Reversing the Colavita visual dominance effect

Abstract

Many researchers have taken the Colavita effect to represent a paradigm case of visual dominance. Broadly defined, the effect occurs when people fail to respond to an auditory target if they also have to respond to a visual target presented at the same time. Previous studies have revealed the remarkable resilience of this effect to various manipulations. In fact, a reversal of the Colavita visual dominance effect (i.e., auditory dominance) has never been reported. Here, we present a series of experiments designed to investigate whether it is possible to reverse the Colavita effect when the target stimuli consist of repetitions embedded in simultaneously presented auditory and visual streams of stimuli. In line with previous findings, the Colavita effect was still observed for an immediate repetition task, but when an n-1 repetition detection task was used, a reversal of visual dominance was demonstrated. These results suggest that masking from intervening stimuli between n-1 repetition targets was responsible for the elimination and reversal of the Colavita visual dominance effect. They further suggest that varying the presence of a mask (pattern, conceptual, or absent) in the repetition detection task gives rise to different patterns of sensory dominance (i.e., visual dominance, an elimination of the Colavita effect, or even auditory dominance).

Appendix

The percentages of correct responses for all three target types (visual, auditory, and bimodal) were rather low as compared to those values observed in previous studies of the Colavita visual dominance effect (e.g., see Ngo et al. 2010). We suspected that this may have been due to the difficulty that participants experienced in trying to simultaneously divide their attention between the visual and auditory modalities in the present study. In order to assess this possibility, we therefore conducted an additional control experiment in which the participants (N = 10) were instructed to detect only unimodal visual or auditory n-1 repetition targets in separate blocks of experimental trials. The procedure was identical to that used in the main experiment reported in the text, except for the fact that the participants completed five blocks of trials detecting each target modality. The order of presentation of each target modality was counterbalanced across participants. At the start of each block of trials, the participants were instructed to respond to n-1 repetition targets in either the auditory or visual modality by pressing a key on the keyboard.

The percentage of correct responses and the RTs from those trials in which the participants responded correctly for each participant and condition were calculated. Separate paired-samples t tests revealed that participants responded significantly more accurately to auditory targets (M = 79.6%, SE = 4.6%) than to visual targets (M = 67.6%, SE = 4.5%), t(9) = 2.71, P = .02. What is more, the participants detected the visual targets (M = 591 ms, SE = 29 ms) significantly faster than the auditory n-1 targets (M = 608 ms, SE = 34 ms), t(9) = 2.40, P = .04. This may have been due to the fact that the auditory stimuli were time-evolving, whereas the visual stimuli were immediately perceptible. That is, the information needed to identify a naturalistic sound such as a cat’s “meow” may not arrive until relatively late in the presentation of the sound. All of the information contained within a picture, on the other hand, is immediately available as soon as it is presented (and is likely processed after approximately 150 ms of stimulus onset; Oram et al. 2002; Thorpe et al. 1996). It is also possible that there were some speed-accuracy trade-offs that might have been present in the data (see Edwards 1965) given that participants were faster but less accurate at detecting visual, as compared to auditory, targets.

The results of this control experiment therefore confirm that participants found the task much easier when they need only attend to one sensory modality (either vision or audition). This was evidenced by the substantially higher percentages of correct responses (68 and 80% for visual and auditory, respectively) as compared to those observed in Experiment 1 (35 and 56% for visual and auditory targets, respectively), where participants had to divide their attention between the visual and auditory modalities in order to detect either type of target n-1 repetitions in the experiments reported in the main text.