Unlisted Household

An unlisted household is a residential unit that does not have a telephone number listed in published directories or one that would be given out by directory assistance. A telephone number may be unlisted for several reasons. Household residents actively may have chosen to keep their telephone number unlisted. Alternatively, although the residents of an unlisted household may not have specifically chosen to keep their telephone number unlisted, the number may be too new to be found in published directories or via directory assistance. This can be a result of a recent move or change in telephone number.

Although the literature suggests important differences between the characteristics of respondents who choose to keep their telephone numbers unlisted and respondents whose telephone numbers are unlisted for other reasons, consistent findings have been reported on the larger differences between the characteristics of households with listed numbers versus those with unlisted numbers. Directory-listed sampling frames have tended to exhibit an overrepresentation of (a) established households, (b) middle- and higher-income households, (c) two or more person households, (d) older householders, (e) married householders, (f) better-educated householders, (g) retired householders, (h) white-race households, and (i) home owners.

In general, sampling frames based on listed telephone numbers exclude unlisted households. This coverage shortfall has the potential to introduce bias into surveys that use a directory-listed sampling frame to construct a representative sample of the general population of telephone households. The coverage issue is most acute in urban areas and the western region of the United States, where unlisted telephone rates are higher. Additionally, as the proportion of households without listed telephone numbers in the general population increases, the potential for bias due to their exclusion from directory-listed sampling frames is likely to increase commensu-rately. Random-digit dialing (RDD) is an alternative sample design that can overcome this coverage limitation. However, although RDD sampling frames include both listed and unlisted households, the higher proportion of nonworking numbers in RDD sampling frames can reduce the efficiency of the [Page 932]sample as compared to that of a directory-listed sampling frame. It also may reduce the external validity of the sample, as respondents with an unlisted number are more prone to survey nonresponse due to both noncontact and refusal.