Hi,

 

Simple question, Can a standard T-test be used for two sets of data which has about 5 values ?

 

I am trying to find diffrences in growth between two bacterial strains at a certain temperature , i was going to do the T-test but i read somewhere that a t-test needs atleast 10 values in your data sets.

Yes you can use the t-test.

The t distribution is defined for a sample size of n = 2 or greater.

 

But remember to check your definition of the 'degrees of freedom' to enter the table

 

I use degrees of freedom = (n-1) and sample size n , but some define it as n and in that case the sample size is (n+1).

Edited August 18, 201510 yr by studiot

The test is valid but you might have difficulty with statistical power depending on the effect size you wish to detect.