Robert has reviewed a number of studies that have manipulated the type of injuries a person receives as a result of a drunk driving accident. He then statistically combines the results from all studies to reach a general conclusion of the effect of type of injuries received. Robert has most likely conducted a :