Log-Linear Analysis

Complete implementation of log-linear modeling procedures for multi-way frequency tables; you can analyze up to 7-way tables, and automatically find the best model.

Element Name Description
General
Detail of computed results reported Specifies the detail of computed results reported. If Minimal detail is requested, only the summary results and marginal tables for the final model will be reported; select All results to review various plots of fitted and observed frequencies; tables of observed and fitted frequencies, as well as various residual statistics, etc. can be requested as an option.
Input file Specifies whether the input data file contains raw data (and needs to be tabulated), or contains (in the dependent variable list) frequency counts for an already computed table.
Automatic model Select this option to have STATISTICA determine automatically the best fitting model for the table. See also the selections on the Automatic Model Search tab for additional options.
All associations models Tests all marginal and partial association models; the partial association between, for example, two factors 1 and 2 (denoted by 12) is computed by comparing the fit (i.e., evaluating the Chi-square difference) of the model that includes all two-way interactions with the model that excludes the interaction between factors 1 and 2. The marginal association between, for example, two factors 1 and 2 is computed by comparing the fit of the model including all main effects (i.e., all effects of lower order than the one of interest) with the model including the interaction between factors 1 and 2.
Model to be tested Specifies the model to be tested, or specify no model (leave blank) to let STATISTICA determine the best model (see also the Automatic Search tab).

 The logic of specifying models in STATISTICA is as follows:

 Refer to factors in the table by their respective numbers. Use those numbers in order to specify main effects and interactions. For example, in order to specify a model with no interactions but only main effects for each of three factors, enter 1 2 3 (separated by spaces or commas). In order to specify a model with a 2-way interaction between factors 1 and 3, and the main effect for factor 2, enter 13 2 (or 2 31; the order does not matter). Higher-order interactions are specified analogously (e.g., 124 134 15).
Parameter delta The constant Delta (0.5 by default) is added to all frequencies in the observed table before it is submitted to the actual analysis. This is a recommended procedure when the table contains several cells with low frequencies.
Number of iterations Specifies the maximum number of iterations to be performed during the parameter estimation.
Convergence criterion Set the convergence criterion value (by default, 0.0001); refer to the Electronic Manual for details.
Automatic Model Search
p(1) Specifies parameter p(1) for the automatic search algorithm; STATISTICA first fits a model with no interactions, then with all two-way interactions, three-way interactions, etc., until a model that fits at the p-value greater than p(1) has been found.
p(2) Specifies parameter p(2) for the search algorithm; STATISTICA eliminates from the initial model in a stepwise manner all effects that do not contribute significantly to the overall fit at the p(2) level of significance.
Tables
Observed frequencies Displays table of observed frequencies (+ parameter Delta; be sure to set Delta to 0 if you want actual observed frequencies instead of adjusted frequencies; see the Electronic Manual for details).
Fitted frequencies Displays table of fitted frequencies for the final model.
Residual frequencies Displays table of residuals (observed - fitted) frequencies for the final model.
Residual standardized freqs Displays table of standardized residuals (observed - fitted) frequencies for the final model.
Components of Chi-square Displays table of components of Chi-square for the final model.
Freeman-Tukey deviates Displays table of Freeman-Tukey deviates for the final model.