Contents - Index


Residuals Plot


Residuals are plotted against their capture history number to assess the fit of the model.  The default for the residual plot icon is deviance residuals.  However, either deviance residuals or Pearson residuals are available from the  Output | Specific Model Output Menu of the Results Browser

A deviance residual is defined as

Sign(Observed - Expected) Sqrt{2 [ Expected - Observed +  Observed Log(Observed/Expected) ] / c-hat }

where Sign is the sign (plus or minus) of the value of Observed - Expected, Sqrt is square root, and Log is the natural log.  

A Pearson residual is defined as 

(Observed - Expected) / Sqrt(Expected * c-hat)

The value of c-hat is the over dispersion scale parameter, and is normally taken as 1.

To see which observation is causing a particular residual value, place your mouse cursor on the plotted point for the residual, and click the left button.  A description of this residual will be presented, including the group that the observation belongs to, the encounter history, the observed value, the expected value, and the residual value.  After you have read the information, close this message box by clicking on the OK button.

To exit the Residuals Plot Window, click on the Close button.