SIZE Statement

SIZE variable ;

The SIZE statement names one and only one variable that contains size measures that are used for PPS selection. The SIZE variable must be numeric.

If you specify a SAMPLINGUNIT statement with a SIZE statement, the procedure computes a sampling unit’s size by summing the size measures of all observations that belong to the sampling unit. Alternatively, if you specify the PPS option in the SAMPLINGUNIT statement and do not use a SIZE statement, the procedure computes sampling unit size as the number of observations in the sampling unit.

When the value of a sampling unit’s size measure is missing or nonpositive, that sampling unit is excluded from the sample selection. See the section Missing Values for more information.

You can adjust the size measure values by using the MAXSIZE= option, the MINSIZE= option, or both of these options in the PROC SURVEYSELECT statement.

All PPS selection methods require size measures, which you can provide by specifying a SIZE statement (or by specifying the PPS option in the SAMPLINGUNIT statement). PPS selection methods include the following: METHOD=PPS, METHOD=PPS_BREWER, METHOD=PPS_MURTHY, METHOD=PPS_SAMPFORD, METHOD=PPS_SEQ, METHOD=PPS_SYS, and METHOD=PPS_WR. For details about how size measures are used in sample selection, see the descriptions of PPS selection methods in the section Sample Selection Methods.

Note that a sampling unit’s size measure, which you provide for PPS selection by specifying a SIZE statement, is not the same as the sample size. The sample size is the number of units to select for the sample; you specify the sample size with the SAMPSIZE= option in the PROC SURVEYSELECT statement.

For METHOD=POISSON, the variable that is specified in the SIZE statement provides inclusion probabilities for Poisson sampling. See the section Poisson Sampling for details. When the value of the SIZE variable is missing, nonpositive, or greater than 1, the sampling unit is not included in the sample selection.