Application of external validity with a calculated indicator
Non-responses caused by refusals became the main reason for surveys’ failure to provide representative data. Refusals do not occur randomly and they distort representation of certain social groups and subgroups in survey data. Criteria for external validity are seldom mentioned in publications dealing with nonresponse bias in surveys. We propose the method for estimating certain biases of sampling by calculating indicators based on sample data, which could be compared with reliable statistical data from some independent source.