Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
T test
Message posted by MARGO QUINN (via 194.125.13.3) on February 26, 2002 at 5:18 AM (ET)
How do I report on t-test where the variance is not assummed equal?
READERS RESPOND:
(In chronological order. Most recent at the bottom.)
Re: T test
Message posted by JG (via 24.42.170.149) on February 26, 2002 at 6:41 AM (ET)
You have to combine the variances to do a proper t-test. Be sure that you combine df appropriately.
Re: T test
Message posted by Jack Tomsky (via 12.144.103.66) on February 26, 2002 at 6:07 PM (ET)
If the data are paired, then you can take pairwise differences and treat it as a single sample. If the data are not paired, then you can use Satterwaite degrees of freedom as an approximation to the t distribution, If the sample sizes are large, then you can use the normal distribution as an approximation.
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.