To date, research about mobile usage to conduct surveys has focused on a comparative analysis with fixed phones (e.g. Keeter et al 2007, Vicente et al 2009, Lynn and Kaminska 2010). Research focused exclusively on mobile phones, specifically in the evaluation of its strengths and weakness as a survey mode is still scarce. 

Non-response is one major problem for surveys’ activity. Mobile phone surveys are no exception to this situation. If nonresponse is unaddressed, the resulting damage to data quality may have serious consequences for data analyses underpinning social science research. Mobile phones have specific characteristics that other modes don’t have – they are of personal use, carried at all time to every places. This specificity may affect the likelihood of getting a successful contact when soliciting people to participate. On one side, the time period for contact is large (theoretically, all day) which can improve the likelihood of contact, on the other side, it is easy to reject a call coming from an unknown number. If the proportion of non-responses is high and/or non-respondents are much different from respondents surveys estimates are subject to non-response error. 

Paradata are data collected during the survey data collection process (Couper 1998). Examples of paradata are call records, interviewer observations, keystroke data. Paradata can also be obtained by means of questions included on the questionnaire (e.g. the location of the respondent in the time of the interview, whether he/she is alone or accompanied by someone else). Paradata is most used to study unit nonresponse (Steeh et al 2001; De Leeuw and De Heer 2002). Nonresponse research is often limited by the small amount of information available for all sample units. However, call records can guide responsive survey design decisions aimed at reducing nonresponse rates and bias (Groves et al 2009), as they provide important clues on when and how best to contact and to achieve cooperation from sample units. Call record data also expand possibilities for post-survey nonresponse adjustment (Kreuter and Kohler 2009; Kreuter et al 2010), to the extent that these data are predictive of nonresponse and survey variables of interest. Paradata collection in the context of mobile phone surest is essential on the research for causes of non-response and the effects of non-response on survey results. Understanding the process that leads to non-response and its effects is indispensable to develop efficient strategies of contact with respondents and to the development of weighting procedures aimed at reducing the effect of non-response on survey estimates.  

This project has three subprojects: (1) Identify the most relevant correlates of non-response in mobile phone surveys among a set of field process and sample unit features, (2) measure the impact of non-response on survey results – to what extent are respondents and non-respondents different and how does non-response affect the quality of mobile phone survey statistics and (3) define data adjustments’ procedures in order to reduce non-response effects on survey estimates. 

Since no database containing data collected via mobile phone and respective paradata is available, data for this project will be collected by setting an operation of data collection. A mobile phone survey shall be conducted by a survey research company. Data shall be collected by means of structured questionnaire administered via CATI facilities using mobile phones.  

The benefits of this research are wide ranging and ultimately affect a wide range of disciplines, such as marketing research, social research, statistics and survey methodology. Research in this area will not only raise further awareness of opportunities and benefits of using mobile phones to conduct surveys, it will also help to identify the potential for improving cost-efficiency and the quality of survey data and stimulate the debate about methodological challenges for the use of mobile phones as a survey mode. Finally, by investigation the use of paradata to understand the problem of non-response we hope that activities in this emerging field will encourage data collection agencies to make paradata more readily available for survey researchers and analysts. The research will inform which type of paradata is useful and should be collected in the future mobile surveys and will contribute to the standardization of such data across telephone surveys in general. The project will in significant ways contribute to the development of a more systematic methodology for the collection and use of paradata aimed at understanding the causes of non-response and adjustment mechanisms.  

No records found.

No records found.

No records found.

No records found.

No records found.