OK, Maybe it has been a long weekend or I am just not making the connection here… Starting on page 190 Schweser book 1 discuses dummy variables. I am able to follow the logic and make sense of the reading until the top of page 192. The paragraph begins "We can also test…EPS in each of the first three quarters is equal to the fourth quarter by testing the individual slope coefficients… Can someone please explain how the Null Hypothesis of b1=0 tests whether fourth quarter EPS = first quarter EPS. This has to be something simple that I am just not getting… Thanks
You might have found the solution by now, but if you think of it each slope coefficient b1,b2 and b3 is the difference between each of the other quarters with the fourth quarter. A simple way of looking at this would be write out the model for finding the EPS. For the first quarter it is given by EPS = b0 + b1 The rest are zero based on the way we construct the dummy variables. the EPS for 4th quarter is EPS = b0 So when will fourth quarter = first quarter? when b1 = 0 HTH.
Coefficients on dummy variables capture shifts in conditional means relative to the “base” case. In this case, the base case is the 4th quarter (the “undummied” one), which is picked up by the intercept. It’s a shift in CONDITIONAL mean because you’re controlling for the other variables’ effects, so the difference is conditional on the values of the existing variabels. Graph the relationship for the 4th quarter and then for the 1st quarter (on the same graph). The two lines are identical in slope, but differ by the amount of the coefficient on 1st qtr eps. The intercept for the 1s quarter is the sum of the “regular” intercept and the value of the coeficient on the dummy. Hope that helps.