Results Browser Window
The Results Browser Window allows you to examine output from models for which you have previously generated output and stored in the Results File. Displayed on the screen is a table of results, including the default columns Model Name, QAICc, Delta QAICc, and Deviance. If c-hat is equal to one, then AICc and Delta QAICc are displayed. In addition, the numerical output for the model is included in a memo field in the Results File data base, and can be viewed by clicking on the Output | Specific Model | Editor menu choices. By selecting File | Preferences, you can also change the model selection method to BIC, and add columns for -2log(L) or a time stamp. Model Notes are also stored in the Results Browser.
Under the Delete menu choice, you can delete a single model from the Results File, or clear the entire file. The clear option can be a disaster, so be careful — you must really want to delete all the models to select clear. If you accidentally delete a model that you want back, use the Undelete Model menu choice from the Delete menu.
Other menu options under the Output menu choice specific to the highlighted model in the Results Browser allow you to retrieve a previous model (including the Parameter Matrices and Design Matrix) for creating a new model, to print the numerical output from a model to the printer, and to produce a table of the model summary statistics in a NotePad Window. Variance components analysis for either the beta estimates or the real parameter estimates. The variance-covariance of either the beta estimates or the real estimates can be retrieved in a NotePad Window. Residuals can also be placed in a NotePad Window for study, or use in a different piece of software. Residuals can also be plotted, either as Pearson residuals or the default deviance residuals. Parameter estimates and their standard errors and confidence intervals can be copied to the clipboard.
Other menu choices under the Output menu choice allow you to Append a model to the current results file, append models from a subdirectory, perform model averaging on models in the results file, or view the input data used to create the models.
Under the Tests menu selection, you can construct Likelihood Ratio and ANODEV tests, parametric bootstrap goodness-of-fit simulations, median c-hat estimation, plus evaluate the probability level of chi-square and F statistics.
Under the Adjustments menu choice, options are available for changing the quasi-likelihood scaling parameter (c-hat, modifying the number of parameters estimable in the model (which changes the QAICc, AICc, BIC, or QBIC value), changing a model’s name, or setting the font size of the window..
Some of the models in the Results Browser Window may not have estimates associated with them because just the model’s structure has been saved. See Save Model Structure to learn how to create models and run them in a batch facility.