Anxiety and Attentional Bias in Preschool-Aged Children: An Eyetracking Study

Prior to conducting any interviews for the present research, all interviewers first watched at least two videos of ADIS interviews. For each, they were asked to assign diagnoses and CSRs. These were then discussed with a senior clinical psychologist, with reference to the diagnoses and CSRs given by the original interviewer. Once the clinical psychologist was satisfied with their performance, trainees observed two ADIS interviews in our clinic. Again they were asked to assign diagnoses and CSRs and these were discussed after the interview with the psychologist who had conducted the assessment. Once they had completed at least two of these and assigned diagnoses and CSRs that were equivalent to those made by the interviewing clinical psychologist, they were then observed conducting at least two interviews. Again, assigned diagnoses and CSRs were discussed with the psychologist who had observed the interview. The clinical psychologist who oversaw the training made a decision regarding the interviewers readiness to conduct interviews for the study based on their overall performance on these tasks. Note that this procedure is followed for training psychologists in our clinic, where excellent reliability is reported (Lyneham, Abbott, and Rapee 2007). Interviewers attended regular supervision sessions with a senior clinical psychologist where difficult diagnostic cases were discussed.

Following a request from an anonymous reviewer, we examined whether the results were driven by participants with a diagnosis of social anxiety disorder. The significant between-group effects reported above for length of first fixation and dwell time were examined comparing three groups of participants: those with a diagnosis of social anxiety disorder anywhere in their diagnostic profile (n = 22), clinically anxious participants without a diagnosis of social anxiety disorder (n = 15) and participants with no anxiety diagnosis (n = 44). These analyses also showed no group by emotion interactions and a significant main effect of group for both length of first fixation, F(2, 80) = 3.93, p = 0.023, and dwell time, F(2, 80) = 6.08, p = 0.003. Follow-up t-tests comparing groups showed that the group with social anxiety disorder had shorter first fixations to faces and shorter dwell time to faces than the control comparison group but neither reached significance (p > 0.08). The clinically anxious group without social anxiety disorder also had shorter first fixations to faces and shorter dwell time to faces than the control comparison group and both reached significance (p < 0.005). The two anxious groups were not significantly different from each other on either dependent variable (p > 0.09).