Date of Award

Summer 2010

Degree Type


Degree Name

Doctor of Philosophy (PhD)

Graduate Group


First Advisor

Paul Rozin, Ph.D.


Food cravings are a common, yet poorly understood phenomenon. Past attempts to explain them with a focus purely on physiological mechanisms have been unsuccessful. Four studies examine the hypothesis that food cravings are best conceptualized in terms of socio-cultural and psychological factors, looking specifically at the example of perimenstrual chocolate craving. Study 1 demonstrates that the word “craving” does not lexicalize in a majority of foreign languages, calling into question the idea that craving is a universally relevant concept. Instead, the notion of craving appears fundamentally culture-bound. Study 2 characterizes chocolate cravers in the United States, and finds that women who link craving temporally to their menstrual cycle are unique in several attributes, most notably in significantly more disordered eating attitudes and behaviors. Differences between menstrual and other cravers and non-cravers hint at possible psychological characteristics associated with perimenstrual chocolate craving. Study 3 assesses the fate of perimenstrual chocolate craving in women post-menopause and finds that cravings remain prevalent, but in the absence of regular menstrual cycles are attributed in self-reports to stressors in the environment. This provides further evidence against a direct causal role of hormones in the etiology of perimenstrual chocolate craving, and gives rise to the hypothesis that cravings are a culturally sanctioned response to stress, and that menstruation is simply one highly salient stressor. Study 4 tests this stress-hypothesis by assessing the affective and physical correlates of chocolate craving in women diagnosed with premenstrual syndrome. Premenstrual chocolate craving, along with a range of affective symptoms, is shown to decrease significantly with the onset of menstruation, in the absence of significant changes in levels of hormones. It is furthermore reduced effectively with placebo treatment. This strongly suggests that perimenstrual chocolate craving is part of a cluster of affective symptoms that emerge prior to menstruation, probably in response to the subjective experience of stress. Results from these four studies constitute compelling new evidence against a physiological basis of food cravings. They provide the basis for a novel model of food cravings that emphasizes socio-cultural and psychological factors, including dietary restraint, ambivalence and culturally-promoted ways of coping with stress.