Telephone Penetration

The term telephone penetration refers to the number of households in a given survey area with one or more telephones. Traditionally, this has meant one or more landline or wired telephones, not including cell phones. A major challenge for drawing proper survey samples is ensuring that the sample represents a very high proportion of the population of interest. Traditionally, bias as a result of undercoverage in telephone surveys was credited to households without phones. As time has gone on, the rate of households without phones has been declining, leading to a decline of said bias. In its infancy, telephone interviewing was used only as a method of support for other interviewing techniques, such as face-to-face interviewing. However, by the 1970s the penetration of telephones in U.S. households had exceeded 90%, and this higher telephone penetration resulted in the evolution of telephone interviewing as it has become a centralized and exact data collection method, evolved further through the use of networked computers and computer-assisted telephone interviewing (CATI).

Telephone penetration in the United States has been on the rise ever since the invention of the telephone in 1861. The percentage of households in the United States with telephones increased from less than 40% in 1940, to above 95% as of 2008. Because of coverage issues, telephone surveys, while often regarded as cheaper than face-to-face interviews, lacked respectability for much of the early 20th century. In the first half of the 20th century, telephone ownership was a privilege for those who could afford it. Its use was not common, and ownership was limited to a select group of citizens. For this very reason, telephone directories and telephone surveys were limited to surveys of special populations.

Perhaps one of the most famous examples of the effect of undercoverage on the results of a survey came in 1936, when the Literary Digest incorrectly predicted a victory for presidential candidate Alf Landon. Although the survey was conducted by mail, its sample was built from telephone directory listings. In the end, Franklin D. Roosevelt handily defeated Landon, and the general consensus was that poor coverage by the telephone directories used to draw the survey's sample was the cause.

As the number of households equipped with telephones increased, other changes to the survey landscape were also occurring. The increasing costs of face-to-face interviewing and a growing resistance to face-to-face data collection led to the need for researchers to find an alternative that was both affordable and adequate in terms of coverage. Increasing telephone penetration, over 90% by 1970, made telephone surveys a more practical alternative to face-to-face interviewing for obtaining information from a representative sample. As telephone interviewing has become more centralized, interviewing has evolved further through the use of networked computers and CATI.

Although telephone penetration has increased enough over time to make telephone interviewing a viable option for survey research, issues surrounding the potential undercoverage of subgroups in the United States still exist. Research has shown that undercover-age exists in certain demographic subgroups (e.g., Native Americans in the Southwest). Because of this, there is still great concern that the exclusion of non-telephone households from research efforts can lead to the underrepresentation of specific subgroups.

...