Psychometric Properties of the Test of Gross Motor Development-3 for Children With Visual Impairments

Abstract

Results of the Test of Gross Motor Development-2 (TGMD-2) consistently show acceptable validity and reliability for children/adolescents who are sighted and those who have visual impairments. Results of the Test of Gross Motor Development-3 (TGMD-3) are often valid and reliable for children who are sighted, but its psychometric properties are unknown for children with visual impairments. Participants (N = 66; M = 12.93, SD = 2.40) with visual impairments completed the TGMD-2 and TGMD-3. The TGMD-3 results from this sample revealed high internal consistency (ω = .89-.95), strong interrater reliability (ICC = .91-.92), convergence with the TGMD-2 (r = .96), and good model fit, χ(63) = 80.10, p = .072, χ/df ratio = 1.27, RMSEA = .06, CFI = .97. Researchers and practitioners can use the TGMD-3 to assess the motor skill performance for children/adolescents with visual impairments and most likely produce results that are valid and reliable.

The validity and reliability of observational assessment tools available to measure fundamental movement skills in school-age children: A systematic review

“…Strong test-retest reliability was evidenced with a CVI of .88 [ 141 ] and Bland Altmann plots found 95% confidence intervals were within one standard deviation [ 77 ], with .96 agreement ratio [ 146 ]. Evidence for internal consistency was more mixed, but there was strong evidence that all items in the TGMD-3, once modified for children with ASD and visual impairments could still measure FMS as an overarching construct [ 56 , 67 ]. Evidence for good internal consistency of the TGMD was also found when testing children with intellectual deficits [ 59 ].…”

“…Results revealed that children with ASD perform better on the TGMD-3 with visual aids compared to the standard assessments [56]. Similarly, modified versions of the TGMD-2 and TGMD-3 were both found to be valid for use in children with visual deficits [67]. Additionally, one study showed significant differences between subtest scores on the second and third editions of the TGMD across year groups and gender, in which participants performed better on the TGMD-2 [69].…”

“…Two studies included participants with Autistic Spectrum Disorder [56,57], and another study recruited children from special educational needs (SEN) schools [58]. Eight defined themselves as sampling children with learning and/or attentional problems [54,[59][60][61][62][63][64][65], three studies recruited children with visual impairments [66][67][68], and the sample of one study included children with a disability or chronic health condition [69]. Information regarding socioeconomic status (SES) was included in one article which stated they sampled from low SES [70], while two studies recruited samples from indigenous populations (in Australia and Canada, respectively) [44,71], the latter of which focused on the recruitment of children whose mothers drank alcohol during pregnancy [71].…”

“…Brian et al [67] Estevan et al [149] Maeng et al [150] Magistro et al [151] Rintala et al [140] Valentini et al [85] Wagner et al [144] Webster & Ulrich [145] NB https://doi.org/10.1371/journal.pone.0237919.t006…”

“…147], some models were hierarchical in nature [77] and in one case a two factor model, whilst best fit, explained only 50% of the total variance [142]. Evidence was however found to suggest that the structural validity of the TGMD is stable across countries, with the data from populations in Greece, Brazil, Germany, the USA, South Korea and Portugal all evidencing a two factor model [67,82,143,144,146,152].…”