Subliminal priming with nearly perfect performance in the prime-classification task

Subliminal priming with nearly perfect performance in the prime-classification task Finkbeiner

Tag layer of the DHSS Meta-Tagging Project. The article text is reproduced from the staged copy named in sourceDesc and is not altered; the tags are held separately in standOff so that overlapping and nested spans are preserved.

Subliminal priming with nearly perfect performance in the prime-classification task 2011 Attention, Perception, & Psychophysics 10.3758/s13414-011-0088-8 Finkbeiner2011-subliminal-priming-with-nearly-perfect-perfo Finkbeiner2011-subliminal-priming-with-nearly-perfect-perfo.txt, the project's staged copy 7dd23cd8f209e00a5850a7b733cda562f0149d9c not-redistributable cognitive-psychology

the paper's own words

the tag layer; a sentence written by the tagger about the paper, always anchored to a quotation but not itself a quotation

0 spans located by searching the stored text for the quote's exact words, allowing any whitespace between them existence Quotation withheld because this record is not cleared for redistribution. subliminal priming Quotation withheld because this record is not cleared for redistribution. behavioural-experiment behavioural-experiment Quotation withheld because this record is not cleared for redistribution. mechanisms using a backward mask that is visually similar to the prime (a letter string masking a letter-string prime) suppresses prime identity strongly enough to reach chance-level identification at durations that still produce priming, whereas dissimilar and repeated masks let the prime pop out Quotation withheld because this record is not cleared for redistribution. mechanisms differences in masking procedure, rather than differences in control of task difficulty, are what drive the opposite conclusions of the two studies status: the paper's own words: We suggest that Quotation withheld because this record is not cleared for redistribution. mechanisms when primes and targets share a stimulus set, learned stimulus-response mappings can be triggered automatically by an unseen prime, so apparent above-chance prime classification may reflect nonconscious motor activation rather than awareness Quotation withheld because this record is not cleared for redistribution. Even with 240 prime-classification trials the design remains open to the low-power criticism levelled at subliminal priming studies. Quotation withheld because this record is not cleared for redistribution. The prime-classification measure used here may itself have been insufficiently sensitive to prime awareness. Quotation withheld because this record is not cleared for redistribution. The .95 posterior criterion used to declare subliminality may be too strict, and the priming result is also reported under a relaxed .90 criterion. Quotation withheld because this record is not cleared for redistribution. the letters in the backward mask were randomly repositioned on every trial eliminates: that the masking effect is specific to one mask arrangement rather than to masking as such · outcome (stated by the tagger, no quotation): held Quotation withheld because this record is not cleared for redistribution. subliminal priming subliminal priming Quotation withheld because this record is not cleared for redistribution. prime-classification task prime-classification task Quotation withheld because this record is not cleared for redistribution. task difficulty task difficulty Quotation withheld because this record is not cleared for redistribution. task-difficulty artifact task-difficulty artifact Quotation withheld because this record is not cleared for redistribution. THE CRITERION STATED AS AN IF-THEN, WHICH IS WORTH HOLDING VERBATIM: `If, in a separate task that assesses PRIME AWARENESS, subjects are not able to identify the prime stimuli at BETTER THAN CHANCE levels (i.e., d' is not statistically different from zero), then subliminal priming is said to have occurred.` `is said to have occurred` - the phenomenon is the conjunction of two measurements, and the paper is explicit that the word is licensed by the second one. dehaene1998 and holcomb2005 argue about WHAT was processed without awareness; this record is about what licenses the word `subliminal` at all, which is the prior question and the corpus had no clean statement of it Quotation withheld because this record is not cleared for redistribution. AND THE CRITIQUE IT ANSWERS IS AN ATTACK ON THE MEASUREMENT, NOT ON THE EFFECT: `In a recent critique of subliminal priming, Pratte and Rouder (2009) suggested that previous claims of subliminal priming may have been due to A failure to control the task difficulty between the experiment proper and the prime-classification task.` If the certifying task is simply harder, chance performance on it certifies nothing. That is the strongest form of methodological criticism - it does not dispute a result, it disputes what the result is evidence OF - and a corpus that records priming findings without recording this cannot represent why the literature is contested Quotation withheld because this record is not cleared for redistribution. THE ANSWER IS BUILT INTO THE STIMULUS RATHER THAN ARGUED IN A DISCUSSION - PRINT THE PRIMES IN COLOUR: `Quotation withheld.` 92.1% on colour, chance on identity, same stimulus, same trial, within subject. The colour report proves the task was easy and the subject was attending, so the identity failure cannot be blamed on difficulty. A control made of the stimulus, not of an argument - and the cleanest reply to an artifact objection I have tagged Quotation withheld because this record is not cleared for redistribution. AND A FOOTNOTE CARRIES THE STRONGEST SENTENCE IN THE PAPER: `The primes that were presented for 40 to ~60 ms were VISIBLE (d' > 2.0); the primes presented for LESS THAN 40 ms were NOT VISIBLE (d' not different from zero), and yet they produced reliable priming effects.