The Diagnostic Accuracy of Eye Tracking Technology in the Assessment of Adult Attachment Style

Abstract

The purpose of this study was to examine the diagnostic value of eye tracking technology in measuring adult attachment style. This investigation of diagnostic accuracy was cross-sectional and was one of the primary observational and quantitative analysis studies. The population of interest for this study included all healthy adults aged 20-45 years with higher education living in Tehran in 2021. For this purpose, 44 adults aged 20 to 45 years who met the criteria of the project were selected by random sampling. Of these 44 individuals, 36 were women and 8 were men. This study was conducted in two phases. In the first phase, a researcher-made questionnaire along with the revised Adult Attachment Scale (R-AAS) as a reference test was created in the Google Form environment, and individuals who met the inclusion criteria were selected and included in the study. In the second phase, participants were presented with one minute and 39 seconds of Adult Attachment Projective Picture System (AAP) images as an activator of attachment, and their eye movements were simultaneously recorded with an eye-tracking device. For statistical analysis, the random forest test, logical regression, and ROC curves were calculated in RStudio 4.1.1. The variable “time to first fixation” in the child’s AOI in the ambulatory image predicted secure and insecure attachment (B = -1.95, sig. 0.01, p≤ 0.05). The correct percentage for classifying securely and insecurely attached adults was estimated to be 54.2% and 75%, respectively. Since the variables “duration of first fixation” and “number of fixations” were important according to the results of the random forest test, it is suggested that they be further investigated in future studies.

Keywords

• attachment
• eye tracking
• diagnostic accuracy