they don’t make it to the ﬁrst date? Yeah, there could be time constraints preventing these potential couples from ever meeting, and yeah, people ‘be on these sites just to kill time’. But there’s a 98% compatibility score here; what about the missing 2%.
Maybe that 2% is a sprinkle of magic dust you cant experience if you don’t meet that persons in real. It’s a sprinkle of DNA, someone’s appearance. And it’s your appearance. It translates to the size of your cheekbones. Yes the size, orientation and angle of your cheekbones, your facial
vocabulary, is going to either going to seal the deal, ‘date-on’ or, be the deal breaker, ‘datecanceled’. It’s a bit shallow, but you know, that’s vanity. It’s a very visual terrain out there.
‘Here’s the proﬁle picture of the person you have been accurately paired with’. Press button to reveal the photo. Two response options. Press the button to accept or decline: ‘Duh, duh, duhhhhhhh…. erghhhh. No, way dude.’ or ‘Duh, duh, duhhhhhhhh… arghhhhhh. Strike’. Shallow, wallow wallow. Such is the vanity.
On this note, what we’re talking about here is how the shape of our cheekbones simply reveal our racial proﬁle. And sadly, subconsciously, yes, for the most part, it is argued that people on dating sites go for an index metric called ‘whiteness’. The experiment reveals, this is true OkCupid Experiment, it revealed what such a skewed metric ‘whiteness’ plays in selecting who gets a hiton and who is left by the wayside.
So the Love Is Blind Day. No proﬁle pics, just text. Here’s an extract from the book Dataclysm: by the experiment’s mastermind Christian Rudder, one of OKCupid’s co-founders. It sums up the what, why and wherefore:
On January 15, 2013, OKCupid declared “Love Is Blind Day” and removed everyone’s proﬁle photos from the site for a few hours. The ideas was to do something diﬀerent and get a little attention for a new service we were launching at the same time. The programmers ﬂipped the switch at nine a.m. It was a bond ﬁde pit of despair. The new service OkCupid was trying to promote was a mobile app called Crazy Blind Date. With a couple of taps on the screen, it would pair you with a person and select a place nearby and a time in the future for the two of you to meet. The app provided an interface to let both parties conﬁrm, but there was no way for anyone to directly communicate before the date. 13 Even after a quarter million downloads, it [the App] failed, because in the end people insist on seeing what they’re getting into. [They needed a proﬁle photo]. Fiasco though it was, Love Is Blind Day gave us a visceral example of what’s people do in the absence of information.
In hiding pictures but changing nothing else, we created a real-time experiment to set against the usual activity by removing what everyone else looked like: [such as the] typical biases, or racial and attractiveness skews. [The pairs of matched people] were giving out phone numbers and email addresses at a higher rate – to people they’d never seen before. [Interestingly] the 30,333 messages sent blindly, [got] a rate of 40% higher than usual. And before a picture was turned back on half of the pairs exchanged contact info (24%) of the time. When you compare them against the control data set, they should have lasted an equivalent of 5 or 6 [messages longer]. But with photos switchedon [it petered oﬀ at around] 4 [messages]. Once it was ‘lights on’ again, threads were down.
Ok then, with that information to hand, in a candy-coated nutshell, a proﬁle pic does dampen desire in people who are otherwise perfectly matched. This isn’t restricted only to dating. What about employment and career opportunities too. There are grave consequences to this.
Media pundits quickly seized the on the buzz around the experiment to draw a parallel with Facebook’s misuse of data and said OkCupid’s decision to do this type of ‘anthropological’ study with its 1 million active users, well it was bordering on unethical. It weirded people out. But the experiment had been designed in part as a way to improve how people used the website, to understand the inherent and societal biases of its members, its users, its subscribers. And so if we are talking business ethics terms, the website remained totally ethical, within the bounds of good business practice.
On the one hand, they were simply manipulating data for the purpose of streamlining its business practice; to improve its functionality of service. But with a bigger aim in mind too, one the other hand the site would provide invaluable insight and measure the impact of racial and attractiveness cues – proﬁle pics – on people’s decision making biases.
There was a lot of levity on the day the experiment hit the news cycle, and in the middle was OkCupid’s co-founder Christian Rudder. Interview after interview he articulately defended the website’s decision to meddle with data, meddle with love, meddle with people’s futures. He is on the record as having said that by removing proﬁle pics, commonly known as mug shots, we remove a layer of bias, we remove racial and attractiveness cues. Basically removing the veil of ‘whiteness’ — peddled through proﬁle pics and names (as I’ll reveal later).
