The null hypothesis of interest is that each of k specific regression coefficients
is zero. For this case K’ is a k ×p' matrix consisting of zeros except
for a single one in each row to identify the βj being tested; m = 0. With
this K’, the matrix multiplication [K'〖(X^' X)〗^(-1) K] extracts from 〖(X’X)〗^(-1)
the k × k submatrix consisting of the coefficients for the variances and covariances
of the k (β_j ) ̂being tested. Suppose the null hypothesis to be tested
is that β1, β3, and β5 are each equal to zero. The sum of squares Q has the
form