Be Careful, It's a Jungle Out There

As the holidays draw to an end, we wrap up this series of light hearted blogs with a few closing thoughts on post-truth and holiday romance, as seen through the lens of data analytics.

The curtain came down on 2016 with the vaguely disquieting word, post-truth, crowned as the word of the year. The Oxford Dictionaries defines post-truth as 'circumstances in which objective facts are less influential in shaping public opinion than appeals to emotion and personal belief'. In the wake of last year's surprising election results in the UK and the US, the Financial Times rightly describes 2016 as 'a bad year for numbers'. As data scientists who work with numbers everyday, we had no choice but to sit up and pay attention.

Is post-truth (the German word is the wonderfully resonant postfaktisch) a phenomenon that only afflicts us at election time or is it a more pervasive, near constant feature of modern lives? We looked at romance in the digital age and asked ourselves to what extent objective facts fall by the wayside in the relentless pursuit of love.

Magic in the air 

There is something magical about walking hand-in-hand along a sandy beach and watching the sun set with someone delightful you’ve just met. But where to find that elusive soul mate who not only ticks all your proverbial boxes, but ticks them with a flourish? These days, more and more people meet their partners online and in the US, one-third of marriages are said to start off with a period of online dating.

There is reportedly a huge spike in online dating between Boxing Day (26th December) and New Year’s Day. Some online websites report increases of up to 350% in online traffic during this time. As data scientists, we take some comfort from knowing that many online dating sites use some degree of data analytics to increase the likelihood of a ‘good’ match. Some online dating sites ask as many as 400 questions, the sheer number of which is probably off-putting in itself. Completing 400 profile questions might be indicative of desirable qualities such as patience and resilience. Or less desirable ones like desperation. It can be hard sometimes to tell the difference. A big problem remains however, and that is how to assess whether the data your prospective partner provides is true.

The truth about lying

Because individuals naturally try to show themselves in the best light, some exaggeration should be expected. Extensive research in sociobiology and evolutionary psychology suggests men look for youth and physical attractiveness, while women look for a partner’s ability to provide. When posting their profile on online dating sites, men and women, subconsciously or otherwise, factor in these considerations.

Jeffery T Hancock and Catalina Toma from Cornell University and Nicole Edison from Michigan State University have done some fascinating research on “The Truth about Lying in Online Data Profiles.” The researchers compared profile information on height, weight and age with actual information. Their research revealed that two-thirds of participants’ weight was inaccurate by 5 pounds or more (roughly 2.2kg) and almost half lied about their height by more than half an inch (roughly one cm). Interestingly enough, age was the least lied about statistic. 

We suffer from what sociologists term social desirability bias. In other words, we tend to give the answer we think the person on the other side wants to hear. No one wants to be unliked. Researchers have found that asking people what they believe other people think of them often yields more accurate responses than asking them the same question more directly. In other words when asked how fit he is on an online dating survey, a man may claim to push Usain Bolt hard in the 100m. But ask him how fit other people think he is, and he may sigh and confess to not being able to walk up a flight of stairs without gasping for breath. 

Now one would think that all would be revealed when the nervous couple eventually meet but some relationships are never intended to get to that stage. There has been an unfortunate spate of men forging online relationships with women and then defrauding them of money through careful emotional manipulation - without ever having met in person. A clear example of appeals to emotion and personal beliefs being more important than objective facts. 

For those who may be feeling twitchy about online dating, there are sites such as truedater.com, a rating platform of sorts, which enable prospective online daters to vet profiles of individuals before deciding whether to meet them or not.

Do you believe in surveys?

Why is all this important beyond the risk of sitting across the table from an incontinent troll you thought from their profile was a Tom Cruise/David Beckham/Zac Efron look alike? (Please select the most appropriate response according to your generational preference). As the Brexit and Trump election victories have shown, sometimes the stakes can be very high indeed. Last year, exit polls and surveys had a torrid time in successfully predicting outcomes. After the Brexit vote, Marine Le Pen, the far right French leader in response to having a barrage of unflattering election survey statistics flung at her by a BBC interviewer, scoffed and asked “… you still believe in surveys?”

For many people looking for love however, the answer to Ms Le Pen’s question is a resounding ‘Oui’. As data scientists, we believe that skillful probing of data sets generally leads to better decisions. So short of enlisting a tame data scientist to hunt for the kernels of truth in your prospective partner’s online profile, all we can say is ‘Be careful, it’s a post-truth jungle out there’.