Linguists and psychologists have been using the content of messages on Twitter to learn more about our society — in areas such as gender, age trends and politics.

Now, a new study by computer scientists at the University of Pennsylvania has found that the behaviors of Twitter users also correlate with their income levels.

The team of researchers, led by Daniel Preotiuc-Pietro of Penn's Positive Psychology Center, analyzed more than 10 million posts of 5,191 publicly available Twitter user profiles. The data was collected in August 2014, and the most recent 3,000 posts from each person were used.

To calculate income levels, the researchers used England's job code system to sort the self-described occupations provided in the profiles and assign a representative, mean income for each code. That data was then used to identify links to a user's wealth and the person's tweeting behavior.

The researchers evaluated how high- and low-income users employ Twitter for different purposes. Those with high incomes used it more for disseminating information and had more followers. Those with lower incomes used it more for social reasons, including sharing links in their tweets.

People earning an average of $69,000 had far more Twitter followers than people making much less.