A study of 75,000 volunteers has found significant demographical differences in how people of different genders, ages, and personality types use language on Facebook. To date, this investigation represents the largest study of language and personality ever conducted. The new “open-vocabulary” research methods offer a fresh and dynamic approach towards understanding how different demographics of people use language differently using Facebook and other social media.
The study of “status” updates offers several advantages over the more traditional research methods of administering surveys and questionnaires. For one, there is an immense volume of content that can be freely accessed. Secondly, obtaining such data does not intrude upon or otherwise influence the subjects. But perhaps most importantly, a subject’s status updates represent unrestricted and spontaneous self-expression.
Whereas previous studies have sought to understand language-use by limiting their subjects to words from fixed lists, researchers from the University of Pennsylvania and Cambridge designed a new method to analyze Facebook status updates. They called this new technique the “open-vocabulary” approach. They examined millions of Facebook status updates and extracted key words, phrases, and topics. The resulting data set was an order of magnitude larger than any other previous study, and provided much richer insights into the kinds of messages posted and the kinds of people who posted them.
After analyzing the data, the researchers reported “striking variations” in the type of language used between different genders, age groups, and personality types.
The results of this study showed that males display significantly higher use of profanity in their Facebook status updates. In addition, they were also more likely than females to refer to objects (e.g. “xbox”). Conversely, females are more likely to use words that refer to emotions (e.g. “excited”) and social processes (e.g. “love you”). Additionally, in this study women tended to use more emoticons, though previous research studies on this have yielded mixed results. Finally, it was noted that males are more likely than females to use the possessive “my” when referring to their significant others (“my wife” or “my girlfriend”).
The language used in Facebook status updates was also examined in relation to age. The results indicated that different words, topics, and phrases peaked in frequency at different points in life. The words “school”, “ugh”, and “homework” were most prevalent with users 13-18 years of age. This group also showed the heaviest use of emoticons and Internet speak (e.g. “idk” and “lol”). Upon transition to the 19 to 22 year-old demographic, words like “semester”, “college”, and “register” became more prevalent. Worryingly, alcohol-related words like “hangover”, “drunk” and “wasted” were also frequently featured in this group. However after transitioning to the 23 to 29-year-old demographic, these alcohol references tended to mellow-out to words like “beer”, “ale”, and “drinking”.
A particularly interesting discovery was how people place increasing importance on relationships as they age. After age 22, people show a nearly linear increase in the word “we”. By contrast, the frequency of the word “I” decreases with age. Other words like “father”, “mother”, “daughter” and “son” continuously increase across the life-span. Though the study had relatively few elderly subjects, this idea that relationships become increasingly important as people age is a subject that researchers are intrigued to follow-up on.
Finally, the study compared how people with different personality types (as assessed by user’s test results from the Big Five questionnaire) correlate with the language of their Facebook statuses. Extroverts showed significantly higher use of words like “party” “boys” “ladies” and “love you”, while introverts were more likely to refer to words like “reading”, “computer” and “Internet”. Introverts also posted more about topics relating to Japanese culture such as “anime”, “manga”, and Japanese style emoticons (e.g. ^_^).
In addition to comparing introversion and extroversion, results also indicated some predictors of emotional stability and neuroticism. Neurotic individuals were more likely to post phrases such as “sick of.” By contrast, people with low levels of neuroticism posted about activities that would seemingly promote emotional stability such “sports”, “church” “team” and “family time”. Though only a correlation, this is yet another intriguing insight that could lead researchers to discoveries about human wellness in relation to behavior.
Open-vocabulary research methods hold much promise for discovering how people from different demographics use language to express themselves on Facebook and other social media platforms. Though still in its infancy, this technique will hopefully create a truly personal and intimate window into the spontaneity of self-expression and the commonalities of the human condition.
By Sarah Takushi