If there is a positive significant trend in the score there is improvement. For simplicity let the trend be the least squares linear trend. If positive then test for the significance of the coefficient of regression by t-test. If significant then there is improvement.

Dear Laura

1. If the number of observation is less than 30, non-parametric test has to be applied.

2. If the number of observation is greater than or equal 30, then T-Test has to be applied.

Regards,

Zuhair

Dear Laura;

As explained by our two colleagues, t-test (student test) can be applied to compare two situations with many observation values

If the t-test value is under 5% we can say that the improvement is significative and it is not due to chance

---------------------

Dr. MATAOUI

Thank you, Ette Etuk, Zuhair and Dr. Mataoui!

I currently have less than 30 data points, so for now I'll use a non-parametric test, as Zuhair suggested. I expect I'll increase to approximately 40 over the next few weeks, so I'll use a t-test after that.

Thanks again, all!