Modified COSMIC 13
child with ASD required a toilet break, at which time filming was paused and resumed upon the
child’s return.
Behaviour coding
Coding was conducted from videotapes using the modified coding scheme (presented in
Table 2). (For orthogonal coding categories, the full record form, and operational definitions of
behaviours, see Appendix). Prior to analysis, event recorded behaviours were transformed into
rates-per-minute to allow for slight variations in the duration of filming for each child.
- Insert Table 2 -
4.5 Assessment of concurrent validity
Concurrent validity of the M-COSMIC was investigated by comparing scores on the
broad M-COSMIC coding categories for form, function, and role with the Communication and
Social Interaction algorithm scores on the ADOS-G. Positive associations were also predicted
between the following specific codes and items (or item combinations):
1. M-COSMIC rates of Initiated Vocalisation/Single words and Phrase Speech and ADOS-G
item Overall Level of Non-Echoed Language;
2. M-COSMIC rates of Gesture/Point and ADOS-G items Pointing and Gestures;
3. M-COSMIC rates of Eye Contact and ADOS-G item Eye Contact;
4. M-COSMIC rates of Follows Gaze/Point and ADOS-G item Response to Joint Attention;
5. M-COSMIC rates of Show/Give and ADOS-G item Showing5;
6. M-COSMIC rates of Initiating Joint Attention and ADOS-G item Spontaneous Initiation of
Joint Attention; and
7. M-COSMIC rates of Request Object/Action and ADOS-G item Requesting (for the subsample
of children for whom Module 1 had been administered; n = 30).