Applying theory of planned behavior to fruit and vegetable consumption of young adolescents

The purpose of this research was to investigate how well Theory of Planned Behavior (TPB) predicted frequency of consumption of fruits and vegetables among seventh graders and to investigate whether gender or socioeconomic status (SES) had moderating effects on the relationships between the constructs in the model.

Two identical self-administered paper and pencil surveys were distributed 6 months apart. Path analyses of the TPB models by gender and SES were performed with AMOS software.

The study occurred in eight middle schools from school districts in the Minneapolis--St. Paul metropolitan area, where a minimum of 20% of the students qualified for free or reduced-price meals.

Of the invited seventh graders, 1406 (74%) completed data for the model constructs at baseline and for behavior at interim data collection.

We assessed frequency of consumption of fruits and vegetables and attitudes, subjective norms, barriers, and intentions related to this behavior. Three levels of SES were created based on questions about free lunch, parents' education, and work.

No significant direct effects of attitudes and subjective norms on behavior were found, but for barriers, both indirect effects through intentions and direct effects on behavior were significant. Seven percent of the variation in the frequency of fruit and vegetable consumption and 31% of the variation in intention to eat more fruits and vegetables were explained by the model. Gender appeared to have moderating effects on the relationships between attitudes and intention and between intentions and behavior in the model. There were no moderator effects of SES, but the larger dropout rates from the low-SES group could have caused the groups to become more homogeneous.

The data fit the model well; still, large proportions of the variance in the frequency of consumption of fruits and vegetables were unexplained. This may be related to the operationalization of the model constructs or to missing important factors in the model. Furthermore, there appeared to be strong relationships between subjective norms or barriers and intentions as well as gender differences in the strengths of the relationships in the model that may have implications for intervention methods or messages.