Answering Machine Messages

Telephone answering machines are devices that automatically answer telephone calls and record messages left by callers when the party called is unable to answer. Within households such devices are often used as “virtual secretaries” to screen unwanted calls or to facilitate communication while away from home. The first automated answering machines became available in the late 1930s in Europe, and the first commercial answering machine was sold in the United States in 1960. It was not, however, until the advent of digital technology in the early 1980s that ownership of telephone answering machines became widespread. Ownership in the United States has increased significantly since then, with more than 70% of households owning a telephone answering machine in 2006. Compared with people who do not have answering machines, owners of these devices typically have higher levels of education and incomes and are more likely to live in households of two or more adults.

Increased ownership of telephone answering machines and their use to screen calls pose a threat to the representativeness of samples in telephone surveys, particularly those based on random-digit dialed designs. More than half of the people who own answering machines say that they or someone else in their household uses the device to screen incoming telephone calls on at least an occasional basis. Households that screen calls are likely to have high family incomes, to be located in suburban or urban areas, and to include young adults with high levels of education. Yet, despite the increased use of answering machines for call screening, many researchers found that households with answering machines can be reached by telephone for survey calls, albeit often after multiple attempts. Fewer than 5% of households appear to screen all of their telephone calls with an answering machine, and when reached, answering machine owners tend to be just as willing to complete surveys as are those without answering machines. Contact with households with answering machines tends to be most successful when calls are made on Saturdays before noon, on Sundays, or on weekdays after 6:00 p.m.

People are not uniform, however, in how they use telephone answering machines. People with on-the-go lifestyles tend to use telephone answering machines to stay in contact and facilitate communication. This [Page 29]finding led some researchers to hypothesize that scripted messages left on such devices may prepare the household for a later call or even encourage a prospective respondent to return the call free of charge to complete the interview. If successful, such an approach would help to reduce the level of nonresponse in telephone surveys.

However, empirical research on the effectiveness of leaving messages on answering machines to improve survey participation is mixed. For surveys that involve a list of sample members whose names are known, leaving messages can be effective at improving survey participation. Such messages appear to work best if the message is tailored to include the sample member's name. Several random-digit dialed telephone surveys conducted in the early 1990s also showed that leaving messages on telephone answering machines could significantly improve response rates by 3 to 4 percentage points. However, more recent studies conducted at the state and national levels using random-digit dialed sample designs found no difference in the contact or completion rates of households that were left a message and those that were not. The strategy does not appear effective for two reasons. First, the percentage of households with which this technique can be used is limited, since messages can be left only at households with answering machines that are set to receive messages. Although telephone answering machines are in more than 70% of households, not all of these machines are ready to receive messages every time a survey call is made. Second, only a small percentage of respondents within households hear the message and are positively influenced to participate in the survey. It may be that people in households with multiple adults or teenagers sort through and listen to telephone messages in much the same way they sort through mail: one person tends to sort and screen for the rest of the household. It is likely that one person (perhaps simply the first person home each day) will listen to all of the telephone messages and relay to others in the household what is deemed to be important information. Unsolicited calls from researchers are probably not at the top of that priority list. As a result, with the exception of the person who sorts the messages, probably few other adults in the household hear them.

...