The next two questions refer to the following information.  …

The next two questions refer to the following information.   A textile company has installed air conditioners in some of its buildings, but not others. It is considering installing air conditions in the remaining facilities, and wants to see how air conditioning affects output. You therefore conduct a difference-in-differences analysis, using data from facilities that do and do not have air conditions, before and after air conditioning installation. The plot below shows output per facility over time, for each group.   

You are analyzing a clustered experiment. You have data on c…

You are analyzing a clustered experiment. You have data on customers in different stores. Stores were the units of randomization, that is, stores were randomly assigned to treatment or control. If you analyze customer level data, what is the consequence of clustering?

A retail company notices a strong positive correlation betwe…

A retail company notices a strong positive correlation between the number of loyalty program members and overall store revenue. For interpreting this correlation as the effect of loyalty program membership on revenue, which of the following would be a confounding factor?

Consider the following figure from an event study. To estima…

Consider the following figure from an event study. To estimate the event study, a researcher fit a trend line from before and after the event, and estimated the event study effect as the difference in predicted values. The estimate is -0.013. What would be a problem with this estimate?