Distortion introduced into a research investigation by incomplete collection of data from all possible respondents. This bias is most likely to influence the results of a study when those who do not participate or who refuse to be sampled have other crucial characteristics that the study is designed to identify.