Suppose your dependent variable is a rating, such as a likelihood to buy: very likely, somewhat likely, somewhat unlikely, very unlikely. What is the relative distance between these categories? Uniform scaling (3,2,1,0) may give very different results than scores such as (3,1,0.5,0), which suggest greater difference between "very likely" and the other levels. GOLDMineR estimates "optimal" dependent variable scores simultaneously with the estimation of the regression coefficients.
This model is an extension of Leo Goodman's RC association model.

Select effects coding (or other coding) for your nominal predictors with a simple mouse click AFTER you estimated the model (no need to re-estimate). See a unique graphical display which interprets the effects for you. See how the log-odds ratio compares each category with the average category (reference) under effects coding. Then click on any category in the graph to instantly transform to dummy coding, where the selected category becomes the new reference.

Select the most important predictors to include in model. If you have many correlated predictors, GOLDMineR's Search feature will leave out the ones that are collinear with those included in the model. Search ranks all predictors according to their unique p-value. You can then enter the most significant, one step at a time, reviewing the results of the model whenever you want. Click "search all" to run Search automatically until all significant predictors have been entered.

In addition to the standard regression type output, an interactive gains chart summary is available, which summarizes the performance of the model in deciles or other quartile groupings.

• Use innovative graphical displays and charts to better visualize and assess you model
• Use the table window to access various counts, odds and residual tables
• Use the search facility for automated predictor variable selection
• Update your results immediately for changes in contrast coding of qualitative variables
• Model and predict dichotomous and ordered categorical dependent variables
• Generate code for scoring external files
• View GOLDMineR's logit model graphically vs. a competing linear model applied to the same data
• Explore and assess competing models through visual displays and statistical measures of fit
• Use different scoring systems for your categorical variables
• Assess the relative impact of individual predictors on the response variable
• Include both categorical and continuous predictors into the model
• Build regression models for dependent variables containing ordered outcome categories (where the ordering can be prespecified or determined by the model)