Submission status
Reviewing
Submission Editor
Submission editor not assigned yet.
Title
An Independent Analysis of the RIOT IQ Test
Abstract
The Reasoning and Intelligence Online Test (RIOT) reports a Full Scale IQ (FSIQ) score and six Cattell-Horn-Carroll (CHC) based index scores, but over a year since its release, its yet to be validated by an independant party. This study examines RIOT’s scoring, norming, reliability, and factor structure using publicly available data. Exploratory and confirmatory factor analyses were conducted on the 16 subtest intercorrelation matrix (N = 417), with RIOT's published hierarchical six-factor model used as a point of comparison. EFA best supported a four-factor model of Verbal Reasoning, Processing Speed, Reaction Time, and a merged factor containing the former Fluid Reasoning, Spatial Ability, and Working Memory subtets. CFA confirmed that a four-factor model inspired by the EFA provided the best fit (CFI = .966, RMSEA = .048, SRMR = .037) with one fewer free parameter than the published six-factor model. RIOT measures a moderately strong general cognitive factor, with FSIQ loading .86 on g in the best supported model. However, Fluid Reasoning and Spatial Ability did not separate cleanly, and the three Working Memory subtests failed to form a coherent factor. Overall, the evidence supports greater confidence in RIOT’s FSIQ, Verbal Reasoning, Processing Speed, and Reaction Time scores than in its Fluid Reasoning, Spatial Ability, and especially Working Memory indices.
Keywords
intelligence,
reliability,
exploratory factor analysis,
psychometrics,
factor analysis,
Confirmatory factor analysis,
Cognitive assessment,
RIOT IQ Test,
IQ testing,
Construct validity
Preprint link
https://doi.org/10.5281/zenodo.21972420