If I asked you to delete your proﬁle photo? And collectively, we all hit the delete key, in synchronicity. Then all our proﬁles would lack these visual triggers. The only remaining attractiveness metric to go on, a way to ﬁgure out a person would be in what they write, the ‘about text’, their likes, hobbies and their repulsions. And that’s just what is fed into the OKCupid algorithm to reveal a ‘match percentage’ between two people. Here’s another extract from Dataclysm that helps identify:
A match percentage is the site’s term for compatibility. There is visual component to a match percentage: The number between two people only reﬂects what you might call their inner selves – everything about what they believe, need, and want, even what they think is funny, but nothing about what they look like. The largest racial groups on OkCupid [all in equal measure are] Asian, black, Latino and white. [And in a blind algorithm] race has less eﬀect on [the] ‘match percentage’ than [does say] religion, politics, or education. [In fact] to a computer [racial] categories [are neutral]: ‘Asian’, ‘black’ [‘Latino’], ‘white’ could just as easily be [astrological categories like] ‘Aries’, ‘Virgo’ and ‘Capricorn’. [However no matter how neutrally] racial [an algorithm may be, it’s us as] the users [who bring our] opinions [and prejudices with us and they] come into [major] play.
As much as the experiment made a buzz and hit the news cycle before ﬁzzling out, the wider more long lasting question the founders were hoping to answer related more to human nature: like, can scientists successfully pair people together? The short answer being, no; of course scientists can’t do that.
Reﬂecting on the idea of there being a magic formula, a love algorithm, the 3 co-founders were not at all convinced. Ultimately, love is about chemistry. But still, pitching themselves as modern
day alchemists, the data engineers set to exploiting the data that had now become their valued property and needed to convert it into more of a precious commodity. After all, they were in the business of love, not love itself, not love per se. However later on, you’ll see how in fact, one of the online dating website co-founders Christian Rudder seems to be one of the kindest of men, with the biggest of hearts.
And I only know this having eagerly read Rudder’s analysis of his time at OKCupid through the book aforementioned ‘Dataclysm’. The title is yet another zeitgeisty word evidently mashing data with cataclysm. And as usual these buzzy words usually lurk about the Urban Dictionary. Dataclysm is deﬁned literally as the intersection where data meets cataclysm; more poetically though it is an endless amount of data from which endless amounts of meaning can be extracted. And that’s Rudder’s skills as a mathematician, who thinks in anthropological terms.
So onto Rudder. Rudder then, being a mathematician, and there’s no surprise there. Isn’t it the case that most of today’s proﬁling sites stem from the minds of geeky maths nerds? Rudder like many of these nerdy types though, are completely socialized creatures. In his book the cofounder shows a sensitivity in the way he draws conclusions from datasets. If you were to date Rudder, and even though he probably identiﬁes as straight and married to a woman, without the bat of an eyelid. I get the feeling he’d equally date a guy as much as a women. He totally seems open-minded. That’s a vibe I’m getting from just the manner in which he presents and interprets data.
Anyway, back to the thread. A had a skeptical motivation for reading the book too; to determine the motives for doing such an experiment. If you search google for OkCupid Experiment, Love Is Blind Day, there are such superﬁcial hits in the return results. There’s very little analysis of Rudder’s or OkCupid’s deep motivations for running the experiment. And I wanted to ﬁgure out if the website’s co-founder Rudder, was he just another guy with a completed ego to exploit, people’s data as just another ‘get rich quick’ scheme? But no, such a crude use of this precious data was not part of his agenda, nowhere in sight.
Since the experiment, he has ensured, at least through publication of the book, that much of the data and its interpretation are public and which you can read about in Dataclysm.
By contrast to OkCupid’s open access policy and its willingness to share research ﬁndings; companies like Facebook and Google consistently obscure and muddy the water, they enshroud their data sets in mystical and slippery algorithms, so much so, that it becomes impossible to draw any conclusions about demographics.
Let’s check some demographic biases of Google for instance. Right now. In a minute I’m going to ask you to tap the pause button and search for something in Google, just to see how much Google skews demographics be it on a personal, geographical or possibly religious scale. As a repository for humanity’s collective id, our socio-psychological identity, the auto-complete is plenty revealing.
