Of course, Dr. Lee used time series data! The break point was a study of operations from prior to May 1st and afterwards. He averaged the data into weekly numbers, and he examined trends comparing various metrics before and after that date. Still don't believe me? Look at Dr. Lee's dozen of graphs and what does the hortizontal axis represent? Time usually in weekly averages. That is prima facia evidence that time series data was used!
You would have a valid point if Dr. Lee was only saying "the average of X prior to May 1st compared to the average of X after May 1st", and he did that to some degree, but there were have been huge flaws to the study, insofar the number of data points prior to May 1st would have over the entire year vs. the data points after May 1st covering just a few months. As a consequence, Dr. Lee rightfully, included the vast amount of data points into time series data, and used a dummy variable to represent operations after the May 1st cut-off point, as such, the analysis of a Durbin-Watson test would have been required. (For all we know, the DW test could be a non-issue, but not reporting the number creates doubt.)
The main body of the study was not cross sectional data examining a very short period of time like a day, week or month, but rather well over a year and through several seasons. Don't think it matters still? Dr. Lee did, as he provided a chart demonstrating the change of US relative to other airlines performance and what a dramatic change over year-to-year, and during the time of his study. As we all know, much can happen to an airline's operations over the course of one year.
Yes we know what can happen to operations year over year. Ask Franke Air mgt. What happened to the operations when they failed to have enough de icing fluid in stock at the eastern CLT hub a year ago. Was that a pilot conspiracy also??? Did Franke Air conspire to skew quarterly earnings?
So Corrects Jester.