Background: Clinical preventive guidelines recommend that health care providers counsel adolescents on nutrition. Brief, accurate, and reproducible dietary assessments are needed. The purpose of the current pair of studies was to develop a dietary fat screening measure for use with adolescents.
Methods: Two measures were developed-a 21-item and a 4-category measure. The measures differed in the level at which fat consumption was assessed (food item vs food group). Study 1 (N = 231, age M = 15 years, 57% female, 41% Euro-American) evaluated reliability. Study 2 (N = 59, age M = 14 years, 63% female, 37% Euro-American) evaluated construct validity and correct classification rates.
Results: Internal consistencies (alpha > 0.70) and test-retest reliabilities (ICC > 0.60) were adequate for both measures. Neither measure correlated with total fat assessed by a 3-day food record (P > 0.05). The 21-item measure correlated significantly with percentage of calories from fat (r = 0.36, P <.01). Correct classification rate (71%) and sensitivity (81%) of the 21-item measure were good. Specificity (47%) was lower, indicating some subjects with a low-fat diet were misclassified by the screening measure.
Conclusions: The 21-item measure is quick to complete and score, is inexpensive to reproduce, and has demonstrated reliability and validity. The measure could be clinically useful, but further improvements should be attempted to improve specificity.
Copyright 2001 American Health Foundation and Elsevier Science.