Paired t_test_2 code problem

Hi,

I hope this is something basic that someone can quickly point out my error. I am currently on Octave but the problem transfers to matlab also.

I am trying to perform a paired t-test on two data groups sorted by a variable. To test my code I have created a short test spreadsheet that I know the answer to so I can compare.

Below is the code I have used to separate into two groups, it worked fine when I had equal groups in my array but now they are uneven (10 and 9).

Can anyone tell me how I can arrange by the vairable in column 2 and assign to x and y accordingly?

clear
cd "/home/james/mystuff"

 filename= 'testttest.csv'
 num =  csvread(filename);

 Xpos=num(:,1);

 G=reshape(num, 11,4);

 x=G(:,1);
 y=G(:,2);

 [pval,t,df] = t_test_2(x,y)

The layout in the spreadsheet is two columns, 1st is the dependent variable and the 2nd is the independent.

[Thorsten]

The function in the statistics toolbox is called ttest. Try

 [h p ci stats] = ttest(x, y)

[Jimmy]

I think thats a difference between matlab and octave. [pval,t,df] = t_test_2(x,y) is the equivalent unless someone can tell me otherwise. I get an output for pval, t and df, they just arent right.

[Tom Lane]

You mention a paired t-test. In my experience this is a test on the difference between the two samples. I don't understand how you could have a paired t-test when the two samples have different sizes.