Quality & Quantity 37: 239–261, 2003.
© 2003 Kluwer Academic Publishers. Printed in the Netherlands.
PEDAKSI: Methodology for Collecting Data about
Institute for Social and Economic Research, University of Essex, Wivenhoe Park, Colchester CO4
3SQ, U.K. Tel.: +44 1206 874809; Fax: +44 1206 873151; E-mail: firstname.lastname@example.org
Abstract. The effects of unit non-response on survey errors are of great concern to researchers.
However, direct assessment of non-response bias in survey estimates is rarely possible. Attempts
are often made to adjust for the effects of non-response by weighting, but this usually relies on
the use of frame data or external population data, which are at best modestly correlated with the
survey variables. This paper reports the development of a method to collect limited survey data
from non-respondents to personal interview surveys and a large-scale ﬁeld test of the method on the
British Crime Survey (BCS). The method is shown to be acceptable and low cost, to provide valid
data, and to have no detrimental effect on the main survey. The use of the resultant data to estimate
non-response bias is illustrated and some substantive conclusions are drawn for the BCS.
Key words: Unit non-response, bias, validity, survey of non-respondents, British Crime Survey
This paper describes a methodology for the collection of key survey data from
non-respondents to personal interview surveys and subsequent assessment of non-
response bias. The methodology is known as Pre-Emptive Doorstep Administration
of Key Survey Items (PEDAKSI). The outcomes of a large-scale ﬁeld test of the
PEDAKSI methodology are reported and discussed.
We ﬁrst discuss the need for effective methods of assessing non-response bias
and then describe the PEDAKSI methodology. The ﬁeld test is then described and
its outcomes documented in terms of ease of implementation, cost, impact on the
main survey, and estimates of non-response bias. The ﬁnal section of the paper
draws some conclusions and outlines potential implications for survey design and
2. The Need for Non-Response Bias Assessment Methods
The control of total survey error is an important issue for any survey. Survey error
can be deﬁned simply as the (expected value of the) difference between a survey
estimate and the true value of the parameter being estimated. Even using this re-
strictive deﬁnition, survey error has many components, however (Groves, 1989).