A study performed by NYU researchers found that cities with more racist tweets reported more real-life crimes related to race, ethnicity, and national origin. The researchers analyzed 532 million tweets posted in 100 cities between 2011 and 2016 by using a machine learning model to categorize tweets into two categories: targeted discrimination and self-narratives that describe an experience or comment on one.
The team then compared the frequency of those tweets to the number of hate crimes reported in the corresponding city during that time period. The study found a negative relationship between self-narrative tweets and crimes related to those biases. Rumi Chunara, one of the study’s leaders, said, “This trend across different types of cities…confirms the need to more specifically study how different types of discriminatory speech online may contribute to consequences in the physical world.”