I have the following problem I could not solve (I published it before but it seem it did not get through.

The authors is making a randomization distribution with the scaled t distribution of a set of data, here is the image.

The book says that to calculate the difference of means if they where purely random, that is there was no difference between the samples I could swap each values and that I have C(11,5) possibilities.

I did the calculation of the 462 possibilities but I cannot make the same graph here is the excel sheet. I am not scaling the t distribution yet.

So my reasoning is as following, I calculate all the combinations, then I average them. After that, I make the difference between the total average and each combination average which gives me the difference of means, then I make the graphs and it does not correspond. And I cannot point what is wrong.

Thank you for your help,

Nick