Characteristics of Cell Phone Only, Listed, and Unlisted Telephone Households

Abstract

Cell phone only households have increased from about 4% in 2004 to over 14% in 2007. Since cell phone only households are not included in random digit dialing (RDD) samples, they represent a major source of noncoverage error. To assess the extent of this noncoverage error we compared the results of a CATI telephone survey of three different samples of households: (1) cell phone only households; (2) directory listed households; and (3) unlisted households. The survey results are based on samples of at least 400 respondents per group, and show significant differences in demographics between all three groups, as well as differences in attitudinal measures. The results indicate that including cell phone only households in a telephone samples can lead to substantial improvements in coverage. Cell phone only respondents tend to be younger, single, and male, and tend to rent rather than own their home, and have either less or more education than respondents from listed and unlisted RDD samples. Cell phone only respondents are also more likely to have more than one cell phone, and to have full time employment. A significant difference was also observed for political party affiliation. The cell phone only sample also picked up the highest rate of Asian respondents. We found significant differences in the characteristics of directory listed vs unlisted households as well. Listed households are the least likely to have a computer in the home, have fewer male respondents, have the highest proportion of retired, and elderly respondents, the highest proportion of white and married respondents. The paper discusses the implications of these results for obtaining representative samples for telephone surveys.