Statistical analysis: Comparison and false discovery rate.

Hello,

This might be very trivial so please bear with me.

A = [a1 a2 a3 a4] B = [b1 b2 b3 b4]

Ho = [A-B]^2 = 0 i.e. A and B are the same.

AND I also want to correct for False Discovery Rate (multiple comparisons) & report values (eg. [a3-b3]^2) that are significantly larger than 0.

Please advise,

S

Answer 1

[the cyclist]

You need to provide more description of what you want. It's not very clear, at least to me.

Do a1...a4 represent samples from a distribution, and b1...b4 represent samples as well? Are you trying to test if A and B are samples from the same distribution? If so, then you may be able to use ttest() or anova1(). The multcompare() function can be used for multiple comparison testing. (These functions are all in the Statistics Toobox.)

If those are not samples, it is not clear to me what you mean. If you just have a single value for A (but the elements are not a sample), and likewise for B, then I don't know how you can a statistical-style analysis.

Answer 2

[sas0701]

Hi, a1.. a4 are independent variables at 1 time point. Not from the same distribution. eg. student math scores during midterms. b1..b4 represents the same students math scores during finals.

What I am trying to ascertain is whether any student's score has changed between midterms and finals (null hypothesis being that they have not). I don't want to know if there is a general change in the distribution but more of -'which student changed'.

Hope this helps..