Previous Page

|

Next Page

The Plots Tab

You can use the Plots tab (Figure 23.4) to create plots that graphically display results of the analysis. There are plots that help you to visualize the fit, the residuals, and various influence diagnostics.

Creating a plot often adds one or more variables to the data table. The following plots are available:

Predicted probability vs. One continuous covariate

creates a line plot of the predicted probability versus the continuous explanatory variable. This plot is created only if the following conditions are satisfied:

• There is exactly one continuous explanatory variable.
• There are three or fewer classification variables.
• There are 12 or fewer joint levels of the classification variables.

ROC curve

creates a line plot that shows the trade-off between sensitivity and specificity. Models that fit the data well correspond to an ROC curve that has an area close to unity. A completely random predictor would produce an ROC curve that is close to the diagonal and has an area close to 0.5.

Pearson chi-square residuals vs. Predicted

creates a scatter plot of the Pearson chi-square residuals versus the predicted probabilities.

Deviance residuals vs. Predicted

creates a scatter plot of the deviance residuals versus the predicted probabilities.

Change in Pearson chi-square vs. Predicted

creates a scatter plot of the DIFCHISQ statistic versus the predicted probabilities.

Change in deviance vs. Predicted

creates a scatter plot of the DIFDEV statistic versus the predicted probabilities.

Confidence interval displacement (C) vs. Predicted

creates a scatter plot of the confidence interval displacement diagnostic () versus the predicted probabilities.

Confidence interval displacement (C) vs. Observation number

creates a scatter plot of the confidence interval displacement diagnostic () for each observation.

Leverage (H) vs. Observation number

creates a scatter plot of the leverage statistic for each observation.