Journal of Official Statistics, Vol.17, No.3, 2001. pp. 351–367
The Use of Neutral Responses in Survey Questions: An Application of Multiple Correspondence Analysis
Jörg Blasius and Victor Thiessen
Abstract:This article explores whether neutral responses are substantive or nonsubstantive ones. Using data from the 1984 Canadian National Election Study, in the first of two examples, respondents were asked which political party would be best (worst) in dealing with issues such as controlling inflation. A large minority of respondents volunteered a neutral no difference' response. The second example concerns a set of Likert items often used as measures of "political trust and efficacy. Multiple correspondence analysis is used to help determine if espondents who frequently use neutral responses such as "neither agree nor disagree' or "no difference' do so to mask their nonopinionation or for substantive reasons. The results show that respondents using substantive answers differ from those using nonsubstantive answers such as "don't know' and "no opinion' and that both are different from those using neutral responses.
Keywords:"don't know'' and "neither-nor'' responses; political opinions; properties of Likert-type scales; multiple correspondence analysis.
Copyright © Statistics Sweden 1996-2018. Open AccessISSN 0282-423XCreated and Maintained by OKS Group