"I Haven't Got a Clue!?" Do Clueless Respondents Affect Data Quality through Response Behaviour?
Sprache des Vortragstitels:
Facing an Unequal World: Challenges for Global Sociology, IVIII ISA World Congress of Sociology Japan 2014
Sprache des Tagungstitel:
Several papers are dealing with the effect of different numbers of response categories (e.g. Preston & Colman: 2000; Lai et al: 2010). However, irrespective of the number of categories it can be assumed that respondents are facing difficulties in answering questions if they do not already have an opinion for the topic. If a "don?t-know-option" is offered the participant might tend to choose it even if it were possible for him or her to form an opinion. This is "making survey researchers reluctant to offer this option unless absolutely necessary" (DeRouvray & Couper: 2002).
However, if the "DK-option" is missing, respondents may skip the question and increase items missing. Offering a middle category in rating scales may even enlarge biases if perceived as "neither-nor-option". In this connection personal characteristics and the type of question (e.g. opinion, attitude or behaviour) are expected to influence corresponding ad hoc decisions.
According to these consideration, the paper
i) focuses on the effect of different numbers of categories in combination with "don?t-know-options" on missing values,
ii) takes several question types into account and
iii) considers effects of personal characteristics on scale quality.
The results are based on methodical experiments included in three online surveys in Austria and Germany.
DeRouvray, C.; Couper, M. P. (2002): Designing a Strategy for Reducing ''No Opinion'' Responses in Web-Based Surveys. Sage, http://ssc.sagepub.com/content/20/1/3. Download 2013-09-11.
Lai, M; Li Y.; Liu Y. (2010): Determining the optimal scale width for a rating scale using integrated discrimination function. Elsevier, http://www.sciencedirect.com/science/article/pii/S0263224110001867. Download 2013-09-11.
Preston, C., C.; Colman, A. M. (2000): Optimal number of response categories in rating scales: reliability, validity, discriminating power, and respondent preferences. Acta Psychologica, 104. S. 1-15.