Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
independentand dependent variables and t-tests
Message posted by Bernadette (via 64.53.45.49) on April 28, 2001 at 8:15 PM (ET)
I am doing a research paper comparing the effects of hormones on menopausal symptoms. Is my dependent variable the actual drug and independent variables the various symptoms? I am comparing with a placebo group as well. I think I need to do a t-test, paired or unpaired?
I'm so confused at this point, I obviously need some help, please.
READERS RESPOND:
(In chronological order. Most recent at the bottom.)
Re: independentand dependent variables and t-tests
Message posted by Darius (via 200.23.217.10) on April 30, 2001 at 5:15 PM (ET)
Paired test is only if the sample 1 of group1 is related to sample 1 of group2 and you try to determine if there is a difference or if one of the groups is bigger (for example the weight before and after a diet program, sample 1 in group1 is weight before and sample 1 in group2 is the weight after the program of the same guy). If there is not that kind of relationship then you should use non-paired "t" test.
Re: independentand dependent variables and t-tests
Message posted by Phil (via 216.175.114.173) on May 6, 2001 at 11:36 AM (ET)
It sounds like your placebo group are different subjects than the other groups. If so, the t-test is for independent (unpaired) groups. This is a MAJOR factor in comparing the results and a mistake can very easily give you a wrong answer.I would think that the drug is your "independent" variable and the "symptoms" that result are are the "dependent" variables. How are these symptoms measured? If it is counted data in several categories, then maybe the Chi-square test for independence/homogeniety is more appropriate.
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.
