Year 12s: what will be the first way you research your future options? >>

Independent samples T-test

Trying to analyse my results from an experiment I carried out but it's been a nightmare!

Participants were split into two independent groups, but sample sizes were different so I have unequal variances. Therefore, the formula I need is this (Welch's T-test):

Stuck on the next bit now. I thought the only formula there was for working out the degree of freedom was n-1, but apparently with two different samples the formula is (n1+n2)-2. Is this correct?

Then do I just go to the 0.05 value on the T table and find the df I've calculated and minus whatever value is there, from what I got from my test? This bit I'm not really sure about :confused:

Reply 1

ghostwalker

Original post by romaleos2

Stuck on the next bit now. I thought the only formula there was for working out the degree of freedom was n-1, but apparently with two different samples the formula is (n1+n2)-2. Is this correct?

'Fraid I know nothing about this, but as no one else has replied.

From a quick google, the number of degrees of freedom for Welch's T-test is somewhat more complex - see here

You may also find this useful - variations if variances are assumed equal +....