Object choice task (OCT) studies are widely used to assess the phylogenetic and ontogenetic distribution of the understanding of communicative cues, with this understanding serving as a proxy for the discernment of communicative intentions. Recent reviews have found systematic procedural and methodological differences in studies which compare performances across species on the OCT. One such difference concerns the spatial configuration of the test set-up, specifically the distances between the two containers (inter-object distance) and the subject-experimenter distance. Here, we tested dogs on two versions of the task: a central version in which the containers were in the subjects' direct line of vision, and a peripheral version in which the position of the containers was distal to the subject. Half of the subjects were tested with a barrier in the testing environment (as nonhuman primates are tested) and the other half without. We found that dogs tested with a barrier performed significantly better in the central version and were more likely to fail to make a choice in the peripheral version. Dogs tested without a barrier showed comparable performance on the two versions. We thus failed to find support for the distraction hypothesis in dogs. We discuss potential explanations for this, highlighting how methodological differences in the presentation of the OCT can influence outcomes in studies using this paradigm.