• Presentation

Experiments in Cell Phone Nonresponse

Citation

Peytchev, A. A., & Krotki, K. P. (2010, May). Experiments in Cell Phone Nonresponse. Presented at AAPOR 2010, .

Abstract

Inarguably, interactions over cell phones are different from those over landlines. There are multiple factors that can uniquely impact survey requests over cell phones, including (1) the expected absence of unsolicited calls and (2) the financial responsibility of incoming calls. Since only the phone number is available in cell phone random-digit-dialing studies, limiting the channels of communication with sample members (e.g., no prenotification letters and no alternate landline numbers), practitioners have little control over these factors.

We carried out two experiments to reduce the impact of these two factors on nonresponse. Providing study information in voicemail messages can supply sample members with information about the reason for the call and allow them to expect the call – potentially decreasing nonresponse. It may, however, provide the opportunity to make a nonparticipation decision prior to talking with the interviewer, thereby increasing nonresponse. A random half of the sample was assigned to receive voicemail messages on the first call attempt if contact was not achieved.

Yet even when contacted, sample members may refuse in part due to costs incurred for “airtime.” To alleviate this concern, sample members could be given the option to be called on another number. Any reduction in nonresponse, however, may be lost and nonresponse even increased if potential respondents are difficult to contact or refuse when called on the alternate number. In this second experiment, sample members were randomly assigned to receive an option to be called on another number.

The study was fielded in June and July, 2009, on national samples of cell phone numbers from MSG and from SSI. Preliminary findings show weak indication of favorable effect of leaving voicemail messages with the reason for the call and a strong indication of adverse effect of providing an option to be called on another number.