The study looked at different ways to measure the relationship between two sets of data. They found that traditional methods work well when the data is clean, but a method called median absolute deviation correlation is better when the data has outliers.