Weighting & Inserting

Discussions about the usability of web surveys most often mention different nonobservation errors. The first one among them is noncoverage error, which occurs because certain units are not a part of the sampling frame or the list that is used to identify members of the target population. Web surveys that target the general population typically suffer from noncoverage error, because the Internet is not used by all members of the population.

Researchers try to surpass noncoverage errors by using the appropriate weighting methods, which attempt to bring internet users closer to the general population according to their sociodemographic characteristics. With the use of weighting it is possible to adapt the respondents to the actual structure of the internet users.

Weighting and inserting data are methods that help us achieve the desired coverage of responses or to influence the nonresponse and bias rates, which can occur due to reasons that were discussed earlier. It is important to keep in mind that statistical inference – despite weighting and/or inserting data – is only possible if we use a probability sample.