Impute Function

Format:

impute (item-name, expression)

    [stat (item-name1, item-name2,....,item-nameN)]

    [title (alpha-expression)]

    [vset (vset-number)]

    [specific];

 [ ] indicates that this part is optional.

Description:

The impute statement assigns a value to a data item and logs the frequency of assignments. The parameters are:

 • item_name:

The dictionary data item to impute. The item must be numeric, with or without decimals, and can be single or multiple. If the item is multiple and is being used inside its PROC, the current occurrence is assumed. If the item is multiple and is being referenced outside its PROC, an occurrence must be specified. Occurrence numbers are 1-based.

 • expression:

 Is a variable name, a number, or an arithmetic expression; for example, 'TEMPAGE', '5', or '2+3'.

The options are:

 • STAT (item_name1[, item_name2]):

Tells the system to generate the secondary .dcf and .dat files. These files contain references to the data items that were changed; i.e., identifying ID values, the data item being imputed, and each subsequent data item named in the stat parameter list.

 • TITLE (alpha_expr):

Under the "IMPUTE STATISTICS" heading at the top of each page, this line will replace the default line that is generated ("IMPUTED Item (<unique name of data item>): <label name of data item>").

 • VSET (vset_number):

Is the 1-based value set number of the item being imputed (i.e., impute_item_name), and corresponds to the order of listing in the data dictionary (i.e., the first value set of an item will be number 1, the second value set will be number 2, etc). This may yield a different number of frequencies than what occurred when not using this option. For example, if you are imputing "age", and do not use the VSET option, your report will show the total number of imputations that occurred. However, if you use the VSET option, and the value set you choose does not list all possible values, then the total number of imputations listed in the frequency report will most likely be less than that given if you did not use this option.

 • SPECIFIC:

Indicates if the frequency will be generated alone or not. You can have multiple impute statements for a single data item. If you do this, you may want to have frequency reports separated for each impute. If specific is not used, all imputations for a given data item will be lumped together in the frequency report.

The impute command can generate up to three files:

  <xxx>.frq

 <xxx>.dat (only generated if the STAT option is used)

 <xxx>.dcf (only generated if the STAT option is used)

where XXX corresponds to the name of the data file used in the run. These files will be placed in the directory where the .bch application is located.

The format of the report contained in the .frq file is divided up into five columns as follows:

Example:

impute(P04_AGE, TEMPAGE) title("Age updated via TempAge") vset(2);

Code Example:

A code example of this statement can be found in the Examples/Impute folder in the CSPro software folder.