Model Checking

This menu produces diagnostic plots useful for checking the assumptions needed to interpret the analysis of regression models.

Type of Graph

Specify how the chosen statistics are plotted: against the fitted values, or an index variate (the only index currently available is the unit numbers of the observations), or as Normal or half-Normal plots, or as a histogram. The default is a composite diagram, including a plot against the fitted values, Normal and half-Normal plots, and a histogram.

Display in Graph

Select the quantity to be plotted: residuals, Cook's statistics or leverages. Leverages measure the influence of each unit on the fitted model, and Cook's statistics combine the size of a residual together with the influence. Only residuals can be selected after fitting curves.

Type of Residual

When plotting residuals, you can select deviance or Pearson residuals, both of which are standardized, or studentized, by dividing by their standard errors. (These two types of residual are the same for linear models, but not for generalized linear models.) Alternatively, you can choose deletion residuals or simple (not standardized) residuals.