PDA

View Full Version : Testing statistical signifigance



ih8stats
08-02-2010, 12:53 PM
I have a problem that I'm totally stumped on. I have a test tonight and I need to know how to do this problem because one like it will be on the test.

I think I have to use p-values. (this all has to be done by using excel as well btw)

I'm really sorry for the formatting. I couldn't figure out how to get it just right.

I've attached a copy of the S&P report and the important information is on page 1,3,4 and 5. I also attached the work that I've done so far.

So here is the problem in a nutshell. Standard’s & Poor did an analysis of “Qualifying Delisted Stocks” from the Wilshire 5000 index. There were 53 qualifying stocks. They were testing the theory that when a stock is delisted its stock price underperforms with regard to the index for the short term and then rebounds and then outperforms with regard to the index. We are not given the original data but are given the results. And this is our task as stated by the professor:

“Your task is to conduct the same analyses as in the article for exhibit 2 and 4 ---see if you obtain the same results...for the assigned tables/figures/exhibits. Also may want to comment on "could the analyses and figures/tables/exhibit be improved”

Here is one of the results from their testing:

Part of exhibit 2

Day after announcement
excess return of deletes: -9.4
std dev of excess return of deletes: 9.4
t-value: -7.30
Statistically significant? Yes


Here is another one of the results from their testing:


Exhibit 4

SIZE QUADRANTS 1st 2nd 3rd 4th

excess return of deletes: -16.8 -15.4 -11.1 -3.9
std dev of excess return of deletes: 10.2 12.3 16.7 14.2
t-value: -5.9 -4.5 -2.4 -1
Statistically significant? Yes Yes Yes No