Continuous effects cannot be random Can save only if all regressors are continuous This will cause a huge formula. Tested against reduced model: Y=mean Tested against reduced model: Y=0 WARNING: Used Synthetic Denominator from Test WARNING: Non-Testable Contrast Contrast Specification Click on + or - to make contrast values. Contrast Dialog for Effect Odds Upper Odds Lower Odds Ratio Alpha out-of-range Specify alpha for 1-alpha Confidence Level Upper Lower Fit: Confidence Intervals Can't do ChiSquare quantile for confid interval. Likelihood Ratio Tests Probability This could take a long time. Are you sure? Prob F F Ratio Numerator DF Sum of Squares SS Prob>|t| t Ratio Std Error Estimate Contrast Signif. Prob Correlation Bivariate Normal Ellipse P= Can only handle responses with 2 levels Y by X Model Fit Likelihood Ratio Test Max RSq Observations (or Sum Wgts) Mean of Response Root Mean Square Error RSquare Adj RSquare Summary of Fit Specify Smoothness Value for Spline Smoothing Spline Fit, lambda= Sum of Squares Error R–Square Predicted  Residuals  SSE Responses should have the same measurement levels. No Y's selected, or personality unchosen Polynomial Degree must be between 1 and 5 Effect Attributes Macros Effect Prob>ChiSq ChiSquare Nparm Wald ChiSquare Prob>ChiSq ChiSquare -LogLikelihood DF Source Analysis of LogLikelihood Observations (or Sum Wgts) RSquare (U) Summary of Fit Singularity Details Mean Std Error Least Sq Mean Level Least Squares Means Prob>|t| t Ratio Std Error Estimate DF Term Parameter Estimates Total Error Pure Error Lack of Fit C Total Error Model Effect Test Nparm Prob>F F Ratio Mean Square Sum of Squares DF Source Analysis of Variance > Weight > > Censor > {RS} {Variance} {Random} Intercept Model Fit