Web Survey Bibliography
Title Some basic experiments on presentation of questions and response options in Web questionnaires
Author Lozar Manfreda, K., Koren, G., Hlebec, V.
Year 2004
Access date 14.09.2004
Abstract Survey methodologists already performed some tests on Web questionnaire design for efficient data collection. However, in comparison to the amount of research dealing with questionnaires in other survey modes these studies present only a first step in systematic research on the possible effects of the Web questionnaire on non-response and measurement errors.
Our study contributes to this body of knowledge on designing principles of Web questionnaires. We deal with some basic principles on how to present individual questions and response options. We focus on the following questions:
- What are the differences in substantive results and data quality when presenting an open question, a close question with a drop-down menu, and a close question with an ordinary list of responses?
- What is the order effect when response options are rotated and when not?
- How does the size of the text entry box influence responses to an open question?
- How does the presentation of a Likert scale set of questions in a grid table or separately (each question on its own screen) influence data?
- What is the effect of different types of scale response options: an 11- point scale with a middle point (-5 to 5 with 0) vs. a 10-point scale without a middle point (-5 to 5 without 0) vs. a 10-point scale divided to two scales with a filter question first?
- What are the differences in substantive results from ordinary attitude questions vs. questions with a filter question first excluding respondents with no attitude towards a certain matter?
Experiments addressing these issues were implemented in a Web survey among the first grade students at the Faculty of Social Sciences, Slovenia, in fall 2001. 350 respondents were randomly administered to different experimental conditions using split-ballot design. For each of the above research questions respondents were randomly assigned to two or three groups being administered a different version of the individual questions. We will show that some of the basic principles for designing self-administered questionnaires from paper-and-pencil self-administered modes hold true also for Web surveys. However, also specifics - for example the effect of introducing drop-down menus (which cannot be implemented within paper questionnaires) – can occur. We must regard the Web as a special medium with special design options, visual features and required respondent's actions. These all require a special treatment of the Web questionnaire.
Year of publication2004
Bibliographic typeConferences, workshops, tutorials, presentations
