A census is where you poll everybody; a sample is where you choose a group of people to poll. Taking a census is inefficient because you can get almost completely accurate information from a large sample, but a sample would take significantly less time than a census since you wouldn't have to poll absolutely everybody in a population.
