MethodParticipants: 16 University of Arizona students

Materials:76 items, each consisting of a prime word such as market, and a semantic associate such as store. The items were matched for frequency of occurrence, length, imageability, and neighbourhood density. One half of base words and their neighbours contained an e and the other half didn’t. Additional 12 items were presented for 100 ms and were used as fillers.

Task:2-alternative forced choice with masked priming paradigm (Forster & Davis, 1984)2 conditions (tasks): semantic discrimination and e-detection.Participants receive no feedback.Confidence ratings on a scale from 1 to 5; 1 = complete guess, 5 = I saw the word and am confident of my decision

Background

What role does awareness play in priming?

Forster (1987): Form priming yields different results depending on whether the prime is masked or not

How much information does one receive from unconsciously perceived primes?

How does phenomenology correlate with performance on discrimination tasks involving masked words?

Can participants perform Semantic Discrimination and e-detection in the absence of phenomenological awareness?

Cheesman and Merikle (1986): Objective measures of awareness show better performance than subjective measures

Marcel, A. (1983): Visual record of the prime could be wiped out, but not the semantic record

Kouider & Dupoux (2004): Unconscious semantic priming is a result of a perceptual illusion, where a fragment of the stimulus is perceived consciously

Finkbeiner et al. (2005): e-detection task and masked priming

Conclusions• At 50 ms participants are able to discriminate between alternatives semantically and are able to perform the e-detection task better than chance as a group

• Marcel’s (1983) claim was not replicated, but it wasn’t refuted either, since only performance on Semantic discrimination was statistically indistinguishable from that on e-detection

• Semantic discrimination isn’t better than chance when participants are guessing, while e-detection is

Discussion

• Chance performance on Semantic discrimination items rated below 3 suggests that awareness is required in this task

• E-detection results suggest that some information is gathered from masked words even when the participants feel that they are simply guessing - visual record can persist in the absence of subjective awareness

• Subject correlation between tasks suggests that participants are assumed to possess different degrees of an ability (e.g., picking up information from the prime) that affects their performance on both tasks

• The absence of an item correlation between tasks suggests that 100% accuracy on some items in the Semantic Discrimination does not necessarily rely on visibility of those items

Summary of Results Semantic Discrimination

E-detection

% Correct 61.9** 67.2**

Paired Samples Correlation (subjects)

r = 0.597*

Results

• Participants report being unable to identify the primes for most of the time at 50 ms prime durations

• Group performance was better than chance on both Semantic Discrimination and E-detection

• Performance on Semantic Discrimination task did not differ significantly from that on E-detection, but the tasks correlated significantly

• 11 participants were better than chance (63%) on e-detection, and 7 on Semantic Discrimination

*p < 0.05; ** p < 0.01; *** p < 0.001

Purpose• Test Marcel’s (1983) claim that visual masking only obliterates the visual record, leaving the semantic record intact.

• We should expect performance on Semantic Discrimination to be better than chance, but performance on e-detection to be at chance in the absence of subjective awareness

• Look for correlations between semantic discrimination and e-detection

• Look for correlations between confidence ratings and performance

References

Cheesman, J., & Merikle. P. M. (1986). Distinguishing conscious from unconscious perceptual processes. Canadian Journal of Psychology, 40, 343-367.

Finkbeiner, M., Forster, K.I., Nicol, J. & Nakamura, K. (2004). The role of polysemy in masked semantic and translation priming. Journal of Memory & Language, 51(1), 1-22.

Forster, K.I. & Davis, C. (1984). Repetition priming and frequency attenuation in lexical access. Journal of Experimental Psychology: Learning, Memory and Cognition, 10, 680-698.

Forster, K.I. & Forster, J. (2003). DMDX: A Windows display program with millisecond accuracy. Behavioral Research Methods: Instruments and Computers, 35, 116-124.

Kouider, S., & Dupoux, E. (2004). Partial Awareness Creates the 'Illusion' of Subliminal Semantic Priming. Psychological Science, 15(2), 75-81.

Marcel, A. J. (1983). Conscious and unconscious perception: experiments on visual masking and word recognition. Cognitive Psychology, 15, 197-237.

• On Semantic Discrimination, only performance on items rated 3 and higher is significantly better than chance

• On e-detection, performance on items across all ratings is better than chance when ratings of 1 and 2 are taken together