The study looked at how the relationship between data points changes over time. By analyzing how much one data point influences the next, researchers found that this connection can be stronger or weaker depending on the data. They used a method called autocorrelation to measure this relationship. When they looked at the autocorrelation of the autocorrelation, they found that the values were often higher than expected if the data points were completely independent. This means that the likelihood of these values happening by chance is very low. The researchers applied these findings to text analysis, showing that better texts have a lower chance of these unexpected values occurring.