Examining the Relation Between ESCS and Pre-Intervention Baseline Levels of Joint Attention

Background: Joint attention is a core deficit in children with Autism Spectrum Disorders (ASDs) (DSM-IV-TR; APA, 2000; Mundy, 2003). Joint attention is often quantified using structured assessment procedures that incorporate specific activities and prompts to elicit behaviors of interest (Roos et al. 2008). Existing research has almost exclusively focused on the use of the Early Social Communication Scales (ESCS; Mundy et al., 2003) to measure joint attention in children with autism; few research studies have compared rates of joint attention across different experimental contexts. The ability to measure joint attention behaviors in more than one context, particularly naturalistic contexts, may allow researchers to obtain more representative samples of behaviors (Roos et al., 2008).

Objectives: The aim of the present study is to determine the extent to which joint attention behaviors on ESCS are related to joint attention in pre-intervention baseline assessment conditions. It is hypothesized that there will be a positive correlation between joint attention on the ESCS and the quantity of joint attention during pre-intervention baseline assessments. 

Methods: The present study compared scores on the Early Social Communication Scales (ESCS) to those on pre-intervention baseline assessments.  Ten (n=10) 2 to 5 year old children diagnosed with ASD were assessed using the Early Social Communication Scales (ESCS) to measure joint attention behaviors. Joint attention was then assessed in a structured pre-baseline assessment under 2 different conditions, lasting15 minutes each.  Both baseline conditions were naturalistic contexts, where the child and examiner interacted with a variety of toy on the floor.  Similar to the ESCS, during the first baseline condition (baseline1), the examiner did not initiate interactions, but responded appropriately to all child initiations.  During the second baseline (baseline2) condition, the examiner attempted to engage the child in play, initiating interactions as well as responding to child initiations.

Results: Contrary to expectations, joint attention (JA) behaviors on the ESCS were not significantly correlated with either Baseline1 (r=.023, p=.949) or Baseline2 (r= -.009, p=.980) JA acts.  Baseline1 and Baseline2 JA acts were also not significantly correlated with one another (r=.029, p=.936).  Mean JA acts across context were lowest on the ESCS (M=5.36, SD=6.12), followed by Baseline1 (M= 6.60, SD= 7.21), and Baseline2 (M= 7.10, SD= 6.49).  Correlations between ADOS joint attention variables and JA acts across the 3 contexts were also explored.  There was a significant negative correlation between ESCS JA and ADOS Showing (r= -.680, p=.021) and Baseline1 JA and ADOS Pointing (r= -.906, p=.0001).  In addition, Baseline1 showing was correlated with ADOS Showing (r= -.667, p=.035).  There were no significant correlations between Baseline2 JA and ADOS variables.

Conclusions: Results indicate that joint attention behaviors among children with ASD may be quite variable across contexts, and may be somewhat higher with more naturalistic paradigms. This reinforces the idea of utilizing multiple measures of joint attention in order to obtain more representative samples of behavior.  These findings have implications for research utilizing joint attention behaviors as predictor and/or outcome variables, as the context may impact the frequency of joint attention behaviors observed.