The aims of this study were to: describe dietary intakes of obese and nonobese middle-aged women using a validated food frequency questionnaire; to assess dietary restraint, disinhibition, and hunger by the three factor eating questionnaire (TFEQ) in obese and nonobese samples and determine which of the factors are independently associated with obesity; and to examine correlations between selected nutritional variables and the TFEQ factors. Subjects studied included 179 obese Swedish women (BMI > 32) and 147 nonobese population-based controls (BMI < 28). Age-adjusted mean energy intake was significantly higher in obese women (2730 +/- 78 vs. 2025 +/- 85 kcal, p < 0.0001). In absolute and relative terms, fat intake was higher and alcohol intake was lower in the obese subjects. Disinhibition was the strongest TFEQ factor independently differentiating the obese and nonobese states, i.e., after adjustment for restraint and hunger. Within the obese sample, strong associations were seen between energy intake and disinhibition (p = 0.0005) and hunger (p = 0.0004). The association between energy intake and restrained eating was negative and weaker (p = 0.04). No such associations were seen in nonobese women. Thus, using a dietary instrument that is valid and unbiased with respect to obesity, strong psychological correlates, possibly causal, of variability in energy intake were detected in middle-aged women with obesity. Disinhibition is associated with both obesity and high-energy intakes and is therefore an important factor to consider in the treatment of women with obesity.