Finding correlations between subjective and objective measures of awareness using masked words

Document Type: 
ASSC Conference Item
Article Type: 
two-alternative forced choice, visual word recognition, e-detection, confidence rating
Deposited by: 
Anastasia Gorbunova
Contact email:
Date of Issue: 
Event Dates: 
5-8 June 2009
Event Location: 
Event Title: 
13th Annual Meeting of the Association for the Scientific Study of Consciousness
Event Type: 
ASSC Conference
Presentation Type: 
Publish status: 
The current study explores the idea proposed by Marcel (1983) that perceptual analysis can be performed in the absence of conscious awareness. In this experiment, participants were asked to perform two forced choice tasks based on the information gathered from a masked word presented for 50 milliseconds: (1) semantic discrimination in which participants had to choose between a word that was associated in meaning with the masked stimulus and an unrelated alternative, and (2) e-detection where participants responded “yes” if the masked word contained the letter ‘e’, or “no” if it did not. Performance on both tasks was better than chance, but performance on e-detection (mean % correct 67.2) was slightly better than on semantic discrimination (mean % correct 61.9). This provided an objective measure of partial awareness of the stimuli. For both tasks, participants were asked to rate their confidence after each trial. Confidence ratings ranged from 1 (complete guess), to 5 (identified the masked word), and provided a measure of subjective awareness. Pearson correlations were calculated to find that highly rated items generally constituted more correct responses than low rated ones (r = 0.549 for semantic discrimination, and r = 0.365 for e-detection). However, these correlations were far from perfect, suggesting that some information is gathered from masked words even when the participant feels that he or she is simply guessing. Possible mechanisms by which these tasks can be performed when confidence ratings are less than 4 and implications of these findings are discussed.
SAW-SelfRating.ppt4.24 MB