OBJECTIVE: To evaluate the tracking of children's fruit and vegetable intake over 2 years. DESIGN: Children in the control group from the Gimme 5 nutrition education intervention trial completed three 7-day food diaries once a year in 1994, 1995, and 1996, beginning when they were in third grade. SUBJECTS: Five hundred sixty-one students from public elementary schools. STATISTICAL ANALYSES PERFORMED: Tracking was examined through correlation of fruit and vegetable servings over the 2 years as well as the percent of subjects remaining in the same or adjacent quintiles of fruit and vegetable intake, using the kappa statistic. RESULTS: Tracking of fruit and vegetable intake was fair to moderate over the 2 years. One-year correlations for total fruit and vegetable servings were .37 for boys and .46 for girls; 2-year correlations were .48 for boys and .40 for girls. kappa values for remaining in the same quintile generally exceeded chance rates only for the top and bottom quintiles. No values exceeded .50. APPLICATION/ CONCLUSION: The drift in ranking may have been a result of true changes in fruit and vegetable intake or measurement error (i.e., inability to accurately assess actual intake). If this drift reflects true change, it may be difficult for nutrition education programs targeting this age group to demonstrate long-term behavioral effects on fruit and vegetable intake. If the drift is the result of measurement limitations, improved dietary assessment may yield stronger tracking.
OBJECTIVE: To evaluate the tracking of children's fruit and vegetable intake over 2 years. DESIGN:Children in the control group from the Gimme 5 nutrition education intervention trial completed three 7-day food diaries once a year in 1994, 1995, and 1996, beginning when they were in third grade. SUBJECTS: Five hundred sixty-one students from public elementary schools. STATISTICAL ANALYSES PERFORMED: Tracking was examined through correlation of fruit and vegetable servings over the 2 years as well as the percent of subjects remaining in the same or adjacent quintiles of fruit and vegetable intake, using the kappa statistic. RESULTS: Tracking of fruit and vegetable intake was fair to moderate over the 2 years. One-year correlations for total fruit and vegetable servings were .37 for boys and .46 for girls; 2-year correlations were .48 for boys and .40 for girls. kappa values for remaining in the same quintile generally exceeded chance rates only for the top and bottom quintiles. No values exceeded .50. APPLICATION/ CONCLUSION: The drift in ranking may have been a result of true changes in fruit and vegetable intake or measurement error (i.e., inability to accurately assess actual intake). If this drift reflects true change, it may be difficult for nutrition education programs targeting this age group to demonstrate long-term behavioral effects on fruit and vegetable intake. If the drift is the result of measurement limitations, improved dietary assessment may yield stronger tracking.