The article presents a confirmatory approach to the analysis of the Autism Scale developed by the authors earlier, for the rapid identification (screening) of the risk for autism (ASD) in 3-4 year-olds, with an accuracy of more than 85% (Nasledov et al., 2021; Nasledov et al., 2022). Forty binary points of the scale form 4 factors (subscales) by which children with ASD differ from children without ASD: "Emotional impairments", "Sensory disintegration", "Communication impairments", "Disinhibition". When developing the scale, we assumed that it is possible to distinguish "autism vectors" as a continuum, at one pole of which there is a conditional norm, and at the other, there is a vivid manifestation of the corresponding symptom of ASD. The purpose of this study was to verify the structural and measurement equivalence of the 4-factor scale model with respect to different parts of the sample: children with and without ASD, boys and girls, 3- and 4-year-old children. The sample consisted of 828 children, of which 294 were children with ASD, the rest were children without an established diagnosis (Norm) and children with developmental delay (DD). A "batch" approach has been applied, implying the combination (parcelling) of items included in one factor into several packages of items, which allows "strengthening" the initial measurements from binary to quantitative. The invariance of the measurement model was verified using multigroup confirmatory factor analysis. The factor structure is not equivalent for samples of children with and without ASD. However, for children with ASD, the structural and measurement equivalence of the 4-factor model for samples of boys and girls, and children aged 3 and 4 years is confirmed. Our initial assumption about the "vector" structure of ASD symptoms, which is also typical for children without ASD, with less severity of these symptoms, has not been confirmed. The presence of selected vectors was confirmed only for different parts of the sample of children with ASD.