How to test this kind statistical hypothesis ?


TS Contributor
Usually exact test are rare. Even though in general you may form the likelihood ratio test statistic, the exact distribution is intractable and the critical region relies on simulation/numerical method. More common approach is forming the asymptotic tests.

But these are not the most important issues. The most important one is that: what is your data structure? model? assumptions?
Can you define your parameters more clearly? any likelihood function you can write down?
I don't have specific data or assumptions. This is the problem submitted by one of professors as a matter of defending the thesis. But now I see that "+" is not shown properly and easy question has changed to huge one ... :confused:

\({H_{0} \: : \;\;\;\; \frac{{\mu_{1}} + {\mu_{3}}}{2}=\mu_{2}}\)
Last edited:


Ambassador to the humans
The Tex tags are kind of weird here. The 'math' tags work a little better and end up rendering what you would expect.