Twitter is a diverse community of over 310 million active users who tweet over 500 million posts a day. You will expect such a robust community to have the same perception they have offline to this online platform, and it is what this report was looking for.
Scientists from the University of Pennsylvania Positive Psychology Centre, Germany and Australia combed through tweets where they analyzed the various stereotypes formed on Twitter.
The participants were asked to categorize the tweets from the people who wrote them based on their own judgement. The researches then used natural language processing techniques to analyze this across gender, education level, political orientation and age. While generally some assumptions and stereotypes were correct, there were other instances where the participants were quite wrong.
They were able to find interesting results. “For instance people had a decent idea that people who didn’t go to college are more likely to swear than people with PhDs, but they thought PhDs never swear, which is untrue,” lead author, Jordan Carpenter was quoted.
Another finding was how people had a hard time determining someone’s political orientation. If they were unable, people resulted to gender stereotypes where feminine sounding people sounded liberal and masculine sounding people sounded conservative.
The next finding was how people assumed that any tech related language was the sign of a male writer. Although men post more about tech than women, this stereotype led to false conclusions where some posts by women were believed to be written by men.
This is quite revealing since the stereotypes expressed by the populace on Twitter is a huge fallacy that can be a problem to a person who uses Twitter as a data source.
