Unmatched Number

An unmatched telephone number is one that does not have a mailing address associated with it. Typically, it also does not have a name matched to it. The vast majority of unmatched telephone numbers are also unlisted telephone numbers. The proportion of unmatched numbers in a telephone survey that uses matching as a technique will depend upon the extent to which the survey organization uses multiple vendors to do their matching and the quality of the matching techniques and databases the vendor uses. However, in random-digit dialing surveys in the United States, regardless of how many matching vendors are used, a large minority of numbers can never be matched to an address because that information simply is not accessible to any matching company. This is due to the privacy and confidentiality concerns of the households whose numbers cannot be matched to an accurate address.

Whether a telephone number can be “matched” to an address is predictive of the likelihood that a completed interview will be attained with that household in a telephone survey. A greater proportion of interviews are completed with numbers that are matched than are completed with unmatched numbers. A primary reason for this is that those with unmatched numbers are less likely to react positively when they are contacted by a stranger (i.e. the interviewer). Another primary reason for this is that matched numbers have an address associated with them. As such, researchers can send advance mailings to these households when they are sampled for a telephone survey to alert them (“warm them up“) to the fact that an interviewer will be calling them. This cannot be done with unmatched numbers. Instead, calls to them are “cold calls,” and response rates for them consistently have been shown to be lower than with matched numbers.

On average, unmatched telephone numbers require more callbacks than do matched numbers in order for them to reach a proper final disposition. Thus, the calling rules used by a survey center to process unmatched numbers should differ from the rules used to process matched numbers. However, unless a survey center has their telephone samples screened for matched/unmatched status or receives this information for each number in their sample from their sample vendor, it will not be possible for them to take the matched/unmatched status into account as their computer-assisted telephone interviewing system processes the callback attempts.