Believing in food addiction: Helpful or counterproductive for eating behavior?

Abstract

Objective: Obesity is often attributed to an addiction to food, and many people believe themselves to be "food addicts." However, little is known about how such beliefs may affect dietary control and weight management. The current research examined the impact of experimentally manipulating participants' personal food addiction beliefs on eating behavior.

Methods: In two studies, female participants (study 1: N = 64; study 2: N = 90) completed food-related computerized tasks and were given bogus feedback on their performance which indicated that they had high, low, or average food addiction tendencies. Food intake was then assessed in an ad libitum taste test. Dietary concern and time taken to complete the taste test were recorded in study 2.

Results: In study 1, participants in the high-addiction condition consumed fewer calories than those in the low-addiction condition, F(1,60) = 7.61, P = 0.008, ηp (2) = 0.11. Study 2 replicated and extended this finding, showing that the effect of the high-addiction condition on food intake was mediated by increased dietary concern, which reduced the amount of time participants willingly spent exposed to the foods during the taste test, b = -0.06 (0.03), 95% confidence interval = -0.13 to -0.01.

Conclusions: Believing oneself to be a food addict is associated with short-term dietary restriction. The longer-term effects on weight management now warrant attention.

Figure 1

Mean calories consumed from chocolate and crisps as a function of condition. **Significant between‐condition difference at P < 0.01. Error bars represent standard error of the mean.

Figure 2

Mean calories consumed as a function of condition (high, low, or average addiction) and food type (chocolate and crisps). *Significant at P < 0.05. Error bars represent standard error of the mean.

Figure 3

Serial mediation analysis with high versus average condition comparison as the predictor variable, calorie intake as the outcome variable, and eating behavior concern and time‐taken as first and second mediators, respectively. Values are unstandardized regression coefficients (SEs) and associated P values. *Significant at P < 0.05, **Significant at P < 0.001. Bracketed association = direct effect after controlling for dietary concern and time‐taken.