biased sample(redirected from Sampling bias)
Also found in: Medical, Wikipedia.
biased sample[¦bī·əst ′sam·pəl]
biased samplea population SAMPLE which is not a true reflection of the parent population (see BIAS 2), i.e. not a REPRESENTATIVE SAMPLE.
When the incidence of a certain occurrence or piece of behaviour in a population is to be investigated, e.g. voting intention, it is often impossible to examine the total population, so a sample of this population is taken. For this sample to produce acceptable data, it must be a true representation of the parent population, so it is essential that it is selected in a way that ensures this. If this is not managed, bias will result and the information collected will not truly reflect the population being studied. Thus, to select a sample by questioning people in the street will bias it against people who do not walk, do not go shopping, are at work or school all day Postal QUESTIONNAIRES attempt to overcome this type of bias, but are likely to be biased against those who do not bother to fill in questionnaires and re turn them, and against the illiterate. To keep bias to a minimum, if random sampling is not possible, it is necessary to select the sample carefully by matching all relevant parameters of the population, e.g. age, class, residence, etc, and to ensure as high a response rate as possible, probably by personal INTERVIEWS.