Comparing the results of the Test of Gross Motor Development-2 (TGMD-2) to the results of the Furtado-Gallagher Computerized Observational Movement Pattern Assessment System (FG-COMPASS)

Fundamental movement skills (FMS) are considered the building blocks for the development of specialized skills. In addition, fundamental movement skill competency has been linked to decreased levels of obesity and increased levels of physical activity/sport participation. Thus, teachers and practitioners working with younger children must conduct regular assessments to gather evidence about the student's level of achievement in FMS development. This study aimed to collect evidence for concurrent validity, reliability of classification decisions, and inter/intra rater reliability for the FG-COMPASS. The FG- COMPASS is a process-oriented and criterion-referenced instrument that evaluates FMS development in children ages 5 to 10 years. The Test of Gross Motor Development 2 (TGMD- 2) was used as "gold standard" for concurrent validity. Participants were 34 children between the ages of 5 and 10 years. Partial Pearson correlations (controlling for age) comparing the scores of both tests indicate a moderate to strong correlation for locomotion (rxy.z =.52, p .01), object manipulation (rxy.z =.59, p .001), and total scores (rxy.z =.63, p .001). The reliability of classification decisions was assessed by comparing the live score ratings of five raters with the primary investigator's video scores on 10 randomly selected children per skill. The inter-rater reliability was assessed by comparing the scores across the 5 FG-COMPASS raters. The intra- rater reliability was assessed by comparing each FG-COMPASS rater's live scores to their video scores. The weighted kappa scores ranged from .51 to .83, .50 to .89, and .60 to .87 for reliability of classification decisions, inter and intra rater reliability, respectively. The results of this study provide further validity and reliability evidence for the FG-COMPASS. Further studies involving children with different ethnicity backgrounds and a larger sample size is recommended.