Garn AC, Webster EK. Bifactor structure and model reliability of the Test of Gross Motor Development - 3rd edition.
J Sci Med Sport 2020;
24:67-73. [PMID:
32919885 DOI:
10.1016/j.jsams.2020.08.009]
[Citation(s) in RCA: 6] [Impact Index Per Article: 1.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/16/2020] [Revised: 06/26/2020] [Accepted: 08/26/2020] [Indexed: 11/16/2022]
Abstract
OBJECTIVES
This study investigated the structure of the Test of Gross Motor Development - 3rd edition (TGMD-3). Specifically, we examine bifactor structure, which simultaneously models a fundamental motor skills (FMS) general factor and specific factors for locomotor skills and ball skills, compared to other models.
DESIGN
Cross-sectional design using the TGMD-3 normative sample.
METHODS
The sample (N = 862) of children (Mage = 6.51, SD = 2.23) was matched based on United States census data, ensuring appropriate percentages of demographic representation and disability status. Confirmatory factor analyses, exploratory structural equation modeling, model-based reliability estimates including coefficient omega hierarchical, and coefficient omega hierarchical subscale, explained common variance estimates, and relative parameter bias were examined.
RESULTS
Findings revealed bifactor structure produced a better model fit compared to both one-factor and two-factor models. Furthermore, model reliability estimates that parceled true score variance for the general FMS factor, locomotor skills factor, and ball skills factor yielded high internal consistency for FMS (.797) but not locomotor skills (.168) and ball skills (.216). Finally, explained common variance (.852-.879) and relative parameter bias (.018-.072) estimates identified the strength of the run, skip, slide, and dribble skills tests to represent the FMS general factor.
CONCLUSIONS
Our findings demonstrate the advantages of using bifactor structure to examine the TGMD-3 compared to one-factor and two-factor models. Additionally, these results provide further evidence that using the TGMD-3 to examine an overall FMS general factor may explain more variance in performance and provide a better picture for evaluating children's current FMS levels compared to subscales independently.
Collapse