The researchers developed a test to check for different levels of error in high-dimensional data. They adapted a known test for this purpose and used a special algorithm to analyze the results. The test was applied to real-world data on economic growth and supermarket sales, showing that the economic growth data had consistent error levels while the supermarket data had varying error levels.