As in all surveys, the costs of conducting surveys have risen for ALLBUS, while response rates have fallen at the same time. Compared to 1980, technical developments have also made computer-assisted and web-based surveys possible. For the ALLBUS, this means that the switch from interviewer-based surveys to self-administered surveys has become both necessary and possible.
After the ALLBUS had been conducted in face-to-face (F2F) mode since 1980 (first as personal interviews (PAPI), then as computer-assisted personal interviews (CAPI)), it took place for the first time in 2021 as a self-administered mixed-mode survey (PAPI and CAWI).
To accompany this change, an experiment was conducted on the study design (see graphic). In the simultaneous design, both survey modes were made available with the cover letter (link to CAWI, PAPI); in the sequential design, only the link to the CAWI survey was made available first and the paper questionnaire (PAPI) was only offered with the reminder letters. There were significant differences in the response rates between the two designs, with a higher response rate in the simultaneous design. Also, compared to the surveys in F2F mode, the composition of the respondents differs on various dimensions such as age, nationality and gender.
Until the results of this experiment will be published, you are welcome to contact Alexandra Asimov and Michael Blohm if you have any further questions.
Like other surveys in Germany, but also in many other countries, the ALLBUS survey has been confronted with the problem of declining participation rates in recent years and decades. In the recent past, the use of respondent incentives (i.e. incentives to participate in the study) has become established as a scientifically recognized means of countering this problem. For some years now, it has been standard practice in Germany to offer respondents a monetary incentive of 10 euros (sometimes more) for their participation in large-scale social science surveys. The handing out of this incentive is normally linked to participation in the survey. However, a large number of international studies have shown that the use of these so-called "promised" incentives, i.e. money for participation, is not necessarily the most efficient way to improve the willingness to participate. In various studies, so-called "prepaid" incentives, i.e. money with a request to participate, have proven to be superior. These are given to the people selected for the study in advance and are therefore not directly linked to participation in the survey.
In order to empirically test whether the use of prepaid incentives also represents an efficient and cost-effective measure for quality assurance in future ALLBUS surveys, an experiment was conducted in each of the 2010, 2014, 2016 and 2018 surveys in which the incentives were systematically varied in subgroups of the sample (for an overview, see chart on the right). In the 2014 survey, a factor of 1.4 higher take-up was achieved in the EUR 10 prepaid condition compared to the EUR 10 promised condition. In the 2010 survey, a group with €20 prepaid was included as a second comparison - the results of this group were very similar to those in the group with €10. Such an increase in response rate also goes hand in hand with cost savings. In order to realize the same number of interviews, an approximately 30% smaller initial sample is sufficient.
In order to test in a next step whether a similar increase in efficiency is possible if the prepaid amount is reduced and only 5 euros are sent in advance to the people in the sample, a further experiment was implemented in the 2016 survey. This showed that the increase in interviews with the EUR 5 prepaid condition was somewhat lower than with the EUR 10 prepaid condition, but the EUR 5 stake naturally also halved the incentive costs. The results of the experiment clearly indicate that the use of prepaid incentives is an efficient and cost-effective strategy for achieving the required number of interviews. Prepaid incentives lead to a considerable increase in willingness to participate and thus reduce the overall effort required to achieve a certain number of interviews.
To summarize: People who are sent money before they have decided to participate are more likely to participate than people who are offered money for their participation. This results in the surprising effect that a similar amount or even less money has to be spent on a survey overall, while at the same time increasing the quality measures. The results of these experiments have been published in the International Journal of Public Opinion Research.
Blohm, M., & Koch, A. (2021). Monetary incentives in large-scale face-to-face surveys: Evidence from a series of experiments. International Journal of Public Opinion Research, 33(3), 690-702. https://doi.org/10.1093/ijpor/edab007
