Establishing Classification Criteria for an Energy Balance Knowledge Test for Fourth and Fifth Grade Children

Abstract

PURPOSE: Children show knowledge deficiency about energy balance, a key factor underlying body weight fluctuation. We previously developed and validated a written test, which has shown good validity and utility in capturing children’ knowledge about energy balance, but no criteria were established to classify scores for meaningful interpretation. The purpose of this study was to develop criteria to determine the knowledge levels. METHODS: Six Iowa schools participated in this study, with 570 children completed the pre-test and 587 completed the post-test. An obesity prevention program was implemented between the two assessments. Data were screened for outliers and examined for normal distribution. Cluster analysis was conducted to establish three levels of energy balance knowledge: high, moderate, and low. ANOVA was subsequently used to verify the classification. Thresholds for knowledge levels were determined based upon the maximum and minimum values at each level. Scores measured at pre- and post-tests were evaluated by referencing to the new criteria. RESULTS: The cluster analysis resulted in three knowledge levels: high (M: 75.3%, 66.7-96.3%; n=392), moderate (M=57.4%, 51.9-63.0%; n=436) and low (M=39.4%, 14.8-48.2%; n=328). Knowledge sum score was significantly different across the three groups (F(2,1153) =2700.14, p<0.001). Differences between every two levels were clear (high vs. low: p<0.001; high vs. moderate: p<0.001; moderate vs. low: p<0.001). The cut-points were 50.0% between low and moderate levels and 64.8% between moderate and high levels. The proportion of children in high, moderate and low levels were 20.9%, 41.9% and 37.2% respectively at pretest and 46.6%, 33.6% and 19.8% respectively at posttest. CONCLUSION: The knowledge classification was found to be empirically sound. The obesity prevention intervention showed preliminary efficacy in knowledge increase, with more children placed in the high knowledge group and less children in the low knowledge group. This study was supported by the Society of Health and Physical Educators, United States of Agriculture, and Iowa State University College of Human Sciences