Human Food Rejections

Thesis submitted to the University of London for the degree of Doctor of Philosophy, July 1999.
[Passed October 1999]

Henry William Windle Potts

Department of Psychology

Institute of Psychiatry


This thesis covers human food rejections in non-clinical situations.

Section 1 considers the list heuristic: a list of foods, with individuals answering some sort of question(s) about each food. The number of foods which elicit a particular response gives a measure of eating behaviour, usually ‘pickiness’. Lists were used with a sample of 72 men and another of 116 women. Results concur with previous studies in showing an inverse relationship between pickiness and sensation-seeking; but differ as a relationship with anxiety is not apparent.

Section 2 uses the method of Rozin & Fallon (1980) to consider multiple facets of rejection behaviour. Past work suggests a division of rejection behaviour into ‘distaste’ / ‘disgust’ / ‘danger’. Results from 43 subjects replicate the tripartite division and suggest a further substructure among ‘distaste’ items.

Section 3 covers the nature of the aversion in Learned Food Aversions (LFAs) whereby a food followed by nausea is subsequently avoided. LFAs are important for our understanding of learning behaviours and have clinical implications as a cause of cancer anorexia and possibly other learned nausea phenomena, like anticipatory nausea and vomiting in chemotherapy. The orthodoxy is that LFAs produce a hedonic shift in palatability. Anecdotal reports suggest there is a learned nausea: the target food becomes nauseating.

Previous human experiments have shown that LFAs can be produced, but have not asked about the nature of the aversion. In two experiments, subjects tasted a novel food and then had nausea induced by motion sickness or underwent a control procedure. Their reaction to the target food was later assessed. The first experiment failed to produce any aversion. The second, involving 47 subjects, did and there was a significant change in ratings of how nauseating the target food was in those subjects who had been made nauseous previously, demonstrating a learned nausea.

For Alex Westcombe and Kate Wrightson, the class of ’92—and to our supervisor, Jane Wardle.

I would like to thank Flávio Kapczinski at the Institute of Psychiatry; Paula Dawson, Alistair Dunn, John Bithell and Francis Marriott at the Department of Statistics, University of Oxford; Mike Bradburn, Sharon Love and Doug Altman at the Centre for Statistics in Medicine; Jeremy Wyatt and Leigh Gibson at UCL; and Lindsay Keith, Ben Leapman,, Special K (especially Lance Parkin for reasons of reciprocity), my mother, the cats, Oscar and Eiffel, and others too numerous to mention.

Special thanks to Michael Gresty of the MRC Human Movement and Balance Unit, National Hospital, for his help and guidance during my final experimental work.

I gratefully acknowledge the financial support of the Medical Research Council.

Section 1 has been published in an abridged form as Potts & Wardle (1998). The second study in Section 3 (3.5.2) was presented at conference: Potts, Wardle & Gresty (1997). Small parts of Section 2 have also been presented at conference: 2.10.3 as Potts (1993) and 2.9 as Potts (1994). In addition, Potts (1995) also used the data collected in Section 2, but its content is not included in this thesis.

The data collected in Section 2 were used as example data in a thesis submitted in partial fulfillment of an MSc in Applied Statistics (Potts, 1996). Potts (1996) developed the statistical procedures subsequently used in 2.6.3 seq., although the analyses presented here are new to this work. Material from both theses has been used for a later presentation (Potts, 1998).