Running the Analysis

To identify unusual cases, from the menus choose:
Data > Identify Unusual Cases...

Figure 1. Identify Unusual Cases dialog box, Variables tab
Select Age category through Stroke between 3 and 6 months as analysis variables.
Select Patient ID as the case identifier variable.
Click the Output tab.
Figure 2. Identify Unusual Cases dialog box, Output tab
Select Peer group norms, Anomaly indices, Reason occurrence by analysis variable, and Cases processed.
Click the Save tab.
Figure 3. Identify Unusual Cases dialog box, Save tab
Select Anomaly index, Peer groups, and Reasons.
Saving these results allows you to produce a useful scatterplot that summarizes the results.
Click the Missing Values tab.
Figure 4. Identify Unusual Cases dialog box, Missing Values tab
Select Include missing values in analysis. This process is necessary because there are a lot of user-missing values to handle patients who died before or during treatment. An extra variable that measures the proportion of missing values per case is added to the analysis as a scale variable.
Click the Options tab.
Figure 5. Identify Unusual Cases dialog box, Options tab
Type 2 as the percentage of cases to consider anomalous.
Deselect Identify only cases whose anomaly index value meets or exceeds a minimum value.
Type 3 as the maximum number of reasons.
Click OK.