Ahmet Bildiren, Sevinç Zeynep Kavruk
Internal consistency was high across the examined age, grade, and socioeconomic subgroups (ωt = 0.932; α = 0.925). Test-retest reliability (r = 0.847) indicated high rank-order stability across administrations despite a small systematic increase in mean scores, and alternate-form reliability was also high (r = 0.812). TONI-3 scores showed strong correlations with NNAT-I (r = 0.756) and RSPM (r = 0.810), supporting criterion-related validity. Confirmatory factor analysis provided partial support for approximate unidimensionality, with favorable CFI (0.983) and RMSEA (0.048), but an elevated SRMR (0.141) and relatively low loadings for introductory basal items. Scores differed significantly across age and grade, whereas the gender difference favored girls with a negligible effect size (d = 0.128). No significant differences were observed across socioeconomic development levels.
BACKGROUND: The aim of this study was to examine the psychometric properties of Form A of the Test of Nonverbal Intelligence-Third Edition (TONI-3) in Turkish middle school students aged 10 years 0 months to 14 years 0 months. Although TONI-3 has previously been validated in Turkish children and late adolescents, psychometric evidence for early adolescence remains limited.
METHOD: The sample consisted of 1,041 middle school students in Aydın province (50.6% female; M = 11.74, SD = 1.15), selected using stratified cluster sampling based on the Socio-Economic Development Level (SEGE). Internal consistency was assessed using McDonald's omega total (ωt) and Cronbach's alpha. Test-retest (n = 94), alternate-form reliability (n = 114), and criterion-related validity were evaluated using Pearson correlations with the Naglieri Nonverbal Ability Test-I (NNAT-I; n = 118) and Raven Standard Progressive Matrices (RSPM; n = 105). Group differences were examined using independent-samples t-tests and one-way ANOVA.
RESULTS: Internal consistency was high across the examined age, grade, and socioeconomic subgroups (ωt = 0.932; α = 0.925). Test-retest reliability (r = 0.847) indicated high rank-order stability across administrations despite a small systematic increase in mean scores, and alternate-form reliability was also high (r = 0.812). TONI-3 scores showed strong correlations with NNAT-I (r = 0.756) and RSPM (r = 0.810), supporting criterion-related validity. Confirmatory factor analysis provided partial support for approximate unidimensionality, with favorable CFI (0.983) and RMSEA (0.048), but an elevated SRMR (0.141) and relatively low loadings for introductory basal items. Scores differed significantly across age and grade, whereas the gender difference favored girls with a negligible effect size (d = 0.128). No significant differences were observed across socioeconomic development levels.
DISCUSSION: The findings provide complementary evidence supporting the reliability and validity of TONI-3 Form A. Although the structural analyses supported approximate unidimensionality rather than a strictly unidimensional structure, the combined evidence supports the use of TONI-3 Form A for assessing nonverbal intelligence in Turkish middle school students aged 10-14 years, within the scope of the present study. The findings also provide preliminary evidence of comparable item functioning across demographic groups, although formal measurement invariance remains to be established.