Facebook is home to billions of conversations a day. Whether those are status updates, messages, or comments, the social network is a buzzing hive full of our constant chatter. Which makes it a very powerful point of analysis: Why not use Facebook to try and learn more about what our communication styles and language choices mean about us?
A group of researchers have done exactly that. “We analyzed 700 million words, twiphrases, and topic instances collected from the Facebook messages [and status updates] of 75,000 volunteers, who also took standard personality tests, and found striking variations in language with personality, gender, and age,” the paper titled Personality, Gender, and Age in the Language of Social Media: The Open-Vocabulary Approach states.
There have been similar studies about Facebook and language before, but none of this magnitude.
The research team’s psychologist, Dr. Margaret Kern, explains the difference between a Closed Approach and an Open Approach in analyzing this data. “In the closed approach, a group of smart scientists sit down, perhaps with a thesaurus, and create a list of all the words they can think of that seem like they are part of the category. For example, when we first started the project, we wanted to create a list of positive emotion words, so we started with the word “happy,” and found hundreds of synonyms,” she tells me.
“There are a few problems with that. First, it’s our thinking on the category, which is impacted by our our background and biases. Second, the same words can have very different meaning, depending on the context. With the open approach, the computer automatically finds words that often are used together, or words that correlate with a trait or characteristics that we are interested in.”
So what did the study reveal?
Battle of the sexes
The study’s lead research scientist Andrew Schwartz say the degree to which they were able to predict gender was an incredibly high 92 percent – surprising even him. “This is much higher than I expected. When people write enough on Facebook, algorithms can determine their gender simply based on their words.”
According to the research, women use more emotion words, like “excited,” as well as more first-person singulars. We’re also more likely to “mention more psychological and social process; [for example] ‘love you’ and ‘<3.'”
Horrifyingly, the word “we” starts showing up a lot more after age 22, and “I” decreases.
The data found men use more crude language and “object references,” like talking about their Xboxes or sick new Jordans. But they aren’t devoid of text-based affection: Apparently, men are more likely to tack on the word “my” when talking about a partner – so they would say “my wife,” or “my girlfriend” when writing about their significant other (the study only concedes this was true of references to “opposite-sex partners,” for the record).
Women, on the other hand, would just say “husband” etc, sans “my.” However, there’s a big caveat to this, because the fairer sex will add descriptors like “amazing” or “her,” meaning we will talk about someone else’s husband and how amazing he is. Men don’t tend to fawn over anybody else’s partner … or their own, for that matter. But they do attribute the “my,” which suggests possessiveness – interpret that as you will.
Age is just a number
Some of the most significant differences in speech surface when you start to look at age. Here are some primary examples, according to the study:
- Between the ages of 13-18, common phrases include “school,” “homework,” and “ugh.”
- Between the ages of 19-22, common phrases include “semester,” “college,” and “register,” as well as “drunk,” “hangover,” and “wasted.”
Don’t worry, though, fellow mid-20 somethings: We talk about drinking, too. Except for us, it’s a little mellower. Words like “beer,” “drinking,” and “ale” surface for our age group.
The researchers say that mostly what’s observable across the ages is a natural progression. When we’re younger, we post about school, then college, then work, then family – a predictable pattern. As we get older, certain words start coming up more that indicate relationships are increasingly important to us – ie., “son,” “daughter,” “father,” “mother.”
Also, horrifyingly, the word “we” starts showing up a lot more after age 22, and “I” decreases. I think I can speak for my generation when I say this is one of the easiest and most annoying conventions of speech to slip into.
But it’s not necessarily the obnoxious couples out there updating you about their fireside weekend chill sessions or pumpkin patch excursions (plus 18 photos of their two year old inside a pumpkin – it doesn’t look just like those Anne Geddes photos, psychos). “We take this as a proxy for social integration, suggesting the increasing important of friendships and relationships as people age.” So the overuse of “we” means that post-22 year olds are spending more time investing in all relationships, not just with our (shudder) lovers.
A winning personality
There aren’t many huge surprises when it comes to language use and personality type. People who, according to the personality test administered for the purposes of the study, are extroverts use words and phrases like “party,” “love you,” “boys,” and “ladies.”
Introverts? “Computer,” “Internet,” and “reading” were more common.
Those who are more open to talking about music, art, writing, dreams, and the universe use the word “soul.”
While none of this is hugely revealing, the group was able to make a few interesting conclusions about personality type and language.
“Emotionally stable individuals wrote about enjoyable social activities that may foster great emotional stability, such as ‘sports,’ ‘vacations,’ ‘beach,’ ‘church,’ ‘team,’ and a family time topic.” (I find it very strange that “Lakers” and “Kobe” both indicate emotional stability, I must note. I’ve been known to write these words on Facebook but they are often prefaced or followed by obscenities.)
Schwartz says language’s ability to predict emotional stability was a surprise, specifically that “… An assortment of sports and other social activities can lead to or is a result of being emotionally stable. Though in hindsight it makes sense, this was a result that hadn’t been studied in psychology before. In other words, it shows how this type of research can guide the future of psychological and social science work.”
And to make things for Forever Alones even more cliche: “Results suggest that introverts are interested in Japanese media.” Example words are “manga,” “anime,” and Japanese-style emoticons. Damn you, predictability!
Those who aren’t particularly open are largely responsible for the horrible, numerical shorthand that so often peppers text messages – phrases like “2day” and “ur.”
Of course, everyone’s an individual – because according to this study and my Facebook posts, I’m a 22 year old male who might be a sociopath.
Clearly, though, there is much to be gleaned from the study. Another surprise to Schwartz was how much sense it all makes.
“It really confirms that we can learn a lot with social media when we start looking at aspects of people that are less well-studied,” he says. “Being one of the first studies to look at Facebook language, it was important to look at psychological factors that are well understood so that the field can start to gain confidence in using the medium as data.”
- How to use Samsung Bixby: Everything you need to know
- What is artificial intelligence? Here’s everything you need to know
- How to use Google Assistant, plus all the ‘OK, Google’ commands you need to know
- You will never need to charge a smartwatch again, if new EU project succeeds
- Hey Google, why did you kill off Allo, your best messaging app in years?