Today I understood something really well.

I never knew why sample variance needed to be divided by (n-1). Today I spent sometime taking expectations of sample variance and proving that by dividing by (n-1) you make the estimate unbiased.

Also never understood properly the reason for S square * ( n-1) / sigma squared to follow chisquare (n-1) distribution . While working on an exercise problem in Annette Dobson, I got the answer using Cochrane’s theorem

image

Solved c using Cochrane theorem

image