The treatment of "don't know," "no opinion," or other nonsubstantive responses is a problem in many consumer research surveys. This paper looks at the problem in the context of 1980 presidential election opinion polls. During the campaign, a relatively large proportion of those interviewed said that they were undecided as to whom they planned to vote for. Discriminant analysis is used to allocate undecided voters to candidates. The method is validated by a postelection follow-up survey. Ian Fenwick is Associate Professor of Marketing at York University. Frederick Wiseman is the Patrick F. and Helen C. Walsh Research Professor in the College of Business Administration at Northeastern University. John F. Becker is President and James R. Heiman Vice-President of Becker Research Corporation, Boston. This research was supported in part by a grant from the Dean's Research Fund, College of Business Administration, Northeastern University. The authors express their appreciation to the referees for their helpful suggestions. An earlier version of this paper was presented at the annual conference of the Association of Consumer Research, October 1981. Public Opinion Quarterly Vol. 46:383-391 ? 1982 by the Trustees of Columbia University Published by Elsevier Science Publishing Co., Inc. 0033-362X/82/0046-383/$2.50 This content downloaded from 157.55.39.255 on Mon, 01 Aug 2016 05:45:06 UTC All use subject to http://about.jstor.org/terms 384 FENWICK, WISEMAN, BECKER, AND HEIMAN illustrated by applying the new technique to data collected by the Becker Institute from an October 1980 poll of registered voters in Massachusetts. Further evidence is provided by an empirical validation study using data obtained in a November 1980 postelection survey of a sample of previously undecided voters.