I believe the answer is sampling method
During the sampling method, researchers will pick a certain amount of representatives from the population under study.
An example of how this could create a bias would be when the Researchers want to study the general population's opinion on immigration, but only choosing voters from a certain political party as a representative.