Achim Koch

Improving Survey Response. Lessons Learned from the European Social Survey

1. Backgrounds of nonresponse 1.1. Introduction 1.2. Declining response rates 1.3. Total survey quality and nonresponse 1.4. Optimising comparability 2. Survey response in cross-national studies 2.1. Introduction 2.2. Harmonisation models 2.3. Contactability 2.4. Ability to cooperate 2.5. Willingness to cooperate 2.6. Nonresponse bias 2.7. Ethics and humans 3. The European Social Survey 3.1. Introduction 3.2. What is the European Social Survey? 3.3. ESS design and methodology 3.4. Nonresponse targets, strategies and documentation 3.5. Conclusions 4. Implementation of the European Social Survey 4.1. Introduction 4.2. Basic survey features 4.3. Practical fieldwork issues 4.4. Summary and conclusions 5. Response and nonresponse rates in the European SocialSurvey 5.1. Data and definitions 5.2. Response and nonresponse rates in ESS 3 5.3. Response rate differences and fieldwork efforts 6. Response enhancement through extended interviewer efforts 6.1. Introduction 6.2. Previous research on contactability 6.3. Previous research on cooperation 6.4. Sample type and recruitment mode in the European SocialSurvey 6.5. Establishing contact in the European Social Survey 6.6. Obtaining cooperation in the European Social Survey 6.7. Effects of enhanced field efforts in the European SocialSurvey 6.8. Conclusion 7. Refusal conversion 7.1. Introduction 7.2. Previous research 7.3. Refusal conversion in the ESS 7.4. Refusal conversion and data quality 7.5. Discussion and conclusions 8. Designs for detecting nonresponse bias and adjustment 8.1. What is nonresponse bias? 8.2. Methods for assessing nonresponse bias 8.4. Final conclusions 9. Lessons learned 9.1. Introduction 9.2. Standardisation, tailoring and control 9.3. Achieving high response rates 9.4. Refusal conversion 9.5. Nonresponse bias 9.6. Contact forms and fieldwork monitoring 9.7. Into the future

Respondent Incentives in a National Face-to-face Survey Do They Affect Response Quality?

Respondent incentives are a popular instrument to achieve higher response rates in surveys. However, the use of incentives is still a controversial topic in the methodological literature with regard to the possible reduction or increase in response quality. We conducted an experiment in a large-scale German face-to-face study in which the treatment group was promised a modest monetary incentive. We used different indicators of response quality and compared the incentivized group with the control group. Our results indicate that in general there are no systematic differences between the incentivized and the control group concerning response quality. We found some hints that specific subgroups react differently to incentives in terms of response behavior. While response quality usually tends to be lower for older respondents, we found that in the incentivized group the response quality is higher for older respondents as compared to younger ones regarding the level of item nonresponse.

Führt eine höhere Ausschöpfung zu anderen Umfrageergebnissen

Die Ausschopfungsquoten von Bevolkerungsumfragen sinken in vielen westlichen Landern (Atrostic et al. 2001; Brick und Williams 2013; de Leeuw und de Heer 2002; Dixon und Tucker 2010; National Research Council 2013; fur Deutschland siehe Reuband 2011; Schnell 1997). Ausschopfungsquoten unter 50 Prozent stellen bei Haushaltsbefragungen in Europa heutzutage keine seltenen Ausnahmefalle mehr dar (Kreuter 2013; Matsuo und Loosveldt 2013). In vielen Fallen haben Umfrageinstitute versucht, durch vermehrten Aufwand bei der Datenerhebung (z.B. durch haufigere Kontaktversuche oder den Einsatz von Befragtenincentives) den negativen Trend zu stoppen – zumeist nur mit masigem Erfolg. Als Konsequenz sind rucklaufige Ausschopfungen trotz steigender Umfragekosten zu verzeichnen.