` Visible primes prime and invisible primes prime. So the priming function and the awareness function are separable, and a study that finds priming has learned nothing about awareness until it measures awareness separately - which is the whole argument of this paper, stated in a footnote about a different experiment Quotation withheld because this record is not cleared for redistribution. contradicts → Pratte and Rouder (2009) Quotation withheld because this record is not cleared for redistribution. contradicts → Pratte and Rouder (2009) Quotation withheld because this record is not cleared for redistribution. replicates → Finkbeiner & Forster (2008); Kouider & Dehaene (2007) Quotation withheld because this record is not cleared for redistribution. builds-on → Morey et al. (2008) mass-at-chance (MAC) model Quotation withheld because this record is not cleared for redistribution. subliminal priming —distinct-from→ task-difficulty artifact subliminal priming task-difficulty artifact strength of the relation: 0.9 · note: THE RESULT, AND 'distinct-from' IS DOING REAL WORK HERE. The artifact was proposed as an ALTERNATIVE EXPLANATION of subliminal priming; this record separates them by controlling the artifact and keeping the effect. So the edge records not that two concepts differ, but that one was offered as the other and failed. Weight 0.9 rather than 0.95 because the paper itself attributes the contrast with Pratte and Rouder to 'important differences…' - which is a second unresolved difference between the labs, stated by the author. Quotation withheld because this record is not cleared for redistribution. repeated-measures ANOVA repeated-measures ANOVA Quotation withheld because this record is not cleared for redistribution. one-sample t-test one-sample t-test Quotation withheld because this record is not cleared for redistribution. paired-samples t-test paired-samples t-test Quotation withheld because this record is not cleared for redistribution. Bayes factor Bayes factor Quotation withheld because this record is not cleared for redistribution. hierarchical Bayesian model hierarchical Bayesian model Quotation withheld because this record is not cleared for redistribution. posterior probability posterior probability Quotation withheld because this record is not cleared for redistribution. null-hypothesis significance testing null-hypothesis significance testing Quotation withheld because this record is not cleared for redistribution. signal-detection sensitivity (d-prime) signal-detection sensitivity (d-prime) Quotation withheld because this record is not cleared for redistribution. LOESS local regression LOESS local regression Quotation withheld because this record is not cleared for redistribution. confidence interval confidence interval Quotation withheld because this record is not cleared for redistribution. Reliable masked congruence priming was obtained at both 40 ms and 50 ms prime durations. Quotation withheld because this record is not cleared for redistribution. Presenting primes in colour let subjects report the prime's colour almost as accurately as they classified targets, which the author takes as successful equating of difficulty across the two tasks. Quotation withheld because this record is not cleared for redistribution. Prime identification was 51.8%, not above chance, with 40-ms primes and 55.6%, above chance, with 50-ms primes. Quotation withheld because this record is not cleared for redistribution. Subliminality was confirmed for 16 of 21 subjects with 40-ms primes but for only 1 with 50-ms primes. Quotation withheld because this record is not cleared for redistribution. Restricting the priming analysis to only those subject-by-duration cells whose subliminality was confirmed still yielded a significant priming effect. Quotation withheld because this record is not cleared for redistribution. The priming effect did not decrease as the criterion for subliminality increased. Quotation withheld because this record is not cleared for redistribution. The authors conclude subliminal priming can be observed even when guarding against the task-difficulty artifact. Quotation withheld because this record is not cleared for redistribution. subliminal priming effect subliminal priming effect Quotation withheld because this record is not cleared for redistribution. target-classification task target-classification task Quotation withheld because this record is not cleared for redistribution. task-difficulty artifact task-difficulty artifact Quotation withheld because this record is not cleared for redistribution. mass-at-chance model mass-at-chance model Quotation withheld because this record is not cleared for redistribution. null-sensitivity problem null-sensitivity problem Quotation withheld because this record is not cleared for redistribution. prime duration prime duration Quotation withheld because this record is not cleared for redistribution. prime type prime type Quotation withheld because this record is not cleared for redistribution. RTs RTs Quotation withheld because this record is not cleared for redistribution. Prime Color Prime Color Quotation withheld because this record is not cleared for redistribution. prime–target SOA prime–target SOA Quotation withheld because this record is not cleared for redistribution. M = 92.1% — mean accuracy of reporting the prime's colour in the prime-classification task Quotation withheld because this record is not cleared for redistribution. both greater than 90% — colour-report accuracy and target-classification accuracy, showing the two tasks were matched for difficulty Quotation withheld because this record is not cleared for redistribution. 51.8% ... 55.6% — prime identity accuracy at 40 ms and at 50 ms prime durations Quotation withheld because this record is not cleared for redistribution. 16 of the 21 — subjects judged to be performing at chance on prime identity with 40-ms primes by the MAC model Quotation withheld because this record is not cleared for redistribution. 21 — number of participants in the experiment Quotation withheld because this record is not cleared for redistribution. Bayes factor of 0.26 — Bayes factor for the priming effect among subliminality-confirmed cells, i.e. the alternative is 3.85 times more likely than the null Quotation withheld because this record is not cleared for redistribution. 9:1 ... 16:1 — odds in favour of subliminality and the resulting odds in favour of priming under the relaxed .90 criterion Quotation withheld because this record is not cleared for redistribution. approximately 570 — prime-classification trials that would be needed, at n = 21 and 52% true performance, to keep the chance of wrongly accepting the null below 5% Quotation withheld because this record is not cleared for redistribution.

The full text of this article is not redistributable, so it is not included here. The evidence sentences are withheld with it; the tag layer and its grounding status remain. The article itself is at https://doi.org/10.3758/s13414-011-0088-8.