Survey, Comprehensive

in statistics, a survey in which all units without exception of the set under study (object of observation) undergo inspection. Examples of comprehensive surveys include a population census and census of small-scale industry. Current statistical reports based on accounting are also a type of comprehensive survey, since they encompass all socialist enterprises and organizations without exception for the purpose of checking upon the fulfillment of the state plans.

Simultaneous surveys are one kind of comprehensive survey. For example, a simultaneous survey of workers and employees of industrial enterprises was taken in the USSR during the Great Patriotic War; the resulting data were used to rationally distribute and utilize personnel. The Central Statistical Board of the USSR takes twice every five years (1) a simultaneous survey of the number of workers by trade or profession, wage category, and payments system; (2) a simultaneous survey of workers and employees by sex, age, and length of service; and (3) a yearly survey of specialists possessing higher and secondary specialized education.

Nonsampling errors, that is, the divergences between the observed values and the actual values of the phenomena being studied, occur during comprehensive surveys. Such errors appearing as a consequence of different random causes (incorrect statements by the respondents, random interchanging of digits) are called random errors. Errors arising as a consequence of specific causes are said to be systematic errors (for example, rounding off the age of the population). The data from comprehensive survevs are controlled and verified from the standpoint of completeness of coverage of the object and from the standpoint of quality. Thus, a comprehensive survey of livestock will be followed by a sampling taken from 10 to 20 percent of the sovkhozes, kolkhozes, and households in order to verify the completeness of the coverage.