An error in polling introduced by encouraging one outcome over another. There are many reasons why survey results can be biased, from choosing participants who are not a representative sample of the group pollsters want to learn about to quirks of survey wording and the ordering of survey questions.

Summaries of the relationship between the results of different survey questions that allow for easy comparisons of differences across groups of survey takers. For example, cross-tabulations, or “crosstabs,” can reveal how support for political candidates varies depending on survey takers’ age, gender, or race.