Kim, Khloé, and Kourtney Kardashian: They're just like us!
Oh sure - the reality TV stars inhabit a different universe than most Americans. But their names aren't as unusual as you might think. If you have a brother or sister, you're about 57% more likely to share his or her first initial than can be chalked up to random chance.
That's one of the latest insights revealed by Facebook's data scientists, who published their findings in a new post last week.
Overall, the researchers concluded that about one in nine siblings who use Facebook share their initials, pre-marriage, compared to about one in 14 if their first names were randomly assigned.
(The rate of similar-sounding and -spelled names was especially high for twins, which I'll go into at the bottom of this post.)
Facebook's able to generate these sorts of insights because the 10-year-old social networking site has steadily transformed from Mark Zuckerberg's attempt to put Harvard's dorm facebooks on the Internet to a proxy for America online.
Almost 6 in 10 American adults now use Facebook, according to Pew Research data, and most Facebook users check it daily - sharing status updates, making social connections, and 'liking' books, movies, TV shows and so much more.
That user behavior gives Facebook's analysts arguably the richest data set in America. The company's researchers can comb for broad insights, like whether the world really is linked by six degrees of separation, or get so specific as to figure out the unique counties where Dallas Cowboys fans live.
Most of the user patterns that Facebook has revealed have been relatively harmless. (That trend is likely to continue, after Facebook's data science team was heavily criticized several months ago after publishing several experiments with users' timelines.)
Also See: If Facebook's Secret Experiment Bothered You, It's Time To Quit Facebook
But even if the data science team isn't sharing much of what they know, their posts are consistently fascinating - and expose that what we do online predicts our behavior offline.
Predicting Your Politics
For instance, Facebook 'likes' often predict who's going to win a political election, the company's scientists say.
That may be intuitive, but the insight is still powerful. In the 2010 midterm elections, the Congressional candidate with more Facebook likes won about three-quarters of the time.
(External analysis backs this up; Senate candidates with more engaged users won eight of nine toss-up races in 2012, UMass-Amherst researchers wrote at Politico earlier this year.) But Facebook status updates also can hint at your political affiliation, even when you don't think you're being obvious.
In one example, a meme about Obamacare that went around Facebook five years ago - in the heat of the health reform debate - offered political clues, based on users' word choice.
Some of those clues were pretty blatant, Facebook's scientists found:
Users who wrote 'No one should die because they cannot afford health care' in their status updates were staunchly liberal. Meanwhile, users who countered with modified updates that read 'No one should die because Obamacare rations their healthcare' leaned strongly conservative.
But other clues were well-hidden. Did you know that users who posted a variation of 'No one should be without a beer because they cannot afford one' were likely conservatives, too? Or that users who posted a version that read 'No one should die because of zombies' were more likely to be liberal?
(The graph below shows how the words in those status updates correlated with political leanings; strong blue is strongly liberal, strong red is strongly conservative.)
Predicting Your Love Life
Facebook researchers also have shared copious insights, based on aggregated data of users' relationships, on how and where love blossoms.
One of the scientists' more intriguing findings explains how romantic relationships manifest on Facebook: There's an initial build-up of posts shared between a future couple, as they tag each other in photos, comment on each other's status, and more.
'During the 100 days before the relationship starts, we observe a slow but steady increase in the number of timeline posts shared between the future couple,' data scientist Carlos Duik wrote earlier this year.
Essentially, Facebook can spot digital signs of a real courtship.
But that shared behavior is followed by a rapid drop-off once a couple makes their relationship official - at least, in the eyes of Facebook.
'When the relationship starts ('day 0″), posts begin to decrease. We observe a peak of 1.67 posts per day 12 days before the relationship begins, and a lowest point of 1.53 posts per day 85 days into the relationship,' Duik added.
Duik hypothesizes that the drop-off is due to a new couple spending more time together in the actual world, as opposed to the virtual version.
In another interesting finding, Facebook scientists reviewed data of where its users are likely to be single - and where they're in relationships - to graph the cities that are most and least conducive to finding a partner.
The inverse correlation between 'single rate' and 'relationship formation' rate is not surprising, researcher Mike Develinwrites; 'in a city where everyone is paired up, the incentive to pair up is even stronger,' he points out.
And Develin notes that users in cities on the left and top side of the graph - like Colorado Springs and El Paso - have the highest probability of forming relationships.
Meanwhile, cities on the right and bottom of the graph - like Los Angeles and New York City - aren't places to head if a Facebook user's priority is to find a life partner.
Predicting Your Sibling's Name
In the data scientists' most recent post, they noted that some users' names are tip-offs about their siblings' names, too.
If a woman named 'Hope' has a sister, the odds are really, really high - 31 times more likely than expected - that her name is 'Faith.' (And if it isn't 'Faith,' it's often 'Charity.')
The highest name correlation was among twins. Many twins tend to share initials, and that frequency only goes up for older Americans. Almost half of female twins who are in their 50s share a first initial with their sister.
But even among non-twins, certain pairs of siblings tended to have similarly sounding or spelled names. Facebook's scientists found that a woman named 'Yvette' was almost 40 times more likely to have a sister named 'Yvonne' as expected. Other pairs of siblings also tended to have names with related letters and pronunciation, especially like 'Colleen-Maureen' or 'Landon-Logan.'
On balance, Facebook's scientists were struck by these frequent typographical and phonetic similarities.
'It looks like the entire country, at least on this front, is keeping up with the Kardashians,' they mused.
-
Follow @ddiamond
From the archives:
If Facebook's Secret Study Bothered You, Then It's Time To Quit Facebook Report: Amazon May Be Preparing To Make A Move Into Health Care Health Care For $4: Are You Ready For Walmart To Be Your Doctor? The Outrage Over Facebook's Creepy 'Experiment' Is Out-Of-Bounds - And This Study Proves It
Tidak ada komentar :
Posting Komentar