Food cravings discriminate differentially between successful and unsuccessful dieters and non-dieters. Validation of the Food Cravings Questionnaires in German

Abstract

Food cravings have been strongly associated with triggering food consumption. However, definitions and measurements of food cravings are heterogeneous. Therefore, Cepeda-Benito et al., 2000, Cepeda-Benito et al., 2000 have suggested the Food Cravings Questionnaires (FCQs) to measure food cravings as a multidimensional construct at trait- and state-level. In the current study, we validated a German version of the FCQs in an online study (N = 616). The factor structure of the state and trait versions could partially be replicated, but yielded fewer than the originally proposed factors. Internal consistencies of both versions were very good (Cronbach’s α > .90), whereas retest reliability of the state version was expectedly lower than that of the trait version. Construct validity of the trait version (FCQ-T) was demonstrated by high correlations with related eating behavior questionnaires and low correlations with questionnaires unrelated to eating. Most importantly, FCQ-T-subscales were able to discriminate between successful and unsuccessful dieters and non-dieters. Validity of the state version was supported by positive relations with food deprivation and current negative affect. Taken together, the German version of the FCQs has good psychometric properties. Moreover, this study provided first evidence that distinct dimensions of food cravings are differentially related to success and failure in dieting.

Introduction

Craving is an intense desire or longing for particular substances (Weingarten & Elston, 1990), for example alcohol, tobacco, and other drugs, but also food (Hormes & Rozin, 2010). Food cravings are characterized by both appetitive and aversive components (Rodríguez, Fernández, Cepeda-Benito, & Vila, 2005). Accordingly, it has been suggested that food cravings can be viewed analogous to emotions as they have motivational significance for the organism (Shiffman, 2000). In this respect, craving and hunger are closely related and indeed show many similarities (Shiffman, 2000). However, cravings differ from hunger as they tend to be more intense and specific for the kind of food desired (Hill, 2007). Furthermore, although nutritional deprivation can increase food cravings, it is not necessary to elicit them. For instance, Pelchat and Schaefer (2000) could show that a monotonous diet – and therefore sensory rather than nutritional deprivation – is sufficient to stimulate food cravings in young adults. Moreover, psychological factors like external and emotional eating are stronger related to food cravings than dietary restraint or daily caloric intake (Hill, Weaver, & Blundell, 1991).

Attempts to measure craving objectively, e.g. based on physiological data, have been criticized for being unspecific (Shiffman, 2000). Until now, “subjective self-report seems the only viable assessment modality” (Shiffman, 2000, p. S172). The term craving is somewhat vague and often subjects are asked to indicate on a one-item rating scale how strong they crave or desire a specific object (Weingarten & Elston, 1990). Therefore, there is a need to assess craving as a multidimensional construct with standardized questionnaires instead of single questions. This is particularly important in non-English speaking countries because there is no simple equivalent expression for craving (Hormes & Rozin, 2010).

To assess craving multidimensional, different measures have been developed such as the Food Cravings Questionnaires (FCQs, including a state and a trait version; Cepeda-Benito et al., 2000, Cepeda-Benito et al., 2000), the Attitudes to Chocolate Questionnaire (ACQ; Benton, Greenfield, & Morgan, 1998), the Orientation towards Chocolate Questionnaire (OCQ; Cartwright & Stritzke, 2008), and the Food Craving Inventory (FCI; White, Whisenhunt, Williamson, Greenway, & Netemeyer, 2002). Each of these measures represents different approaches to the craving construct. Both the ACQ and OCQ are designed to measure cravings specifically related to chocolate and emphasize the relationship between craving and feelings of guilt (Benton et al., 1998) or the conflict between approach and avoidance inclinations during the experience of craving (ambivalence model; Cartwright & Stritzke, 2008). The FCI measures cravings related to different classes of food (high fats, sweets, carbohydrates/starches, fast-food fats; White et al., 2002). Therefore, all of these instruments assess habitual cravings related to specific kinds of food and are restricted to certain dimensions of food cravings. As opposed to these questionnaires, the FCQs were constructed to assess craving for a variety of foods, without confining them to certain categories or specific foods, e.g., chocolate. Furthermore, the FCQs cover behavioral, cognitive and physiological aspects of food cravings. Finally, the FCQs combine two versions that measure current and habitual food cravings. Therefore, the FCQs are the only currently available food craving instruments that (1) do not refer specifically to chocolate or similar, (2) assess food cravings on a multidimensional level, and (3) measure food cravings as trait and state. Moreover, there is evidence that the FCQs can be used easily as a measure for specific cravings, e.g. by replacing references to food with references to chocolate (Rodríguez et al., 2007).

