#### Description:

This routine computes and normalises the David-Johnson approximation for the covariance matrix of
normal order statistics. The value V11 can be calculated using the PDA_V11 routine and the values of
EX1, EX2 and SUMM2 using PDA_NSCOR.

#### Invocation

CALL PDA_COVMAT( V, N, MDIM,
V11, EX1, EX2, SUMM2, IFAULT )

#### Arguments

##### V( MDIM, N ) = DOUBLE PRECISION
(Returned)

The covariance approximation.

##### N = INTEGER (Given)

The sample size.

#####
MDIM = INTEGER (Given)

First dimension of V as declared in the calling routine.

##### V11 =
DOUBLE PRECISION (Given)

Exact value of the extreme variance V(1,1).

##### EX1 = DOUBLE
PRECISION (Given)

Absolute expected value of the smallest order statistic from a size N
sample.

##### EX2 = DOUBLE PRECISION (Given)

Absolute expected value of the second
smallest order statistic from a size N sample.

##### SUMM2 = DOUBLE PRECISION (Given)

Sum of squares of expected values order statistics for a sample of size N.

##### IFAULT =
INTEGER (Returned)

Failure indicator. Zero for success, otherwise N is out of bounds.

####
Origin

Applied Statistics / Statlib Archive

#### Copyright

The Royal Statistical Society.