SPSS Mixed Linear Models 1
/*******************************************************************************
************************ SPSS Mixed Model Syntax *******************************
*******************************************************************************/
MIXED
*** Outcome variable ***
DV
*** Categorical fixed effects go after BY ***
BY catvar1 catvar2
*** Continuous (covariate) fixed effects go after WITH ***
WITH time contvar
*** Request specific output ***
/PRINT =
*** Case processing summary: factor values, repeated measures variables,
repeated measure subjects, random effects subjects & frequencies ***
CPS
*** Descriptive Statistics: sample sizes, means, SD of DV & covariates
for each distinct combination of these factors ***
DESCRIPTIVES
*** Iteration History (1=print every iteration) ***
HISTORY(1)
*** Solution (estimates) for fixed and random effects ***
SOLUTION
*** Tests for covariance parameters (asymptotic SE's & Wald tests) ***
TESTCOV
*** Estimated covariance matrix of random effects and residuals ***
G R
*** SAVE Subcommand - Saves casewise statistics to working file ***
/SAVE =
*** Fixed Predicted Values, their SE’s, and Satterthwaite DF
Regression means without random effects ***
FIXPRED SEFIXP DFFIXP
*** Predicted Values, their SE’s, and Satterthwaite DF
Model fitted value (includes random effects) ***
PRED SEPRED DFPRED
*** Residuals - data value-predicted value ***
RESID
*** Estimation Criteria & subcommands (defaults) – usually DON’T CHANGE THESE ***
/CRITERIA =
*** Confidence interval level of 95% (0-100%) ***
CIN(95)
*** Maximum # Iterations, Maximum Half-Stepping (positive integer) ***
MXITER(100) MXSTEP(5)
*** Use Fisher scoring up to iteration (positive integer) ***
SCORING(1)
*** Tolerance value in checking singularity (positive value) ***
SINGULAR(0.000000000001)
*** Convergence Criteria (non-negative value, type) -
Hessian, Log-likelihood, Parameter Estimates ***
HCONVERGE(0, ABSOLUTE) LCONVERGE(0, ABSOLUTE)
*** Method of Estimation (REML, ML) ***
/METHOD = REML
*** Fixed Effects ***
*** Intercept included by default, or must be specified first ***
*** Can include interactions between different variables with asterisk ***
*** Cannot include squared terms for categorical variables using asterisk ***
/FIXED = time catvar1 catvar2 contvar time*catvar1 time*contvar |
