Usage
The exchangeable control can be used to force the exploitation
of exchangeability in the special circumstances where (a) the sample
size (fictitious or actual) is unity; and (b) beliefs specified for
exchangeable situations exist. Ordinarily a sample size of unity doesn't
require these beliefs for the evaluation of the adjustment. Use of this
control implies that the modelvar control also will be
referenced. This method of forcing the exploitation of exchangeability
is not available for the iterative adjustment algorithm.
When this method is used, then all the features commonly available to
exploit exchangeability are available even though the original sample
size was unity.
BD>control obs =E
The obs control determines the fictitious sample size that the
program assumes when it is preparing an adjustment in the case that no
actual data is available, or in the case that the usedata
control is switched off. If a mistake is made in defining the equation,
and an error is reported, then the previous value will be retained.
The usedata control determines whether the command uses any
available data to produce observed adjustments. When the control is on,
the program always tries to use data if available. If no data is
available, or if the control was switched off, there is no attempt to
use data, and the prospective sample size used is the one set by the
obs control.
The setting of the matchcd control determines whether or not
mean component and predictive component exchangeable adjustments carried
out via the ADJUST: command are computed using special
algorithms. We recommend that the special algorithms should be used as
this avoids much unnecessary calculation. Except in the case of rank
degeneracy, the results will be identical. In the case of rank
degeneracy, the special algorithms ensure that the canonical directions
for mean and predictive component adjustments are identical except for a
normalising constant; this typically gives canonical directions for the
mean component adjustment which are valid but different (because of
non-uniqueness) to the directions computed by the default algorithm.
Usage 
where E is any valid
equation which, when rounded, results in a positive integer.
.
Usage
Usage