Hello all,
I am not a statistician but I was wondering if someone could help me
figure out how many participants I need in order to carry out a number of
t-tests (I already have the data, so I just want to make sure that it
makes sense for me to run the tests). Here's a brief summary;
I have a number of unequal sized groups who have answered the same
question on a scale from 1 to 5.
They are;
T-test 1: Asian women - GROUP A: Intend to do behaviour X (N=120), GROUP
B: don't intend to do it (N=8)
T-test 2: White women - GROUP A: Intend to do behaviour X (N=91), GROUP B:
don't intend to do it (N=31)
T-test 3: Black women - GROUP A: Intend to do behaviour X (N=40), GROUP B:
don't intend to do it (N=2)
T-test 4: Teenagers - GROUP A: Intend to do behaviour X (N=48), GROUP B:
don't intend to do it (N=23)
T-test 5: Aged 20 and over: GROUP A: Intend to do behaviour X (N=205),
GROUP B: don't intend to do it (N=18)
T-test 6: Low Deprivation: GROUP A: Intend to do behaviour X (N=124),
GROUP B: don't intend to do it (N=16)
T-test 7: High Deprivation: GROUP A: Intend to do behaviour X (N=132),
GROUP B: don't intend to do it (N=25)
So I want to compare the mean score of Group A with Group B in each of the
t-tests.
I guess from looking at the numbers that T-test 3 is going to be fairly
meaningless, but I'd like to know if there is a way that I can work this
out statistically.
Thank you in advance for your help,
Brian
---------------------------------------------------
Dr. Brian McMillan
Research Fellow
School of Psychology
University of Leeds
Leeds
LS2 9JT
phone: 0113 3435714
fax: 0113 2431751
www.psyc.leeds.ac.uk
|