The online survey in its classic form is a frequently used data collection method. However, many aspects of questionnaire design, in particular the use of graphical, interactive and multimedia elements in online questionnaires, have not been clarified. The Chair of Empirical Social Research is conducting laboratory and field-experimental studies on these topics.
Digital voice assistants (Alexa, Siri, Google, etc.) are becoming ubiquitous. However, are they suitable to minimize problems of population surveys? And which population groups can be surveyed with digital voice assistants? Besides these questions, we are also determining how the interview should be initiated and which known and possibly new measurement errors occur with sensitive and non-sensitive questions as well as different question types. In a series of six laboratory studies, each with 20 to 200 students or people from the general population, the willingness to participate in an interview with a digital voice assistant and the response quality will be analyzed. A comparison will be made with face-to-face interview, a telephone interview and an online survey.
Since 2007, various secondary analyses on the data quality of old and very old people in standardized surveys have been conducted. Based on available data sets (Allbus, ESS and SOEP as well as SHARE), data quality indicators (item nonresponse and extremity bias, as well as degree of differentiation) are generated and correlated with measures of cognitive abilities of the elderly people. The basic assumption of the secondary analyses is that the declining cognitive abilities of old and very old people are factors causing the lower data quality in surveys.
Expired projects
Over the next few years, a methods laboratory will be set up at the Chair of Empirical Social Research, which will include both a CATI laboratory (call center for conducting telephone surveys) and an online panel. The CATI laboratory and the online panel will be used for content-related studies as well as for methodological research in cooperation with internal and external partners. There will also be equipment for experimental studies on data quality in surveys.
Since 1994, a cross-sectional survey of around 4,000 school students at general and vocational schools on the subject of “violence at schools” has been carried out every five years in the Free State of Bavaria. Based on self-reports on victimhood and perpetration, statements on the frequency of violence at schools and its explanation are derived. In recent years, multi-level analyses on context and composition effects have been carried out with increasing intensity. The next survey wave is planned for spring 2010.
Based on a DFG-funded project, around 230 nine- to fourteen-year-old children were filmed while conducting standardized oral personal interviews. On the basis of these video recordings and subsequent detailed behavioral coding, indications were analyzed that could explain the possibly limited data quality of responses from younger children. Following this study, experimental and secondary analytical studies are conducted with younger children.
As part of a project together with colleagues from Slovenia, Austria, Lithuania, Great Britain, Latvia and Greece – funded by the EU under the seventh framework program – a study is being carried out on the respective vocational education and training systems. In addition to country reports, a core task is to conduct a survey among secondary school students on their perception of the advantages and disadvantages of the various paths through the vocational education and training systems.
The project, which has been funded by the DFG since 2008 as part of the “Survey Methodology” priority program, is setting up a mobile phone panel of around 1,200 participants. With the help of this mobile phone panel, changes of usage habits of mobile phone communication technology are analyzed. Another focus is the identification of households that can only be contacted using a mobile phone (mobile onlys). A total of seven survey waves are planned over a period of four years.
