The REPLICATE statement provides a way to accommodate models in which different subjects have identical data. This occurs most commonly when the dependent variable is binary. When you specify a REPLICATE variable, PROC NLMIXED treats its value as the number of subjects that have data identical to the data for the current value of the SUBJECT= variable (specified in the RANDOM statement). Only the last observation of the REPLICATE variable for each subject is used, and the replicate variable must have only positive values.
The function of a REPLICATE statement is related to but not quite the same as the function of either a FREQ or a WEIGHT statement in other statistical modeling procedures, such as the GLM, GENMOD, GLIMMIX, and LOGISTIC procedures. A FREQ or a WEIGHT value essentially multiplies the log likelihood or sum of squares contribution for each observation. On the other hand, a REPLICATE value multiplies the log likelihood contribution of each subject, which consists of one or more observations. When no SUBJECT= variable is specified, the REPLICATE value behaves like a weight, multiplying each observationâ€™s log likelihood contribution.
NOTE: The REPLICATE statement is not allowed when there is more than one RANDOM statement.