Abstract
OBJECTIVE: To validate an adaptation of a short questionnaire measuring behaviour related to selecting low fat diets. The questionnaire was adapted for telephone use in a low income, low education population. DESIGN: The factorial structure of the 38 item adaptation was studied in a population based random sample of 1432 adults. Seven day test-retest reliability was measured in a convenience sample of 93 adults, and criterion related validity in measuring fat was assessed against a dietitian administered diet history in another convenience sample of 81 adults. SETTING: Adults aged 18-65 years living in low income, inner city neighbourhoods in Montreal, Canada. RESULTS: Principal components analysis identified five food factors: avoid fat, junk food, high fat traditional foods, low fat substitutes for high fat foods, and modification of meat to reduce fat. Two factors were similar to those of the original version. Internal consistency of the subscales ranged from 0.49-0.72. Test-retest reliability ranged from 0.72-0.90. Validation of the subscales against usual dietary intake indicated that the "junk food" factor, arising from questions added to the original questionnaire to reflect local dietary habits, was most closely related to fat intake (r = 0.48; p < 0.001). CONCLUSION: This telephone adaptation provides an inexpensive and valid method of measuring fat intake. However, these results suggest that adaptations of existing dietary instruments should be validated in the populations for which they are intended before they are used.