If you’re not familiar with autocomplete, it’S when you begin typing a phrase, for example, “Who is the…” the ubiquity of Google oﬀers to ﬁnish your thought with the text from other popular searches. Tinker and peek at humanity in wonder. For instance let’s ﬁsh out some stereotypes using the Google’s Autocomplete feature then. Rudder explains a littler bit:
[Dataclysm’s copy editor had a mess of a time pinning down the Google autocompletes for prompts like “Why do women…” Google had given each of us each slightly diﬀerent results (for instance “… wears thongs?” Was my third result, presumably because that’s a typically male question? Hers was… “wear bras?”) It becomes humanity’s [then, Google’s] inner dialogue [that men want to know ‘why women wear thongs’ but women want to know ‘why women wear bras’]: Google autocomplete ranges from the hurtful [through to] the ridiculous [revealing] as it does, part[s] of our [collective] selves. [Our collective prejudices].
I invite you to pause the audio and go to google and trigger the auto complete by typing the follow half statements… ‘Why women…’ and ‘Why do men…’ and see the bursts of judgement that Google spews out. Play along.
Pause and type in Why women… and screenshot the auto completion options which it oﬀers. Mine are: Why women… …cheat?…have periods?…wear high heels?
And also Pause to check the auto complete on Why do men… and again screenshot the auto completion options which pop up. Mine are: Why do men… …pull away? …fall in love? …lie?
Though, then, although OkCupid is deﬁnitely an advertising-revenue driven website, for Rudder his whole involvement with a website from day dot, has been to break through the crystallising stereotypes peddled and which get embedded as societal tropes by the likes of, let’s say, the Google search engine.
From the day he ﬁrst co-launched the dating website in January 2004, he was simply interested in people, not stereotypes, but the individuals. He seems more into the art of anthropology, and his tool is mathematics. The mechanics of this experiment then were to simply remove members’ proﬁle photos, and let the data do the talking.
In my humble opinion. I’m no mathematician, nor data scientist. However I’m well immersed in studying the ethics and implications of data usage, and ready to jump on the heels any lack of ethics by companies and governments involved in data manipulation.
In my opinion, The Love Is Blind Day experiment was one of the most ethically-minded. It was also one of the least intrusive of online data experiments in the history of the Internet. And for the purpose of improving how data can be used as a tool for comparing prejudices and breaking stereotypes. I think it sets a god precedent that data regulation is in fact not the solution, but we need more Rudders in the world, literally a guide; for instance anthropologically-minded, mathematicians who boast loads of empathy for improving the world around them.
Christian Rudder was the perfect guy to master the experiment then. These days, post OkCupid, Harvard educated Rudder describes himself as an American entrepreneur, writer and musician. When you read Dataclysm you can see his love for humanity shine through. He’s some kind of living cyber-folk hero, of the commercial variety, a rarity. But why big him up with such a fancy rosette? Really, it is that OkCupid generated valuable anthropological data. And over more than a decade, as a data generating company, it managed to hoard and analyze responsibly. Responsibly, the key word! Here is a younger college-aged Rudder, a one-line extract from Dataclysm, which sums up how his people-ﬁrst world view got shaped early on:
I was an exchange student in Japan: The name of my exchange program captured the timbre of the visit in three words: Youth for Understanding. A 90s pre-internet. It’s goal: to see the country as much as it would see us. [Rudder goes on to recount story of dismay, about having to take the podium during a school assembly there]… I rose from the ﬂoor to the podium and said something dumb. The next person due up was the only blonde in our little troupe [of students]. And as she stood, and I’ll never forget it, there was an audible gasp. The person standing there was just a regular girl. A shudder went through the crowd as if Pamela Anderson were there in the ﬂesh. People. Many people have taken that shudder at face value. And for decades, phrenologists, racialists and quacks have jumped through hoops to give that essential cultural response, a biological (and therefore immutable) basis.
Rudders does highlight the inherent racism in many thing. So in-keeping with Rudder’s formative educational experiences as an exchange student, the OKCupid site reﬂects his morality. And is programmed to localise similarities we share rather than emphasising diﬀerences. Sound like a bit of an echo chamber, maybe, but that’s another story. The company’s analysis of this data goes beyond the dating metaphor. Today, the datasets collected by OkCupid give an invaluable insight into how racism, sexism and bigotry pervades at a societal level.
