Settings for All Function Types

In the case of each of the three function types, the model field parameters offer options for controlling how outliers and missing values are treated.

For discrete model fields, values not belonging to those entered explicitly or to the most frequent values are considered outliers. For continuous model fields, outliers are those values falling outside of the outer limits that were either entered explicitly or determined automatically when the value ranges were defined. You can set one of the following system reactions for dealing with occurrences of a data record containing an outlier:

Cancel processing

Ignore the data record

Assign the default score value.

With continuous model fields, you can specify that outliers are extrapolated. With discrete model fields, you can specify that all outliers are treated like a single value (as a remainder).

You first need to identify missing values before the system can treat them. If a model field, such as Occupational Group, takes a two-digit number code and code 99 stands for unknown occupational group, then you can enter 99 as the missing value to treat such values separately. You can set one of the following system reactions for dealing with occurrences of a data record containing a value defined in this way:

Cancel processing

Ignore the data record

Assign the default score value.

You can also set a substitute value.