A MEASURE OF STAGES OF CHANGE IN FRUIT AND VEGETABLE CONSUMPTION AMONG FOURTH-GRADE AND FIFTH-GRADE SCHOOL-CHILDREN - RELIABILITY AND VALIDITY

Objective: We developed, pilot-tested, and field-applied a stages of c hange questionnaire regarding fruit and vegetable (F&V) consumption am ong fourth- and fifth-grade students. Methods: The design included cro ss-sectional assessment of internal consistency and construct validity , and 2-week and 7-week longitudinal assessment of test-retest reliabi lity. Subjects included 134 students from one school for pilot-testing and 252 from two schools for field application. Thirty-two questions from McConnaughy et al's stages of change questionnaire were adapted a nd pilot-tested; minor revisions were made for field application. Stat istical analyses included principal components analysis to identify su bscales; cluster analysis to identify subgroups within students; Cronb ach's alpha coefficient to assess internal consistencies; Pearson prod uct-moment correlations to assess test-retest reliabilities; and onewa y ANOVA's by F&V stages of change clusters with actual F&V consumption , F&V self-efficacy subscales, F&V preferences, and F&V outcome expect ations subscales to determine construct validity. Results: Principal c omponents analysis from the field application indicated two subscales (precontemplation and beyond precontemplation) accounting for 39.5% of variance. Cluster analysis indicated 6 interpretable clusters; 2 (n = 63) provided responses inconsistent with the stages of change theory and 4 (n = 189) provided responses consistent with the theory. interna l consistencies and test-retest reliabilities were acceptable. Student s in the ''beyond precontemplation'' clusters had higher levels of sel f-efficacy and outcome expectations regarding eating F&V. Conclusions: Measuring stages of change other than the precontemplation stage in F &V consumption among elementary school children is problematic. Perhap s the theoretical concept does not apply to children, or elementary sc hool children lack the ability to comprehend the questions measuring t he concept, or the approach used was nor entirely appropriate.