Self-selection bias

Self-selection (also self- selectivity) is a form of Arbitrary sampling. In the opinion research and empirical social research we speak of when examined in a sample of persons are not selected by a random or systematic sampling, but even make the decision to be part of the sample. Self-selection leads compared to random selection method means that the sample is not representative and draw conclusions about the characteristics of the population, therefore, only to a limited extent are allowed, even if the socio- demographic characteristics of the subjects studied should match those of the population.

Example of self-selection: A survey on smoking would not target a random sample from the population to be contacted, but a state be established in a pedestrian zone, ausliegen to the questionnaires. It might now want to join particularly Non smoking. The opinion of smokers would be under-represented it.

Self-selection is often used as improper selection in online surveys. The survey participants are passively recruited, so as to respond on a banner. Such persons are particularly interested in the subject of the survey usually, therefore, their data can not be transferred to the population of Internet users or site visitors.

Accruals

Self-selection as the selection is to be distinguished from systematic Unit Nonreponse at a random selection. If, for example, in empirical social research a particular group of people systematically fail to participate in an inquiry (unit nonresponse ), there will be systematic non-response and thus distortions (see item non-response bias ). These distortions can be corrected but with knowledge of appropriate characteristics ( Missing at Random ).

The self-selection is to be distinguished from the frame error, have no or a lesser chance in which certain people to get into the sample. An example would be here homeless people who have a chance at a telephone survey hardly to be reached by landlines.