A new tool for computer language analysis can evaluate your mind based on your Tweets (and might help psychologists, too)
Unless you've been living under a rock or among the molemen, you've probably enjoyed the humor of @s--tmydadsays, the popular Twitter account of Justin, who describes himself like so: "I'm 29. I live with my 73-year-old dad. He is awesome. I just write down s--t that he says." That s--t consists of cranky honesty like "I need to change clothes? Wow. That's big talk coming from someone who looks like they robbed a Mervyn's" and "Oh please, you practically invented lazy. People should have to call you and ask for the rights to lazy before they use it."
Most agree that s--tmydadsays is funny, but did you realize his emotional style is angry, his social style is personable, and his thinking style is analytic, sensory, and in-the-moment? These psychological insights can be gleaned by plugging s--tmydadsays into Analyze Words, a new Twitter-analyzing tool put together by James W. Pennebaker, his colleagues Roger Booth and Chris Wilson, and his daughter Teal. Pennebaker-a University at Texas Professor of Psychology-is a longtime innovator in using computer analysis of language to study how we think.
I asked Pennebaker by email for insight into the s--tmydadsays results, and though he said the sample size was a bit small, "...the analyses catch the emotional tone perfectly. Some serious hostility, depression, and anxiety is in the air. Socially, the writing suggests someone immersed in his social world, with constant references to other people-wife, mother, father, son. In other words, very different from someone who writes about computer components. Low in arrogance because he does not use big words and complex sentences and a high rate of articles-all of which are markers of psychological distance. The valley girl language probably reflects his high use of present tense verbs and punctuation."
Yes, s--tmydadsays scored high in the social style category "Spacy/Valley Girl," which is kind of a brain-bender. If you're equally surprised that this category is included at all, it's because it can measured-not every emotional, social, and thinking style has reliable linguistic symptoms. As Pennebaker said in a phone interview, "I know what I can measure and what I can't." It would be wonderful to measure something like "guilt-riddenness," for example, but that tendency can't be quantified yet.
Pennebaker has worked for decades on figuring out just how words and mental states are associated, in an effort to "come up with a way to measure healthy writing." Starting in the early nineties, he first collaborated with grad student Martha Francis and later with New Zealand immunologist Roger Booth to create LIWC-Linguistic Inquiry and Word Count, pronounced "Luke"-which provides the methodological basis for the Analyze Words site. Using language as a window into the mind is as old as listening for Freudian slips, but Pennebaker's work is groundbreaking in how it links, as he puts it, "low-level words with broad psychological processes." It turns out that style words (such as articles and prepositions) actually reveal more about what's on our minds, psychologically and socially, than content words (like dog, airplane, etc).
Many of Pennebaker's discoveries are counterintuitive, to say the least-particularly with regard to that pesky pronoun "I." To many, "I" feels like a word of the powerful and arrogant, but it isn't really: It turns out that women, followers, young people, poor people, depressed people, crappy students, and sick people all use "I" more than men, leaders, older people, rich people, happy people, good students, and healthy people. That paints a clear overall picture: "I" is a marker of low status, mainly because people who are lower status are more self-conscious. ("I" is also used more often by people telling the truth, as well as the worried more than the angry). In looking extensively at President Obama-who critics have incorrectly accused of being in love with the word "I"-Pennebaker found just the opposite: Obama is an infrequent I-user, reflecting self-confidence, coolness, and psychological distance.
President or peon, our words give away emotions and thoughts we might prefer to conceal. As Pennebaker wrote in "The Psychological Meaning of Words: LIWC and Computerized Text Analysis Methods" (co-authored with Yla R. Tausczik), "The words we use in daily life reflect what we are paying attention to, what we are thinking about, what we are trying to avoid, how we are feeling, and how we are organizing and analyzing our worlds." Digital tools like LIWC allow those symptoms to be collected and quantified with tremendous ease. As Pennebaker puts it, "In the amount of time it takes to run a single participant in a social psychology language study, we can now download thousands of personal writings, interaction transcripts, or other forms of text that can be analyzed in seconds."
That said, Pennebaker emphasizes that while style words are "reflections of what is going on in people's heads," but they're not a tool for getting someone to change their way of thinking. In other words, you can't ask someone to mindlessly repeat more "positive" words and expect them to become less depressed or suicidal. LIWC's real use is in detecting problems such as excessive worry or anger and then showing when progress has been made. When we become more mentally healthy, our language changes unconsciously, because we are changing perspectives. The internal world manifests in the lexical world.
Let's just hope Pennebaker detects minimal "progress" in s--tmydadsays. When it comes to humor, anger and worry are pure gold.