Quantitative Methods in the
Social Sciences 2

Survey Design and Quality

Seminar: Answer/Response Scales in Comparative Surveys

3 - 5 December, 2009, Leinsweiler, Germany

Introduction

General population surveys rely on closed answer formats using response scales of various kinds. Despite some methodological criticism, Likert scales are very often used for attitudinal questions, and frequency scales of various kinds used for behavioural self-reports. There may be considerable variation in the number of response options, in polarity, or in visual representation, etc. across surveys and also within a single questionnaire. However, in comparison to the endless number of questions asked in surveys, the forms and wordings of response scales are far more limited. The seminar aims to bring together review insights and research into answer scales as these are currently used in comparative research, both in terms of “source instrument design” and with regard to “target” versions produced in different languages.
Tradititionally, localisation of response scales from (usually) Anglo-Saxon cultures for other languages, cultures or nations has relied heavily on translation and adaptation. The research into version production procedures and their consequences is relatively sparse. Judging the effects of answer scale formats can be quite complex in that respondents’ use of response categories may be influenced by many factors including the substantive question, true values, response styles, and features of the question and answer event. Our understanding of interdependencies across these issues is only beginning to emerge. More in-depth research is needed.
The prominent role of response scales makes them a prime target for an advanced seminar within the QMSS 2 framework. Our 2009 winter seminar targets this topic. Researchers who have been involved in relevant research projects will discuss general and specific issues of response scale design and localization.

PROGRAMME

Final Report