Measuring Early Communication in Classroom Contexts

Poster Presentation
Friday, May 3, 2019: 5:30 PM-7:00 PM
Room: 710 (Palais des congres de Montreal)
N. Brady1, K. K. Fleming2, K. Muller3 and K. Matthews1, (1)University of Kansas, Lawrence, KS, (2)Life Span Institute, University of Kansas, Lawrence, KS, (3)The University of Kansas, Lawrence, KS

There are few standardized tools or strategies designed specifically to measure communication in students with autism who have minimal verbal skills in natural contexts (Kasari et al., 2013; Tager-Flusberg & Kasari, 2013), Standardized or norm referenced assessments often yield scores that are at the floor of norm-referenced assessments. New measures, like the Communication Complexity Scale (CCS; Brady et al., 2012; 2018), may help fill the void of appropriate assessments for this population. The CCS is a 12-point scale that measures pre-linguistic and early-linguistic communication in individuals with autism or intellectual and developmental disabilities who have minimal verbal skills. The CCS can be used to describe students' communication during a standardized play-based assessment and during naturalistic-classroom contexts. This poster presents CCS data from observations in naturally occurring interactions in classroom environments.


  1. Present student’s CCS scores obtained during classroom observations and standardized play-based assessments.
  2. Present CCS reliability results from classroom observations.



  • Six students with a diagnosis of autism spectrum disorder and expressive vocabularies of 30 or fewer words, signs, or symbols, between the ages of 5-16 years participated.


  • A 20-30-minute scripted play-based assessment was administered to each participant. The standardized play-based assessment consists of 12 activities. Research assistants viewed videos of the assessment and assigned a score to the highest communication act produced in each activity. Intentional communication acts were also assigned a function of behavior regulation (BR) or joint attention (JA)
  • Each participant was observed during four-to-six, 10-minute naturally occurring classroom contexts, using Noldus Pocket Observer and a time-sampling procedure. During the 12-minute observations, research assistants observed participants for 30-seconds and then had 15-seconds to code the highest communication act that occurred during the 30-second observation period. Function was also recorded for the highest communication act. In addition to BR and JA, we coded Response to Questions or Prompts. We also recorded if particpants’ were communicating with teachers or peers.
  • Inter-rater reliability during naturalistic classroom observations was computed by comparing scores and functions within each interval, and the average of the highest scores across intervals.
  • The following summary scores were derived:

1) Optimal CCS score = average of 3 highest observed communication acts (regardless of function)

2) Typical CCS score = the average of scores from the middle of the distribution

2) Highest BR = the highest communication observed with a BR function

3) Highest JA = the highest communication observed with a JA function


  • Reliability: The range of percent agreement between observers was 56.25 – 100% and the average was 77.08. The Optimal CCS scores from the two observers were all within a point of each other.
  • Participants’ communication scores varied by context. Most participants’ scores were highest in 1:1 interactions and the scripted communication samples.
  • For 5/6 participants, almost all communication was directed to adults as opposed to peers.
  • Communication to adults was more complex than communication to peers


  • Reliable coding was more difficult during academic instruction.
  • Information from the CCS can help identify communication goals and monitor progress.