MODEL Statement
MODEL dependent = <effects> </ options> ;

The MODEL statement specifies the dependent (response) variable and the independent (regressor) variables or effects. The dependent variable must be numeric. Each term in a MODEL statement, called an effect, is a variable or a combination of variables. You can specify an effect with a variable name or with special notation by using variable names and operators. For more information about how to specify an effect, see the section Specification of Effects in Chapter 41, The GLM Procedure.

Only one MODEL statement is allowed for each PROC SURVEYREG  statement. If you specify more than one MODEL statement, the procedure uses the first model and ignores the rest.

You can specify the following options in the MODEL statement after a slash (/):

ADJRSQ

requests the procedure compute the adjusted multiple R-square.

ANOVA

requests the ANOVA table be produced in the output. By default, the ANOVA table is not printed in the output.

CLPARM

requests confidence limits for the parameter estimates. The SURVEYREG  procedure determines the confidence coefficient by using the ALPHA= option, which by default equals 0.05 and produces 95% confidence bounds. The CLPARM option also requests confidence limits for all the estimable linear functions of regression parameters in the ESTIMATE statements.

Note that when there is a CLASS statement, you need to use the SOLUTION option with the CLPARM option to obtain the parameter estimates and their confidence limits.

COVB

displays the estimated covariance matrix of the estimated regression estimates.

DEFF

displays design effects for the regression coefficient estimates.

DF=value

specifies the denominator degrees of freedom for the F tests and the degrees of freedom for the t tests. For details about the default denominator degrees of freedom, see the section Denominator Degrees of Freedom for details.

I | INVERSE

displays the inverse or the generalized inverse of the matrix. When there is a WEIGHT variable, the procedure displays the inverse or the generalized inverse of the matrix, where is the diagonal matrix constructed from WEIGHT variable values.

NOINT

omits the intercept from the model.

PARMLABEL

displays the labels of the parameters in the "Estimated Regression Coefficients" table, if the effect contains a single continuous variable that has a label.

SINGULAR=value

tunes the estimability checking. If is a vector, define ABS() to be the largest absolute value of the elements of . For a row vector of the matrix , define

     

If ABS() is greater than *value, then is declared nonestimable. Here, is the matrix . The value must be between 0 and 1; the default is .

SOLUTION

displays a solution to the normal equations, which are the parameter estimates. The SOLUTION option is useful only when you use a CLASS statement. If you do not specify a CLASS statement, PROC SURVEYREG  displays parameter estimates by default. But if you specify a CLASS statement, PROC SURVEYREG  does not display parameter estimates unless you also specify the SOLUTION option.

VADJUST=DF | NONE

specifies whether to use degrees of freedom adjustment in the computation of the matrix for the variance estimation. If you do not specify the VADJUST= option, by default, PROC SURVEYREG  uses the degrees-of-freedom adjustment that is equivalent to the VARADJ=DF option. If you do not want to use this variance adjustment, you can specify the VADJUST=NONE option.

X | XPX

displays the matrix, or the matrix when there is a WEIGHT variable, where is the diagonal matrix constructed from WEIGHT variable values. The X option also displays the crossproducts vector or .