Elements of Survey Sampling

This is because in social sciences the factors that cause variation in the measurements on the study variable for the population units can not be controlled, whereas in physical sciences these factors can be controlled, at least to some extent, through proper experimental design.