Modified COSMIC 15
3. We also combined rates of the M-COSMIC function of Initiation within certain
communicative forms (i.e., Language, Joint Attention), in order to permit more meaningful
comparison with relevant ADOS-G items which combine these forms and functions together.
4. We also compared the broad category M-COSMIC codes with the standardised
language/communication measures. MCDI receptive and expressive raw vocabulary counts, and
communication domain Standard Scores (SS) from the VABS and T-VABS showed good spread
within this sample, and as such, were employed as appropriate metrics for comparison.
Receptive, expressive, and total SS from the PLS were not considered useful, however, due to
lack of sensitivity for preschoolers with autism and very low verbal ability (i.e., with an artificial
floor at SS = 50) of 22 children in the current sample. By contrast, age-equivalence (AE) scores
showed good spread, and these were therefore employed for comparison of the M-COSMIC with
this measure.
4.6 Assessment of cross-contextual agreement
As the spontaneous communication behaviours of children with ASD were considered
likely to differ across activity settings, with different types of behaviour more or less easily
promoted by a structured teacher-led ACT vs. FP, comparison of the various M-COSMIC form,
function, role and partner codes across the ACT and FP settings was conducted. Arising
differences would provide information about the sampling context on child social
communication acts, validating the decision to sample more than one context systematically
across all children in modification of the original COSMIC protocol.
5. Results
5.1 Assessment of inter-rater reliability