The Talking Heads Attentional Bias Assessment Task is a recently developed, freely available, methodology for assessing anxiety-linked attentional bias to ecologically relevant emotionally negative information. The task employs video-based stimuli conveying emotionally negative or benign information concerning issues that individuals commonly encounter in daily life. It utilizes a dual-probe methodology with the aim of indexing the degree to which viewers’ attention is automatically captured by the more negative videos. When introducing this task, Mazidi et al. (2025) demonstrated that the resulting attentional bias index i. has excellent internal consistency; ii. is significantly correlated with trait anxiety; and iii. mediates the association between trait anxiety and the state anxiety elevation elicited by the emotional videos. However, this prior study did not verify that the anxiety-linked attentional bias indexed by this task is automatic, in the sense that it is independent of volitional viewing preference. The present study was designed to determine the replicability of Mazidi et al.’s original findings, while overcoming this prior limitation. A sample of 170 undergraduate students completed a variant of the Talking Heads Attentional Bias Assessment Task, that now included an additional condition designed to directly measure volitional viewing preference. All of the three effects described above were replicated in full. Moreover, the association between trait anxiety and the resulting index of attentional bias was independent of volitional viewing preference. These findings attest to the capacity of this task to yield a psychometrically appropriate index of trait anxiety-linked bias in automatic attentional capture by emotionally negative information. • Developed a reliable, ecologically valid attentional bias task using video stimuli • Replicated prior findings linking anxiety to attentional bias with high reliability • Findings confirm anxiety-linked attentional bias operates automatically • Task materials and code are openly available for replication and extension
Mazidi et al. (Tue,) studied this question.