Quality & Quantity 32: 77–91, 1998.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.
Handling Missing Data by Re-approaching
, BOUDIEN KROL
& ERIC VAN SONDEREN
Department of Statistics, Measurement Theory & Information Technology, University of
Groningen, Grote Kruisstraat 2/1, 9712 TS, Groningen, The Netherlands;
Northern Center for
Healthcare Research, University of Groningen, The Netherlands
Abstract. When handling missing data, a researcher should be aware of the mechanism underly-
ing the missingness. In the presence of non-randomly missing data, a model of the missing data
mechanism should be included in the analyses to prevent the analyses based on the data from
becoming biased. Modeling the missing data mechanism, however, is a difﬁcult task. One way in
which knowledge about the missing data mechanism may be obtained is by collecting additional
data from non-respondents. In this paper the method of re-approaching respondents who did not
answer all questions of a questionnaire is described. New answers were obtained from a sample of
these non-respondents and the reason(s) for skipping questions was (were) probed for. The additional
data resulted in a larger sample and was used to investigate the differences between respondents and
non-respondents, whereas probing for the causes of missingness resulted in more knowledge about
the nature of the missing data patterns.
Key words: missing data, follow-up, cause of missingness, scale data.
Missing data is a problem a researcher is often confronted with. There are fre-
quently persons not answering all questions in a questionnaire and the resulting
item non-response can cause serious problems (Little & Schenker, 1995). The
procedures to treat the missing data can be grouped into three categories (Little &
Rubin, 1987): (1) weighing procedures, (2) imputation procedures, and (3) direct
analysis of the incomplete data. An important feature determining the success of
a treatment procedure is the mechanism underlying the missing data. When data
are non-randomly missing, analyses may be seriously biased due to differences
between respondents and non-respondents, and the missing data mechanism should
be modeled in the analysis.
Insight into the missing data mechanism requires knowledge about which fac-
tors contribute to not answering questions. When the cause of missingness is known
the missing data mechanism is called accessible, and when included properly in the
analysis it will cause no bias even if the data is non-randomly missing (Graham &
Donaldson, 1993). Inaccessible, non-ignorable missing data mechanisms, however,