Quality & Quantity 38: 17–33, 2004.
© 2004 Kluwer Academic Publishers. Printed in the Netherlands.
Response Effects in Surveys on Children and
Adolescents: The Effect of Number of Response
Options, Negative Wording, and Neutral Mid-Point
and JOOP HOX
Catholic University Brabant
Abstract. Social researchers increasingly survey children and young adolescents. They are con-
vinced that information about perspectives, attitudes, and behaviors of children should be collected
from the children themselves. Methodological expertise on surveying children is still scarce, and
researchers rely on ad-hoc knowledge from ﬁelds such as child psychiatry and educational testing, or
on methodological knowledge on surveying adults. Regarding adults, empirical evidence shows that
respondent characteristics (cognitive abilities) as well as question characteristics (question difﬁculty)
affect response quality.
This study reports on a methodological survey experiment on the effect of negatively formu-
lated questions, the number of response options and offering a neutral midpoint as response option
question characteristics on the reliability of the responses, using children and young adolescents as
The study shows no effects of negatively formulated questions on the reliability measures,
although children respond consistently differently on negatively formulated questions than on pos-
itively formulated questions. Taking all results on the effects of number of response options and
offering a neutral midpoints on the different reliability measures into consideration; it would appear
that offering about four response options is optimal with children as respondents.
Key words: question characteristics, stability over time, internal consistency, response quality,
Surveying the general population involves a set of complex activities, although
procedures to increase general response quality are well documented (Biemer et
al., 1991; Groves, 1989; Lyberg et al., 1997). There is an increasing body of empir-
ical evidence that both respondent and question characteristics affect the reliability
of responses in surveys (Alwin & Krosnick, 1991; Krosnick, 1991; Krosnick &
Alwin, 1987; Krosnick & Fabrigar, 1997; Narayan & Krosnick, 1996; Schwarz &
Author for correspondence: University of Utrecht, Methodology & Statistics Department, P.O.
Box 80.140, NL-3580 TC Utrecht, the Netherlands. E-mail: firstname.lastname@example.org