I ve had and am having the hardest time trying to figure out which hypothesis test to use for the following:
A 2 tailed test
Known: n = # of trials
mean for sample 1
mean for sample 2
n's are <30 AND population variances are NOT known. Since variances are not known, I should use a t-test, but then in order to get Sp^2 I need the variances. Am I trying to use the right test?
Do sample 1 and sample 2 follow any law? n: # of trials seems like it could be a binomial law? In this case you can find the std deviation with the definition of the variation in a binomiale law (mean = E(x) = np; then find var = np(1-p) => std dev = sqrt(var))
If you give the whole question it might be easy to figure out
well, here's the original problem:
An experiment is conducted to determine weather the average dose of vaccine exceeds that of another by .2 mg.
x1(bar) = #
x2(bar) = #
s1 = #
alpha/2 = 0.025
For this I used the z-test. Then question asks to test under same conditions, except that this time, n's are small and population variances are unknown. So how would I approach this?
I am closing this thread, I have just about given up on trying to find a solution to this problem.
what are the hypotheses?
It seems that you can either do the s pooled test if the pop variances are equal or the uequal case via Satterthwaite,
see Student's t-test - Wikipedia, the free encyclopedia
Equal sample sizes, equal variance
This test is only used when both:
the two sample sizes (that is, the n or number of participants of each group) are equal;
it can be assumed that the two distributions have the same variance.
Unequal sample sizes, equal variance