The rate of depression in women tends to be about twice that of men. A graduate student took a sample of 100 cases of depression from area psychologists and found that 61 of them were women. You can model what the data would look like over repeated samplings when the probability of a case being a woman by creating 1,000 samples of 100 cases each when p(woman) 5 .50. This is easily done using SPSS by first creating a file with 1,000 rows.

(This is a nuisance to do, and you can best do it by downloading the file http://www.uvm.

edu/~dhowell/methods8/DataFiles/Ex4-27.sav, which already has a file set up with 1,000 rows, though that is all that is in the file.) Then use the Transform/Compute menu to create numberwomen 5 RV.BINOM(100,.5). For each trial the entry for numberwomen is the number of people in that sample of 100 who were women.

a. Does it seem likely that 61 women (out of 100 clients) would arise if p 5 .50?

b. How would you test the hypothesis that 75% of depressed cases are women? An R program to do these calculations is available on the book’s Web site.