The researchers developed a new method called Modified Information Criteria (MIC) to improve the selection of lag length in unit root tests. They found that the MIC performs better in choosing the right lag length compared to traditional criteria like AIC and BIC. Using Monte-Carlo experiments, they showed that the MIC significantly enhances the accuracy of unit root tests. Additionally, they extended existing tests to include detrending of the data, resulting in tests with better size and power properties.