In the PATH statement, you specify the paths in your model. You can specify at most one PATH statement in a model, within the scope of either the PROC TCALIS statement or a MODEL statement. To complete the PATH model specifications, you might need to add some subsidiary model specification statements. The following is the syntax for the PATH modeling language:

PATH path lists ;

PVAR partial variance parameters ;

PCOV partial covariance parameters ;

MEAN mean parameters ;

In the PATH statement, you specify paths of variables in your model, excluding paths from the errors or disturbances. Paths in the PATH statement are separated by commas. In the PVAR statement, you specify parameters for either variances or partial variances. In the PCOV statement, you specify parameters for either covariances or partial covariances. In the MEAN statement, you specify parameters for either means or intercepts. For details about these subsidiary model specification statements, refer to the syntax of the individual statements.

In each path, you have one outcome variable and one predictor variable. The variables in a path can be manifest or latent. The specified direction determines the predictor and outcome variables. The variable being aimed at is the outcome variable and the other variable is the predictor variable. The path coefficient parameter is specified after each path in parameter-spec. For example, in the following statement you specify a model with four paths, each with predictor variable F1 and one outcome manifest variable.

   PATH 
      V1  <-  F1    1,
      V2  <-  F1    b2,
      V3  <-  F1    b2,
      V4  <-  F1    b4 (.4);

As shown in the example, the parameter-spec is one of the following three types:

• fixed path coefficients

  The parameter-spec for the first path to V1 is a fixed path coefficient at 1.

• free or constrained parameters without starting values

  The parameter-specs for the second and third paths to V2 and V3, respectively, are the same path coefficient parameter b2 with no starting value given. Because the same parameter name is used for the two path coefficients, the coefficients are constrained in the model.

• free or constrained parameters with starting values

  The parameter-spec for the last path to V4 is a free parameter named b4 with a starting value at .

If an outcome variable has more than one predictor variable, you have more than one path specification for the outcome variable. The following PATH statement shows an extension of the previous specification.

   PATH 
      V1  <-  F1    1,
      V3  ->  V1    b1 (.2),
      V2  <-  F1    b2,
      V3  <-  F1    b2,
      V4  <-  F1    b4 (.4);

As compared with the previous example, an additional path is specified in second path specification of the current PATH statement. In the specification V3 is a predictor of V1, in addition to the predictor F1 in the first path specification. The additional path V3 –> V1 has a path coefficient parameter named b1, with a starting value at 0.2.

Note that because the display of path results follows the way you specify the paths, you should arrange the order of the input paths in the most desirable way.

   proc tcalis;
      group 1 / data=d1;
      group 2 / data=d2;
      model 1 / group=1;
         path
            V1 <- F1   1.,
            V2 <- F1   load1,
            V3 <- F1   load2,
            F1 <- V4   b1,
            F1 <- V5   b2,
            F1 <- V6   b3;
         pvar 
            E1-E3    = ve1-ve3,
            F1       = vd1,
            V5-V6    = phi4-phi6;
         pcov
            V1 V2    = cve12;
      model 2 / group=2;
         refmodel 1;
         path 
            V3 <- F1   load1,
         pcov 
            V1 V2    =  .,
            V2 V3    = cve23;
      run;