# Help: Stats Project

#### StatsKills

##### New Member
The project is to run regression analysis on 2 stocks compared to the S&P 500 and then do a write up about it. I did the regression analysis on excel but do not understand the professors instructions for what is needed to get full credit. Here is what he said. Can anyone tell me what I am supposed to do? Everytime he tries to "clarify" I get more confused. Oh, this is an online course too.

Three major analysis I will look for in your project is as follows:

(1) Test if Ha: beta >1 (or beta<1). In Chapter 14, most of cases, we were testing Ha: beta>0, beta<0, or beta not equal to 0. And EXCEL stats functions can only outputs on the later cases (in another words, you need to modify a little bit to what you get from the EXCEL function). Therefore, make sure you are in the right track of testing beta > 1 (or beta<1).

(2) Test if Ha: beta for stock 1 > beta for stock 2 (or vise versa). In Chapter 10, we have a test for mu1 > mu2 (or vise versa) by using the T-test (xbar-ybar)/sqrt(var(xbar)+var(ybar)). You need to come up a similar T-test for testing betas by the T-test (b1 from stock1 - b1 from stock2)/sqrt(var(b1 for stock 1)+var(b1 for stock 2)). Think a little bit deeper and you can get all of those figures from your EXCEL printout.

(3) Residual analysis for (a) y-yhat follows a bell shaped distribution (draw its histogram as learned in Chapter 3, and the textbook's discussion was sort of wrong) (b) mean is 0 (from Chapter 9), (d) variance = constant (we did not learn it in this semester, but we can look into the residual plot as discussed in the textbook) (e) independent ( we did not learn it in this semester either, but we can look into the residual plot as discussed in the textbook also.)

E

#### elnaz

##### Guest
Hello

Every tests for performing must have existed its special conditions.
for example Before that you start to do regression
you must survey that its conditions exist or not.
the conditions for doing regression is normality and independence and hemogeny of variance for residuals