Notice: the WebSM website has not been updated since the beginning of 2018.

Web Survey Bibliography

Title A literature review on constructing answer formats
Year 2009
Access date 12.08.2009

The apparently easy task of constructing answer formats poses many design decisions with consequences for the reliability and validity of requests. In accordance with Parduccis theoretical framework, the range of scale points and their frequency build a frame for the respondents’ understanding of what was asked and how the answer should be edited. The consequent research shows that answer scales influence respondents’ understanding regarding 1) the meaning of the concept to be measured and 2) their assumptions about the distribution of the associated behaviour or opinion in the population. Consequently, the task of designing answer scales is a complex one. While designing answer scales one should consider the properties of dimension to be measured, the appropriate scale characteristics, the modus of measurement and the characteristics of respondents. This presentation gives an overview of the literature addressing different questions related to the design of answer scales. Is an open ended or a closed answer format appropriate? What is the optimal number of scale points? What about the response order effect? How should labelling be applied? Are fully labelled scales preferable to numbered scales or end-labelled scales? Should the answer scale be polar or bipolar? Should a scale have a middle point and “no opinion” option? While regarding some of these aspects, e.g. numbers of scale points or order effects similar findings and design suggestions are available in the literature (e.g. maxim of reliability in the case of 5-7 scale points), other aspects are controversial. For example prefer some researchers an open ended question format while asking about the frequencies of certain behaviours since the answers are more exact and the context effect of answer scales is absent here. Others discuss the psychological problems associated with an open answer format, e.g. that the respondents round their answers off and build their own answer categories. Considering the presented review of literature the consequences for the design of answer scales and open research questions are summarized.

Access/Direct link

Conference Homepage (abstract)

Year of publication2009
Bibliographic typeConferences, workshops, tutorials, presentations
Full text availabilityFurther details