okay, so the formula for the t-test is:
(mean_1-mean_2)/sqrt[ var_1/(n_1-1) + var_2/(n_2-1)]
So look at the number of, say, women, in the two groups you're trying to compare (this is where I didn't understand the instructions.... are you comparing >400 schools with <400 schools?). Anyway, I would first sort the dataset (Tools -> Sort) by the criteria you want.... so (assuming I'm right about what you're supposed to do) sort by # of students. Take all of the schools with >400 students, and move them slightly down the page so that you can have a good separation between the two groups (or copy-and-paste to another page in the spreadsheet if you feel like it and are comfortable working across pages).
Then find the mean proportion of women in both groups of schools. Say it's .25 in <400 schools and .35 in >400 schools. Then those are your mean_1 and mean_2. Then find the variance of both groups.... using the formula for variance, or maybe Excel can do it for you. Anyway, that's your var_1 and var_2. Then just count the # of schools in the <400 group and the >400 group, and those are your n_1 and n_2. Then just use the formula. You can do this in Excel or with a calculator, whatever.
Facebook
Twitter