And if we can see it displayed in data, this is possibly an opportunity to move society to a better position, to eradicate it across many sectors including; of course in matching people, but also in employment, but more widely in reducing inequality. The data extracted from the experiments expose inherent bias against major parts of the population, ‘us’: Rudder’s calls this the ‘whiteness factor’. What is your whiteness factor, then? What are the indicators? The shape of your cheekbones, the tone of your voice, the density of your hair follicles, your postcode, the school you attended, your grade sheet, the people who you connect with on social networking sites?
How do we get around this data driven ‘whiteness’ bias? To expose the ‘whiteness’ biases Rudder posits three examples: it’s as if there’s a lottery, a cultural, employment and geographical lottery. But is your gamecard already voided before game on?
Okay, let’s emphasis this ‘whiteness factor’ in three details. We have a culture detail, and employment detail and a geographical detail. Ok, the ﬁrst Culture:
Culturated whiteness: in the non-sighted, registered blind population: ——- Sociologist Professor Osage K. Obasogie recently produced some ingenious research – Blind from birth. Obasogie interviewed people blind from birth and found the same attitudes about race as in the sighted world. Obasogie asserts that blind people’s attitudes on race reﬂect a lifetime of cultural absorption, as opposed to any visual reality.
Institutionalised employability whiteness: ———- Are Emily and Greg More Employable Than Lakisha and Jamal? Which got a resounding Yes from the nation’s HR professionals.
Expatriate geographical whiteness: ———- It is not my main claim that white men are unusually good-looking. Nor am I claiming that the data ‘proves’ black people aren’t attractive. In fact, OkCupid’s patterns change in places outside the United States. [For example] in the UK, the site’s black members get 98.9% of the messages white members do. In Japan, 97.8 percent. In Canada, 90%. (Many are Amercian’s abroad).
It seems ‘whiteness’ is actually a good old-fashioned American problem, which an unresponsible manipulation of data simply exacerbates and then slips over into being reﬂected in our subconscious attitudes and actions. I urge you to challenge any of the ‘whiteness factors’ you see ﬂoating about the Internet and in person to person conversations. ‘Whiteness’ also aﬀects the poor-working class white folk too.
So it’s less about skin colour, it is post-racial. Rather it’s about calling-out the power structures and the status quo but also saying ‘no’ to passing casual ‘isms’: sexism, racism, classism, mysogynism. Let these ideas slips through, and the faster we ingest and repeat them: calcifying and corroding egalitarian values. Let’s keep ‘whiteness’ in check. And stop negative ideas spreading.
That’s your call to action!
So please allow me this ﬁnal digression on appearance, self-appearance: if maybe you upload a better looking person’s picture, passing it oﬀ as your own, believe me it’s more common than your think; a little bit like cat ﬁshing; or you posting one of your much younger self, you’ll probably get more dates. But why are people likely to do it? Maybe they’ve no intention of ever meeting, moving beyond the realm of safety of the black mirror, your screen?
To understand this dilemma we have to then question what motivates people join dating websites in the ﬁrst place? And there’s a two pronged response to this. It’s either that you are the type of person who seeks quick validation: enjoying a kind of foreplay of messaging, even a causal hookup. Or conversely you’re the type seeking a long-term commitment and so keep dating until you ﬁnd ‘the one’ in a million.
There’s interesting highlights on Rudder’s data related to gay, lesbian and bisexual compatibility matching criteria too. And I’ve included this on the episode page over at upyourenglish.net. That’s it, so that’s it, that’s the end of this podcast episode. Thank you very much for listening again. I hope it interested you in discovering more about what good data science can achieve. Also do leave a comment on YouTube if you listen there: is the ‘whiteness factor’ something you’ve encountered? If you’ve any comments. Otherwise, don’t hesitate to write to me and tell me and let me in on your thoughts.
As usual I would like to remind you that if you like this podcast and want to help me, you can leave an evaluation on iTunes (if you listen to podcasts on iTunes) or on the other platforms like Spotify. That way, I can reach more and more listeners and help more and more people up their English.
Of course there will be a new podcast episode soon, and during that I’ll be proﬁling King Krule whose album is due late Feb 2020, and who I’m going to be catching play at Paris Olympia. And a ﬁnal word from the 26 year King Krule:
‘When I realised I could write lyrics and let someone that I knew listen to them, like I could write this song about this girl, I could play it to them. I just loved it, because all of the words would speak to them. I could see them slowly falling in love with me.’
Thanks lovers, I’m out of here! Gotta check that black mirror. Biiiiiig up, and biiiiiiiiiig lerrrvvvve