The FCQs are arguably the most extensively validated food craving measures and are available in Dutch (Franken & Muris, 2005; modified version from Nijs, Franken, & Muris, 2007), English (Cepeda-Benito, Gleaves, Williams et al., 2000) and Spanish (Cepeda-Benito, Gleaves, Fernández et al., 2000). Excellent psychometric properties could be demonstrated for healthy participants (Cepeda-Benito et al., 2003, Cepeda-Benito et al., 2000, Cepeda-Benito et al., 2000) and patients with eating disorders (Moreno, Rodríguez, Fernández, Tamez, & Cepeda-Benito, 2008). However, the factor structure could only be partially replicated in a sample of overweight and obese persons (Vander Wal, Johnston, & Dhurandhar, 2007). The trait version of the FCQs (FCQ-T) has been positively associated with disinhibited eating behavior, habitual hunger ratings, eating disorder symptoms, sensitivity to reward, and body-mass-index (BMI) in healthy participants (Cepeda-Benito et al., 2000, Cepeda-Benito et al., 2003, Franken and Muris, 2005). Moreover, female participants had higher FCQ-T-scores than male participants (Cepeda-Benito et al., 2003). In patients with eating disorders, FCQ-T-subscales have been found to discriminate between anorexia and bulimia nervosa subtypes (Moreno, Warren, Rodríguez, Fernández, & Cepeda-Benito, 2009). However, only cravings that were associated with a lack of control over eating, preoccupation with food, negative affect, and guilty feelings were predictive of bulimic symptomatology (Moreno et al., 2008). Morbidly obese patients consistently reported higher food cravings than controls, except for cravings associated with positive reinforcement that result from eating (Abilés et al., 2010).

Based on these findings, we expected the FCQ-T to be associated with a loss of control in eating behavior as an indicator of convergent validity. Specifically, we expected food cravings to be highly correlated with self-reported binge eating, food addiction symptoms, low perceived self-regulatory success in dieting, and dieting strategies that have been previously connected to low dieting success or high eating-related psychopathology (rigid control; Shearin et al., 1994, Stewart et al., 2002, Timko and Perone, 2005, Westenhoefer, 1991, Westenhoefer et al., 1999).

Relevant, but not directly eating-related constructs were used as an indicator of divergent validity. Substance craving and impulsivity have been positively, but weakly associated with BMI (Meule et al., submitted for publication, Meule, A., Vögele, C., & Kübler, 2011). Furthermore, impulsivity was weakly correlated with food addiction symptoms (Meule, Vögele, & Kübler, in press). Accordingly, we expected small positive correlations between food craving and both, substance craving and impulsivity. Sensitivity to reward or punishment has been inconsistently linked to dysfunctional eating behavior (see Bijttebier, Beck, Claes, & Vandereycken, 2009 for a review). For instance, while sensitivity to reward was positively correlated with the FCQ-T (Franken & Muris, 2005), the Behavioral Inhibition System (BIS), but not the Behavioral Activation System (BAS), was positively correlated with food addiction symptoms (Gearhardt et al., 2009, Meule et al., in press). Therefore, we expected no or small positive correlations between food craving and BIS/BAS-reactivity.

We further expected an association between food craving and dietary restraint. However, evidence for such a relationship is ambiguous. For instance, Polivy and colleagues (2005) found restrained eaters – as measured with the Restraint Scale (RS) – to experience more food cravings than unrestrained eaters. Restrained eating as measured with the Dutch Eating Behavior Questionnaire (DEBQ) and the Three-Factor Eating Questionnaire (TFEQ), however, was not correlated with craving frequency (Hill et al., 1991). These contradictory findings could be due to the different questionnaires used, because the RS has been found to measure unsuccessful restrained eaters while the DEBQ and TFEQ identify successful restrained eaters (Heatherton et al., 1988, Williamson et al., 2007). Yet, another study using the RS did not find an association between restraint status and food cravings (Rodin, Mancuso, Granger, & Nelbach, 1991). Furthermore, using restraint scales is critical because the population of restrained eaters consists of unsuccessful and successful ones (Van Strien, 1997, Van Strien, 1999). Most recently, attempts have been made to distinguish more explicitly between successful and unsuccessful restrained eaters (Papies et al., 2008, Van Koningsbruggen et al., 2011, Van Koningsbruggen et al., 2011). In those studies, the subscale concern for dieting of the RS is used to identify dieters and non-dieters. Additionally, the Perceived Self-Regulatory Success in Dieting Scale (PSRS; Fishbach, Friedman, & Kruglanski, 2003) is used to evaluate successful or unsuccessful dietary restraint. We adapted this procedure to classify participants as non-dieters and successful and unsuccessful dieters and explored food cravings in these subpopulations. Specifically, we expected food cravings to be increased in unsuccessful dieters compared to successful ones who in turn were expected to experience more food cravings than non-dieters.

The state version of the FCQs (FCQ-S) has been found to be sensitive to meal consumption in normal-weight (Cepeda-Benito, Gleaves, Williams et al., 2000) and overweight (Vander Wal et al., 2005, Vander Wal et al., 2007) participants. State cravings decreased after breakfast and increased afterwards during the first 3 h (Cepeda-Benito et al., 2000, Vander Wal et al., 2007). Accordingly, length of food deprivation uniquely and exclusively predicted state cravings (Cepeda-Benito et al., 2003).

Thus, we expected the FCQ-S to be positively correlated with the hours that have elapsed since the last meal. Furthermore, we also predicted current cravings to be associated with less positive and more negative current affect because mood has been suggested as a possible antecedent to or consequence of craving (Hill et al., 1991).