Do Initial Respondents Differ From Callback Respondents? Lessons From a Mobile CATI Survey. (October 2017)

• Record Type: Journal Article
• Authors: Vicente, Paula; Marques, Catarina
• Is Part Of: Social science computer review. Volume 35:Number 5(2017:Oct.)
• Page Start: 606
• Page End: 618
• Publication Date: 2017-10
• Subjects: mobile phones -- CATI surveys -- callbacks -- two-level ordered logit model; Social sciences -- Data processing -- Periodicals; Computers -- Social aspects -- Periodicals; Microcomputers -- Periodicals; Sciences sociales -- Informatique -- Périodiques; Micro-ordinateurs -- Périodiques; 300.285
• DOI: 10.1177/0894439316655975
• Languages: English
• ISSNs: 0894-4393

Abstract

Mobile phones are increasingly being used to conduct mobile computer-assisted telephone interview (CATI) surveys and some say it is just a question of time before they replace fixed phones. Although mobile phones allow their users to be called anywhere and at any time, a 100% response rate (RR) is unlikely to be achieved with a single call attempt. Callbacks are costly and delay fieldwork completion, but they are worth the effort when better overall RRs as well as those for specific subgroups of the population are obtained. Using data collected on a nationwide mobile CATI survey, this article investigates the differences between initial and callback respondents. Findings reveal that although the majority of interviews were achieved with a single call attempt, the initial sample differed from the callback sample in terms of the age and residence of respondents. Additionally, callback respondents were more likely to be interviewed outside home.