State of Transportation Statistics - Box 4

State of Transportation Statistics - Box 4

Telephone Surveys: Who Gets Left Out?

Some important transportation data are collected through telephone surveys. However, telephone surveys run the risk of not accurately accounting for the transportation patterns and needs of lower income, minority, non-English speaking, and other segments of the population. A review of the U.S. Census Bureau's American Housing Survey for the United States: 1999 bears this out. While only 4.4 million (4.2 percent) of the 102.8 million U.S. households do not have telephones available, the distribution is not evenly spread across the population. Renters, for instance, make up only 33 percent of households but constitute 51 percent of households without telephones. Similar disparities exist for blacks and Hispanics who make up 13 percent and 9 percent, respectively, of the total households but 20 percent and 13 percent, respectively, of households without telephones. Additionally, while 14 percent of households have incomes below the poverty level, they comprise 25 percent of the homes without phones. The percentages are higher in central cities.

A slightly different issue exists for those who only have mobile phones. With the popularity of cellular telephones rising because of their convenience and declining cost, some households have opted to forgo having a regular (i.e., landline) telephone. Since mobile phone owners may incur charges for incoming calls, most telephone surveys exclude cellular numbers. This leaves another segment of the population unrepresented in surveys even though they have telephones.

Lastly, telephone surveys may require responses in English. The 1990 census data show that nearly 8 million, or a little more than 3 percent of the 230 million people in the United States over the age of 5, live in households where no person 14 or older speaks English well enough to respond to an English-only survey. Therefore, it is important that survey design take into consideration the impact of who might be missed and find ways to include these segments of the population either through alternative sampling or statistical methods.