OBJECTIVE: To establish reliability and construct and discriminant validity of selected factors adapted from the Health and Taste Attitudes Questionnaire developed in Finland and the Food Choice Questionnaire developed in England for use with low-income mothers in the United States. DESIGN: Cross-sectional. SETTING: Head Start and the Special Supplemental Nutrition Program for Women, Infants and Children sites in south-central Wisconsin. PARTICIPANTS: A convenience sample of 211 nonpregnant non-Hispanic black and white women. VARIABLES MEASURED: Predisposing, enabling, and reinforcing constructs of the PRECEDE-PROCEED model. ANALYSIS: Factor analyses. RESULTS: After modification, the resulting measurement models of predisposing, enabling, and reinforcing constructs demonstrated good reliability and construct and discriminant validity. These models showed excellent model fit (chi-square/df range 1.5-2.0, root mean square error of approximate range.05-.07, non-normed fit index range.97-1.0, comparative fit index range.98-1.0, incremental fit index range.98-1.0), with all factor loadings significant (P <.001). Items previously developed with European adults demonstrated different pattern structures in predisposing and enabling constructs but the same structures for reinforcing construct when applied to low-income mothers in the United States. CONCLUSION AND IMPLICATION: This modified instrument provides a tool for researchers and educators to further explore factors affecting fat intake behaviors in low-income mothers.
OBJECTIVE: To establish reliability and construct and discriminant validity of selected factors adapted from the Health and Taste Attitudes Questionnaire developed in Finland and the Food Choice Questionnaire developed in England for use with low-income mothers in the United States. DESIGN: Cross-sectional. SETTING: Head Start and the Special Supplemental Nutrition Program for Women, Infants and Children sites in south-central Wisconsin. PARTICIPANTS: A convenience sample of 211 nonpregnant non-Hispanic black and white women. VARIABLES MEASURED: Predisposing, enabling, and reinforcing constructs of the PRECEDE-PROCEED model. ANALYSIS: Factor analyses. RESULTS: After modification, the resulting measurement models of predisposing, enabling, and reinforcing constructs demonstrated good reliability and construct and discriminant validity. These models showed excellent model fit (chi-square/df range 1.5-2.0, root mean square error of approximate range.05-.07, non-normed fit index range.97-1.0, comparative fit index range.98-1.0, incremental fit index range.98-1.0), with all factor loadings significant (P <.001). Items previously developed with European adults demonstrated different pattern structures in predisposing and enabling constructs but the same structures for reinforcing construct when applied to low-income mothers in the United States. CONCLUSION AND IMPLICATION: This modified instrument provides a tool for researchers and educators to further explore factors affecting fat intake behaviors in low-income mothers.