CAEFFECT Procedure

OUTCOMEVAR Statement

OUTCOMEVAR outcome-variable <(EVENT='event-level')> </outcome-options> ;

OUTCOMEVAR events / trials </outcome-options> ;

The OUTCOMEVAR statement specifies information about the outcome variable. Information about the outcome variable is required for all estimation methods except the regression adjustment method. You can specify the outcome as a single variable or as a ratio of two variables that are denoted as events/trials. The first form applies to response variables that are continuous or categorical; the second form applies only to summarized binomial response data.

When you have binomial data, the events variable contains the number of positive responses (or events) and the trials variable contains the number of trials. The values of both events and (trials – events) must be nonnegative, and the value of trials must be positive. For binomial response data, the potential outcomes correspond to the event probability.

When you have categorical data, you must specify the following option:

EVENT='event-level': specifies the outcome event-level of interest.

For categorical response data, the potential outcomes correspond to the probability of the event level occurring.

You can specify the following outcome-options after a slash (/):

COUNTMISSING

treats missing values (".", ".A", . . . , ".Z" for numeric variables and blanks for character variables) as valid values of a categorical outcome variable.

LEVELIZERAW

bases the levelization of a categorical outcome variable on raw values.

MAXLEV=number

specifies the maximum number of levels for a categorical outcome variable. The number 0 means an unlimited number of levels.

TYPE=BINOMIAL | CATEGORICAL | CONTINUOUS

specifies the type of the outcome variable.

You can specify one of the following types:

BINOMIAL: specifies a binomial outcome. For a binomial outcome, the event probability is modeled, and predicted outcome values must be between 0 and 1, inclusive. If the specified estimation method requires the observed outcome values, you must use event/trial syntax. Event variable values must be nonnegative, trial variable values must be positive, and the number of events must be less than or equal to the number of trials.
CATEGORICAL: specifies a categorical outcome. For a categorical outcome, the probability of a designated event level is modeled, and the observed outcome and predicted outcome values must be between 0 and 1, inclusive. If the specified estimation method requires the observed outcome values, you must specify the event level by using the EVENT= option.
CONTINUOUS: specifies a continuous outcome. For a continuous outcome, the observed outcome and predicted outcome values can be any real values.

Last updated: May 14, 2026