Quality & Quantity 32: 31–45, 1998.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.
The Impact of Respondent and Interviewer
Characteristics on the Number of
“No Opinion” Answers
A multilevel model for count data
JAN PICKERY & GEERT LOOSVELDT
Department of Sociology, University of Leuven, E. Van Evenstraat 2, B-3000 Leuven, Belgium
Abstract. In this article we analyse the number of “no opinion” answers to attitude items. We
argue that that number can be considered as a count variable that should be analysed using Poisson
regression or negative binomial regression. Since we’re interested in the effect of both respondent and
interviewer characteristics on the number of “no opinion”’s we use multilevel analysis that takes into
account the hierarchical structure of the data. As a consequence multilevel Poisson regression and
multilevel negative binomial regression are applied. Our analysis shows that answering “no opinion”
is related to some sociodemographic respondent characteristics. In addition we ﬁnd a signiﬁcant
interviewer effect, but we are not able to explain that effect in terms of interviewer variables.
Key words: multilevel analysis, interviewer effects, Poisson regression model, “no opinion” answer.
It is generally accepted that the use of a “don’t know” or a “no opinion” ﬁlter
increases the proportion of respondents who give this answer, and that the increase
itself is a function of the nature of the ﬁlter used (Schuman and Presser, 1981: 143).
Krosnick argues that answering “don’t know” is one form of satisﬁcing. Satisﬁcing
occurs when a respondent is not motivated to expend the mental effort necessary
to generate optimal answers. A “no opinion” answer is an acceptable answer but
it is the result of a ‘weak’ cognitive process. Satisﬁcing is a function of task
difﬁculty, respondent’s ability and motivation. This theoretical reasoning is con-
sistent with the ﬁnding that offering a “don’t know” response option increases the
proportion of respondents who select it, particulary among respondents with little
formal education and people who consider an issue to be less personally important.
(Krosnick, 1991). Following this argumentation, answering “don’t know” is mainly
explained by respondent characteristics that can be related to the cognitive aspect
of answering questions. Previous research points us to the following characteristics
of interest: age (see e.g. Groves, 1989: 441–443), education (e.g. Sudman and