Reverse Directory

A reverse directory has two general definitions. The first is a residential telephone directory, which has been converted from a surname alphabetical ordered listing to a street name ordered listing. The second is a listing of addresses in a city, also known as a city directory.

A telephone directory covering a city or other geographic area consists of residential listed telephone numbers. The directory is in alphabetical order by surname. A first name or initials generally accompany the surname. The street address is then listed, although in some areas a household may decide to not have their address listed in the telephone directory. The telephone number follows the address. Because telephone directories are key-entered into a database by commercial firms, it is possible to manipulate the order of the listings.

In a reverse directory, the telephone directory database for a city is sorted into an alphabetical street order and then by address number within street name. In some cases the commercial firm will add additional information such as zip code, census tract, and census block number. This type of reverse directory makes it possible to sample households from very small geographic areas for a telephone survey or an in-person interview survey. However, there are several sampling issues that must be taken into account. One key issue is that the reverse directory is a listing of residential addresses with directory-listed landline telephone numbers, and therefore households with unlisted telephone numbers and households without landline telephone service are excluded. These types of households are likely to be demographically different from households that are listed in a reverse directory, thus leading to possible coverage error in a survey that is measuring the residential population in a local area.

The selection of a sample from a residential telephone directory, which has been converted from a sur-named alphabetically ordered listing to a street name ordered listing, involves the random selection of address listings. This can be accomplished by determining the number of pages in the directory and first selecting a systematic random sample of pages. For each selected page, one or more address listings can then be selected also through systematic random sampling. This approach assumes that the pages in the reverse directory contain approximately the same number of address listings. The survey researcher must keep in mind that such samples exclude telephone households with unlisted telephone numbers.

City directories, on the other hand, provide a listing of residential addresses in a city from which a sample can be drawn. RL Polk is the original commercial firm that sold city directories for many of the cities in the United States. The street guide section of the typical city directory provides a listing of streets in alphabetical order and addresses in numerical order. City directories are sometimes used to select dwelling units for area probability sampling of cities.

Sampling from city directories has its origins in area probability sampling, as discussed by Leslie Kish. Kish described sampling procedures for selecting an element sample of address listing from a city directory. For a large city, the selection of an element sample of [Page 763]address listings does not take advantage of two-stage sampling methods developed to reduce data collection costs for in-person surveys. This approach involves first selecting clusters of addresses and, for each sample cluster, drawing a sample of address listings.

...