Response bias

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Response bias (also known as the Hawthorne effect) is a type of cognitive bias which can affect the results of a statistical survey if respondents answer questions in the way they think the questioner wants them to answer rather than according to their true beliefs. This may occur if the questioner is obviously angling for a particular answer (as in push polling) or if the respondent wishes to please the questioner by answering what appears to be the "morally right" answer. An example of the latter might be if a woman surveys a man on his attitudes to domestic violence, or someone who obviously cares about the environment asks people how much they value a wilderness area.

This occurs most often in the wording of the question. Response bias is present when a question contains a leading opinion. For example, saying "Given that at the age of 18 people are old enough to fight and die for your country, don't you think they should be able to drink alcohol as well?" may yield a response bias, in comparison with simply "Do you think 18-year-olds should be able to drink alcohol?" Or, "Should the killing of babies be legalized?" as opposed to, "Should abortion be legalized?"

It also occurs in situations of voluntary response, such as phone-in polls, where the people who care enough to call are not necessarily a statistically representative sample of the actual population.

Non-response bias is not the opposite of "response bias" and is not a type of cognitive bias: it occurs in a statistical survey if those who respond to the survey differ in the outcome variable (for example, evaluation of the need for financial aid) from those who do not respond; or in other words, the statistical sample is not sufficiently random. For example, if a telephone survey is conducted, the results will only include people who are willing to sit and answer questions from a stranger over the telephone, and if those responses differ from the rest of the population, the survey results may be skewed. Often, the differences, which may include race, gender or socioeconomic status, are reported and/or accounted for through statistical modelling in any publication of the results.

See also[edit]

External links[edit]