t test for two samples with different sizes

Hi,

Which function should i use if I want to test whether two samples with different sizes have different means? I see many ttest functions in matlab

Answer 1

[the cyclist]

 >> ttest2(x,y)

where "x" and "y" are the vectors of your two samples.

[Shafa-At Sheikh]

As per notes, x and y comes from independent random samples from normal distributions with equal means and equal but unknown variances. What about the case, once the variances are unequal?

Is this function also checking the variances before calculating the significance?

[the cyclist]

The documentation does state that the simplest syntax assumes equal variances. But if you read a bit further in the description, you will see that there is a calling syntax for unequal variances.

Specifically, see the VarType name-value pair

[Jules]

you responded nearly 9 years later. dedication.

[Jules]

if you see this, i am trying t do a t-test with two sets of unequal size. i do not care about the size difference, is there anyway to bypass this - other than doing it manually ?

[the cyclist]

Yes, still dedicated.

I'm unsure of what you mean, specifically I am not sure what you are trying to "bypass".

When you say "size", I assume you referring to the number of observations in each sample. ttest2 doesn't require equal-sized samples. (If you input two samples into ttest, it will do a pairwise test, and will require equal-sized samples.)

What specifically are you testing for, between the two samples? Whether or not they have different means? Or something else?

A careful reading of the documentation should answer your question.