Non-sampling errors in household surveys

Both sampling errors and non-sampling errors need to be controlled and reduced to a level at which their presence does not defeat the usefulness of the final survey results. In the preceding chapters on sample design and estimation methodology, much attention was focused on sampling error and somewhat less on other sources of variation in surveys, such as nonresponse and non coverage, which make up the class of errors known collectively as non-sampling errors. Non-sampling errors are particularly harmful when they are non-random, because of the bias they introduce in survey estimates.

Related Subject(s): Economic and Social Development