RESUMO
PURPOSE: Test the applicability of the transtheoretical model (TTM) to adult fruit/vegetable consumption. DESIGN: Cross-sectional random-digit dial survey. SETTING: Hawaii. SUBJECTS: 700 (62.6% female; age [mean +/- SD], 47 +/- 17.1 years; education [mean +/- SD], 14.6 +/- 2.8 years; 35.0% white, 31.1% Asian, 22.1% native Hawaiian/Pacific Islander, 11.8% other). MEASURES: Stages, processes, self-efficacy, decisional balance, and self-reported fruit/vegetable consumption. ANALYSIS: Confirmatory factor analysis tested the factor structure. Analyses of variance were used to explore stage differences in constructs. RESULTS: Stage distribution was precontemplation (33%), contemplation (4%), preparation (37%), action (3%), and maintenance (23%). A 10-factor process model with two higher-order correlated factors (experiential and behavioral) provided the best data fit (chi2 = 1446.12; df = 366; p < .0001; comparative fit index [CFI] = .89; standardized root mean square residual [SRMR] = .05). The self-efficacy structure fit the data well (chi 2 = 81.86; df = 9; p < .0001; CFI = .94; SRMR = .04), as did the decisional balance structure (chi2 = 37.42; df = 19; p = .007; CFI = .99; SRMR = .02). Processes, self-efficacy, decisional balance, and fruit/vegetable consumption behavior differed significantly by stage, with medium effect sizes for most variables. CONCLUSION: The variables revealed adequate fit to the theorized measurement models. TTM predictions regarding stage differences in self-efficacy, pros and cons, and fruit/vegetable consumption were confirmed; however, most experiential and behavioral processes increased in the early stages and then leveled off.