Bias is anything that produces systematic error (i.e., errors that are consistent, not random) in research or evaluation findings. Random error tends to balance out in the long run, whereas systematic errors continue to grow in the long run and skew the results in a particular way.

For example, bias can be introduced in sampling. If you conducted a survey of Edmonton residents by standing on a busy corner on Saturday mornings, your sample would likely be biased. You would systematically exclude people who work on Saturdays (perhaps people working more than one job, or in the retail and restaurant industries), people who attend a Shabbat service on Saturday mornings, and people who live in parts of the city far from that one street corner. As a result, your survey sample will not be representative of the population of Edmonton.

See also: sampling

Return to the Evaluation Dictionary