# Thread: Stats: independent random samples from 2 populatsions

1. ## Stats: independent random samples from 2 populatsions

So, I think it's fair to say that I suck at stats. Hopefully some of you guys can shed some light onto how to do this problem.

Problem: A farmer tests two diets, say 1 and 2, for 12 weeks on samples of young turkeys to see if either offers any weight gain advantage. The results are summarized in the table below

Diet Sample size Mean St. Dev.
1 15 40.2 7.5
2 15 36.4 7.1

Assuming the weight gains by each diet are approximately normal and have a common variance, test the hypothesis that neither diet offers a weight gain advantage. Use alpha = .05.

I'm not really sure what I should be using as the hypothesis. Should I set it up as mu(1)-mu(2) = 0, and have the alternative hypothesis be mu(1)-mu(2) =/= 0?

If so how do I find the rejection region for that?

Also, is the test statistic t= [mean(1)-mean(2)]/[(S-pooled)(sq. rt((1/15)+(1/15)))]?

2. Originally Posted by monsieur fatso
So, I think it's fair to say that I suck at stats. Hopefully some of you guys can shed some light onto how to do this problem.

Problem: A farmer tests two diets, say 1 and 2, for 12 weeks on samples of young turkeys to see if either offers any weight gain advantage. The results are summarized in the table below

Diet Sample size Mean St. Dev.
1 15 40.2 7.5
2 15 36.4 7.1

Assuming the weight gains by each diet are approximately normal and have a common variance, test the hypothesis that neither diet offers a weight gain advantage. Use alpha = .05.

I'm not really sure what I should be using as the hypothesis. Should I set it up as mu(1)-mu(2) = 0, and have the alternative hypothesis be mu(1)-mu(2) =/= 0?

If so how do I find the rejection region for that?

Also, is the test statistic t= [mean(1)-mean(2)]/[(S-pooled)(sq. rt((1/15)+(1/15)))]?
See here

RonL