Assessing fat-related dietary behaviors among black women: reliability and validity of a new Food Habits Questionnaire.

Objective: To describe the development of the SisterTalk Food Habits Questionnaire (STFHQ).

Design: Formative research was conducted to adapt previous tools for the study's target population. A pilot tool (168 questions) was tested. The new 94-question tool was then used for evaluation of the SisterTalk project. Lastly, a 4-week reliability calibration study of the revised STFHQ was conducted in comparison with a food frequency questionnaire (FFQ).

Analysis: Reliability was assessed using test-retest correlations. Validity was assessed by correlations between STFHQ scores with FFQ calculated calories, total fat (g) and percentage of calories from fat. Three scoring methods (ie, introductory, product, and detail) were calculated along with inclusion or exclusion of dining out questions and alternate methods of scoring for food items not consumed.

Results: Reliability (correlation) was 0.87. Inclusion of dining out questions and imputation of zero for food items never consumed were more highly associated with fat intake than other scoring methods. The introductory score was most highly correlated with fat (g), whereas the product and detail scoring methods correlated highest with percentage of calories from fat. Responsiveness to the SisterTalk intervention was highest with the detail score.

Conclusions And Implications: The STFHQ is a reliable and valid tool that may be useful for evaluating dietary change for black women.