Laboratory measurement of cumulative food intake in humans: results on reliability.

The universal eating monitor (UEM) of Kissileff for laboratory measurement of food intake was modified and used with a newly developed special software to compute cumulative intake data. To explore the measurement precision of the UEM an investigation of test-retest-reliability of food intake parameters was conducted. The intake characteristics of 61 normal weight males and females were measured repeatedly in the laboratory with a measurement interval of 1 week. Yogurt of preferred flavor served as test meal. Test-retest-reliability of intake characteristics ranged from .64 (number of bites) to .83 (bite-size) and was .73 for initial eating rate, but only .16 for change of eating rate. Sex, BMI and eating habits according to TFEQ-factors had no significant effects on reliability of intake characteristics. The retest-reliability for all but one of the laboratory intake measures is comparable to standards of personality questionnaires and seems not to be influenced by sex or trait characteristics of eating behavior.