(3) [This example is loosely based on actual studies] Researchers have gathered daily data on new COVID infections from two states, which we'll call State A and State B. During the time period studied, State A imposed no restrictions related to COVID at any time, while State B imposed a "shutdown" (closing bars and restaurants and limiting the size of gatherings) halfway through this time period, which remained in effect for the rest of the time included in the data. The researchers want to use a differences in differences research design to estimate the effect of shutdowns on new daily COVID infections. The graph below plots daily infections in the two states over the time period (40 days). The red vertical line indicates that State B's shutdown went into effect starting on the 21st day. (Note that in reality we might expect the impact of such as shutdown to take a week or two to show up in the counts of new infections, which are really new diagnoses of previous infections, but here let's assume we actually have the number of people who caught COVID on each date, not the number of people who were previously infected and diagnosed on