I cannot read that scan.

I bet I can answer your questions if you would retype it for me.

It looked like a linear combination of normals and also squares of normals, creating Chi-Square rvs.

If so, that should be rather easy.

There might be some t and F's in there too.

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

I blew it up and I had to laugh.

I'm teaching out of Wackerly (I also knew Dennis when I was at Florida)

this semester and I just assigned those two problems this year.

I even have the sol manual, but thats not necessary

37a each Y_i is a N(0,1)

so the square is a Chi-square with one df

the sum of indep chi-squares is a chi-square with the df each to the sum

of those df so the answer is chi-square with 5 dfs

37b U divided by sigma squared is a chi-square w n-1=4 df

but since sigma=1 here we don't need to see sigma.

Hence we have a chi-square w 4 dfs.

37b Thats U (from part b) plus an indep chi-square w 1 df

hence it's a chi-square w 5 df (add the dfs)

38a that's a t with 5 dfs, because its a standard normal divided

by a chi-square divided by its dfs and square rooted

38b that's a t with 4 dfs, for the same reason as in (a), just take

the 2 and rewrite him as square root of 4 and place him under the denominator inside a square root.

38c thats an F with 2 and 4 dfs, because you can write him as a chi-square (with 2 df)divided by two OVER a chi-square (with 4 dfs) divided by 4.