The CALIS Procedure

MODEL Statement

  • MODEL i </ options>;

where i is an assigned model number between 1 and 9999, inclusively.

A MODEL statement signifies the beginning of a model specification block and designates a model number for the model. All main and subsidiary model specification statements after a MODEL statement belong in that model until another MODEL or GROUP statement is encountered.

The MODEL statement itself does not serve the purpose of model specification, which is actually done by the main and subsidiary model specification statements that follow it. The MODEL statement serves as a "place-holder" of specification of a single model. It also makes the reference to a model easier with an assigned model number. For example, consider the following statements:

proc calis;
   group 1 / data=women_data;
   group 2 / data=men_data;
   model 1 / group=1 label='Women Model';
      {model 1 specification here}
   model 2 / group=2 label='Men Model';
      {model 2 specification here}
run;

This example illustrates a two-group analysis with two models. One is model 1 labeled as 'Women Model' in a MODEL statement. Another is model 2 labeled as 'Men Model' in another MODEL statement. The two groups, group 1 and group 2, as defined in two separate GROUP statements, are fitted by model 1 and model 2, respectively, as indicated by the GROUP= option of the MODEL statements. Within the scope of model 1, you provide model specification statements by using the main and subsidiary model specification statements. Usually, one of the following main model specification statements is used: FACTOR , LINEQS , LISMOD , MSTRUCT , PATH , RAM , or REFMODEL . Similarly, you provide another set of model specification statements within the scope of model 2.

Hence, for an analysis with a single group, the use of the MODEL statement is not necessary because the model that fits the group is unambiguous.

You can set model-specific options in each MODEL statement. All but two of these options are also available in the PROC CALIS statement. If you set these options in the PROC CALIS statement, they apply to all models, unless you respecify them in the local MODEL statements. If you want to apply some options only to a particular model, specify these options in the MODEL statement that corresponds to that model.

You can also set group-specific options in the MODEL statement. These group options apply to the groups that are specified in GROUP= option of the MODEL statement. See the section Options Available in the GROUP and PROC CALIS Statements for a detailed descriptions of these group options.

Options Available Only in the MODEL Statement

LABEL=name
NAME=name

specifies a label for the model. You can use any valid SAS names up to 256 characters for labels. You can also use quoted strings for labels.

GROUP= [int-list]  |  {int-list}
GROUPS= [int-list]  |  {int-list}

specifies a list of integers which represent the groups to be fitted by the model.

Options Available in the MODEL and PROC CALIS Statements

The following options are available in the MODEL and PROC CALIS statements. If you specify these options in the PROC CALIS statement, they are transferred to all MODEL statements. These options might be overwritten by the respecifications in the local MODEL statements.

Option

Description

DEMPHAS=

Emphasizes the diagonal entries

EFFPART | TOTEFF

Displays total, direct, and indirect effects

EXTENDPATH | GENPATH

Displays the extended path estimates that include variances and covariances

INEST=

Specifies the data set that contains the initial values and constraints

INMODEL | INRAM=

Specifies the data set that contains the model specifications

MEANSTR

Analyzes the mean structures

MODIFICATION

Computes modification indices

NOMEANSTR

Deactivates the inherited MEANSTR option

NOMOD

Suppresses modification indices

NOORDERSPEC

Displays model specifications and results according to the input order

NOPARMNAME

Suppresses the printing of parameter names in results

NOSTAND

Suppresses the standardized output

NOSTDERR

Suppresses standard error computations

ORDERSPEC

Orders the model output displays according to the parameter types within each model

OUTEST=

Specifies the data set that outputs the estimates and their covariance matrix

OUTMODEL | OUTRAM=

Specifies the output data set for storing the model specification and results

PARMNAME

Displays parameter names in model specifications and results

PDETERM

Computes the determination coefficients

PESTIM

Prints parameter estimates

PINITIAL

Prints initial pattern and values

PLATCOV

Computes the latent variable covariances and scoring coefficients

PRIMAT

Displays results in matrix forms

READADDPARM

Instructs generated default parameters be read in the INMODEL= data set

STDERR

Computes the standard errors

Options Available in MODEL, GROUP, and PROC CALIS Statements

Some options in the GROUP statement can also be specified in the MODEL statements. Group options that are specified the MODEL statements are transferred to the GROUP statements that define the groups that are fitted by the associated models in the MODEL statements. This is a little more convenient than setting the common group options individually in the GROUP statements for all fitted groups by a model. See the section Options Available in GROUP, MODEL, and PROC CALIS Statements for a reference of these options.