The researchers used a machine learning algorithm to predict whether anglers would keep or release fish they caught. They analyzed a large dataset from recreational fisheries and found that fish length was the most important factor in predicting whether a fish would be kept. Other key factors included the number of fish caught, where the angler lived, how far they traveled, and how long they fished. The algorithm was able to accurately predict the outcome for most fish caught, showing that these factors play a crucial role in anglers' decisions. Understanding these factors is important for conserving and managing fisheries effectively.