Contactability

The ease or difficulty with which a sampled respondent can be contacted by a survey organization is referred to as her or his “contactability.” It can be expressed as a quantity (or “contact propensity”) and ranges from 0.0 to 1.0, with 0.0 meaning it is impossible to contact the respondent and 1.0 meaning it is certain that the respondent will be contacted.

Contactability will vary by the mode that is used to attempt to contact a respondent in order to recruit her or his cooperation and/or gather data. Contactability also will vary according to the effort a survey organization expends to reach the respondent and what days and times these contact attempts are tried.

For example, take the case of young adult males, ages 18 to 24 years, who are among the hardest of demographic groups for survey organizations to make contact with. The mode of contact that is used will affect the contactability of this cohort, as they are far less likely to be contacted via a traditional random-digit dialed (RDD) landline survey. If the telephone mode is used, then researchers trying to contact this cohort also need to sample cell phone numbers, as nearly one third of these adults in the United States were “cell phone only” in 2007 and their proportion is growing each year. If the mode of contact is the postal service (mail), [Page 139]this young adult male cohort also will have relatively lower contactability, as they are likely to move from address to address more than other demographic groups.

The number of days, which days of the week, and what times of day a survey organization uses its interviewers (telephone or in-person) to make contact with respondents also will affect the contactability of respondents. In the case of the young adult cohort, fielding the survey for only a few days (e.g. a weekend poll, Friday through Sunday) will greatly lower the contactability of this cohort, especially if no late evening hours are included.

In a telephone survey, contactability also will vary by whether or not the survey organization sends out some form of name identifier to be shown on caller ID or on the privacy manager devices that many households use to decide whether or not to answer their incoming calls. (Yet, even if the survey organization's name is displayed on such a device it will not help raise contactability unless the respondents know the name and think positively toward it.) Leaving a message on an answering machine when it is first encountered at a household is thought to aid contactability, assuming the message is a persuasive one, given that many household use these machines to screen their incoming calls.

Low levels of contactability within a sample will lead to higher nonresponse due to noncontact. Thus, it behooves researchers to think explicitly about cost-effective ways to increase the contactability of their sampled respondents.