The PLM Procedure

SHOW Statement

SHOW options ;

The SHOW statement uses the Output Delivery System to display contents of the item store. This statement is useful for verifying that the contents of the item store apply to the analysis and for generating ODS tables. Table 69.10 summarizes the options available in the SHOW statement.

Table 69.10: SHOW Statement Options

Option

Description

ALL

Displays all applicable contents

BYVAR

Displays information about the BY variables

CLASSLEVELS

Displays the Class Level Information table

CORRELATION

Produces the correlation matrix of the parameter estimates

COVARIANCE

Produces the covariance matrix of the parameter estimates

EFFECTS

Displays information about the constructed effects

FITSTATS

Displays the fit statistics

HESSIAN

Displays the Hessian matrix

HERMITE

Generates the Hermite matrix $\bH = (\bX ’\bX )^{-}(\bX ’\bX )$

PARAMETERS

Displays the parameter estimates

PROGRAM

Displays the SAS program that generated the item store

XPX

Displays the crossproduct matrix $\bX ’\bX $

XPXI

Displays the generalized inverse of the crossproduct matrix $\bX ’\bX $


You can specify the following options after the SHOW statement:

ALL | _ALL_

displays all applicable contents.

BYVAR | BY

displays information about the BY variables in the source item store. If a BY statement was present when the item store was created, the PLM procedure performs the analysis separately for each BY group.

CLASSLEVELS | CLASS

displays the Class Level Information table. This table is produced by the PLM procedure by default if the model contains effects that depend on classification variables.

CORRELATION | CORR | CORRB

produces the correlation matrix of the parameter estimates. If the source item store contains a posterior sample of parameter estimates, the computed matrix is the correlation matrix of the sample covariance matrix.

COVARIANCE | COV | COVB

produces the covariance matrix of the parameter estimates. If the source item store contains a posterior sample of parameter estimates, the PLM procedure computes the empirical sample covariance matrix from the posterior estimates. You can convert this matrix into a sample correlation matrix with the CORRELATION option in the SHOW statement.

EFFECTS

displays information about the constructed effects in the model. Constructed effects are those that were created with the EFFECT statement in the procedure run that generated the source item store.

FITSTATS | FIT

displays the fit statistics from the item store.

HESSIAN | HESS

displays the Hessian matrix.

HERMITE | HERM

generates the Hermite matrix $\bH = (\bX ’\bX )^{-}(\bX ’\bX )$. The PLM procedure chooses a reflexive, $g_2$-inverse for the generalized inverse of the crossproduct matrix $\bX ’\bX $. See Important Linear Algebra Concepts of Chapter 3: Introduction to Statistical Modeling with SAS/STAT Software, for information about generalized inverses and the sweep operator.

PARAMETERS<=n>
PARMS<=n>

displays the parameter estimates. The structure of the display depends on whether a posterior sample of parameter estimates is available in the source item store. If such a sample is present, up to the first 20 parameter vectors are shown in wide format. You can modify this number with the n argument.

If no posterior sample is present, the single vector of parameter estimates is shown in narrow format. If the store contains information about the covariance matrix of the parameter estimates, then standard errors are added.

PROGRAM<(WIDTH=n)>
PROG<(WIDTH=n)>

displays the SAS program that generated the item store, provided that this was stored at store generation time. The program does not include comments, titles, or some other global statements. The optional width parameter n determines the display width of the source code.

XPX | CROSSPRODUCT

displays the crossproduct matrix $\bX ’\bX $.

XPXI

displays the generalized inverse of the crossproduct matrix $\bX ’\bX $. The PLM procedure obtains a reflexive $g_2$-inverse by sweeping. See Important Linear Algebra Concepts of Chapter 3: Introduction to Statistical Modeling with SAS/STAT Software, for information about generalized inverses and the sweep operator.