SUBROUTINE PDA_NSCOR( S, N, N2, IER )
algorithm as 177.3, applied statistics, v.31, 161-165, 1982.
calculates approximate expected values of normal order statistics.
claimed accuracy is 0.0001, though usually accurate to 5-6 dec.
arguments:
s(n2) = output, the first n2 expected values (double precision).
n = input, the sample size.
n2 = input, the number of order statistics required; must
be <= n/2.
ier = output, error indicator
= 0 if no error detected
= 1 if n <= 1.
= 2 if n > 2000, in which case the order statistics
are still calculated, but may be inaccurate.
= 3 if n2 > n/2 (n.b. this differs from the
published algorithm which returns an error
if n2 is not equal to n/2.)
calls PDA_PPND16 = a variation of PPND7 in algorithm AS 241.
author: royston, j.p