Feasibility of Ecological Momentary Assessment of Emotion in Adolescents with Autism Spectrum Disorders

Background:   Ecological Momentary Assessment (EMA) is a method of obtaining subjective information from respondents in a natural setting (Shiffman & Stone, 1998).  EMA has been implemented successfully with different populations, including adolescents and severely mentally ill individuals (aan, Hogenelst, & Schoevers, 2012; Forbes et al., 2012); however, the feasibility of using EMA with adolescents who have Autism Spectrum Disorders (ASD) has only recently been investigated (Khor et al., 2014). 

Objectives: EMA was used in the present study to gather subjective reports of affect and behavior in order to: (1) determine the feasibility of using a low-cost EMA protocol with this population (2) examine the frequency of social and nonsocial behavior and (3) examine positive affect (PA) among adolescents with and without ASD in real-world contexts.

Methods: Thirty-nine adolescents participated in the present study.  Sample characteristics are displayed in Table 1.

Table 1: Sample Characteristics for Participants with EMA Data

	N	Mean Age	%Male	Mean IQ
ASD	19	14.11 (SD=3.33)	74%	102.2 (SD=16.9)
TD	20	14.5 (SD=1.96)	80%	113.1 (SD=10.6)

Participants with autism met criteria for ASD on the ADOS.  Technology used in the EMA protocol was free to researchers in order to optimize accessibility and minimize cost.  Prompts were delivered using automated Outlook messaging and surveys were accessed on Qualtrics by participants’ home computers or smartphones. Six EMA prompts were delivered over the course of four days.  Each prompt contained a link to an online survey which asked participants where they were when they were contacted, whom they were with, and what they were doing.  Participants then completed the Positive and Negative Affect Scale for Children (PANAS-C) (Laurent et al., 1999) online.

Results: Participants completed 210 (out of 234) surveys, yielding a 90% response rate (94% among typically developing participants, 85% among participants with ASD).  There was no correlation between age and the number of surveys completed (r=-.23, p>.16) or IQ and the number of surveys completed (r=.27, p>.05). With regard to behavior, there were no group differences in frequency of social or nonsocial behavior.  However, when the nonsocial category was further divided into circumscribed interests for the ASD group and “Primary Interests” for the TD group, a group difference emerged, such that the ASD group participated in more circumscribed interests (relative to time spent engaging in primary interests in the control group), t=2.59, p=.01 (see Figure 1).  There were no group differences in self-reported positive affect (p>.05).  

Conclusions:   The results of the present study confirm that EMA is a promising tool for research with individuals with ASD.  Participants were highly adherent with the protocol, and there were no correlations between age and survey completion rates or IQ and survey completion rates.  Notably, the EMA protocol was low-cost and accessible and thus has great potential for use in under-resourced settings.  Participants with ASD did not engage in nonsocial behaviors significantly more than their typically developing peers; however, they did engage in activities related to their circumscribed interests more than their peers engaged in hobbies or interests they identified as “primary.”