The tutor shows an example of the matched pairs t-test.

In my post from Dec 11 I began about the matched pairs t-test. It’s used to determine if two samples come from the same population.

In my previous post I mentioned generating two lists of random numbers. The second one, I pointed out, has all its entries increased by 5.

Here are the two lists:

Testing against a supposition that the samples come from the same distribution, the mean difference should be 0. The t-statistic is the observed mean difference, minus 0, all divided by the standard deviation of the observed differences:

t = (diff – 0)/(Σ(diff)²/(n-1))^1/2

It has n-1 degrees of freedom: 14, in this case.

Crunching the numbers (this time, with the Casio fx-260Solar) gives t=-0.174, which is virtually the middle of the distribution. In this context, the matched pairs t-test cannot distinguish that one list differs from the other.

HTH:)

Source:

Harnett, Donald L. and James L. Murphy. Statistical Analysis for Business and Economics. Don Mills: Addison-Wesley, 1993.

Jack of Oracle Tutoring by Jack and Diane, Campbell River, BC.