Noncontact Rate

Unit-Level Noncontact Rate

At the unit level (household or business), the numerator of the noncontact rate can be computed by tallying up all those sampled units for which the researchers are certain contact attempts were made. In addition to these cases, the researchers must make an informed decision about what portion of the units for which it is uncertain if contact was made also should be included in the numerator (and the denominator). This uncertainty differs when the survey is interviewer-administered versus when it is done via mail or Internet. In the case of in-person surveys, interviewers who approach homes or businesses can make informed judgments about whether the unit looks to be occupied. If it is determined to be occupied and no contact is ever made with an occupant, then that unit must be included in the numerator (and the denominator) of the noncontact rate calculation. If it is determined the unit is not occupied then that unit is not counted as a non-contact and thus not counted in the numerator (and may or may not be counted in the denominator depending on whether all sampled cases or only “eligible” cases are included in the denominator). This estimation of eligibility (referred to as e) of which additional units to count as noncontacts is further complicated when a survey has unusual eligibility criteria (e.g. only adults aged 35–49 years), because some of the apparent noncontacts would have been found to actually be ineligible had contact been made. In these surveys, the researchers must make informed (and defensible) decisions about how to estimate which of these cases should be included in the numerator of the unit-level noncontact rate calculation. The denominator of the unit-level noncontact rate can be calculated either by including all cases, or by including all known eligible cases, or by including all known eligible cases plus an estimate (e) of the portion of unknown eligibility cases that are judged to be eligible. As noted above, for mail and Internet surveys, sampled cases from which there is no reply whatsoever to the researchers are very difficult to classify as to whether contact ever was made. Again, the researchers need to make a reasonable judgment about what proportion of these cases should be counted as eligible and what portion of these should be counted as being implicit refusals rather than as noncontacts in the noncontact rate calculation. Any of these cases that are counted as refusals should not enter into the noncontact rate numerator.

...