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

Stuck on what statistical tests to use..?

Hi all,

I’m currently analysing some data that I have collected in my postgraduate research and I am struggling to nail down which statistical tests I need to use.

I conducted a study on treatment seeking of heavy menstrual bleeding using two groups (menstruating women and practice nurses) and administered an identical survey to both.

The sample size for the menstruating women is 66 and the sample size for the practice nurses is only 33. Could I still use an independent t-test for comparison between the two groups, as the sample sizes differ quite a bit? If not, what alternatives could I use please? I really appreciate any input in this to help me. Also, if you could supply any academic references to support your point or to help me that would be appreciated also.

I feel like I should really know this stuff at this point, but statistics has always been a weak point of mine!

Thanks in advance. :-)

Reply 1

iammichealjackson

Original post by abbi_lloyd

You can use a t-test even if you have unequal sample sizes. The main issue is just that to have the maximum amount of power to detect any differences between groups it would be good to have them as balanced as possible (though having two smaller, but equally sized groups isn't particularly better!).

I'd use Welch's t-test which automatically is calculated by spss under the "equal variances not assumed" row.