(8 points) Difference In Differences. Suppose we want to estimate the effect of streetlight intensity on crime (e.g: does installing streetlights in a particular area change the amount of

crime in the area). We have two neighborhoods, A and B. We build streetlights in neighborhood A.

(A) Two ideas for estimating this effect. One: Compare average crime in neighborhood A to average crime in neighborhood B after the streetlights were installed. Two: compare average crime in A after the streetlights were installed to average crime in neighborhoodBto before the streetlights were installed. Why are these comparisons problematic? You need to provide one reason for each comparison. (2 points)

(B) Now suppose we decide to estimate the treatment effect using Difference in Differences. All remaining parts of this question refer to the following graph: