The article tests different models to see if they are better at predicting data than a simple linear model. They look at a specific type of model called a smooth transition autoregressive model. The researchers compare three tests to see which one is best at detecting when the data follows a more complex pattern than a straight line. They find that one test works well in certain cases, while another test is generally more powerful in small samples. This research helps us understand how to choose the right model for predicting data accurately.