7) (12 points) A key assumption for interpreting the slope S, in the regression function y = Bo + Bix + u as a causal effect is that E[u | x]=0 a. (4 points) Suppose y is first year university grade point average and x is usual hours of sleep per night. Explain the assumption E[u | x]=0 in this context. b. (4 points) Imagine that you have reason to believe that E[u | x]=2. Does this mean that the OLS slope is now biased? c. (4 points) Imagine now that you would like to test whether E[u | x] = 0, and you decide to do the following: 1) estimate the OLS regression of y on x, and compute the residuals; 2) estimate the OLS regression of res on x and check to see if the slope and intercept are zero. Comment on whether this is a good or bad strategy