TY - JOUR

T1 - Revising beliefs about the merit of unconscious thought

T2 - Evidence in favor of the null hypothesis

AU - Newell, Ben R.

AU - Rakow, Tim

PY - 2011/12

Y1 - 2011/12

N2 - Claims that a period of distraction-designed to promote unconscious thought-improves decisions relative to a period of conscious deliberation are as multifarious as they are controversial. We reviewed 16 experimental studies from two labs, across a range of tasks (multi-attribute choice, creativity, moral dilemmas), only one of which found any significant advantages for unconscious thought. The results of each study were analyzed using Bayesian t tests. Unlike traditional significance tests, these tests allow an assessment of the evidence for the null hypothesis\-in this case, no difference between conscious and unconscious thought. This is done by computing the likelihood ratio (or Bayes factor), which compares the probability of the data given the null against the probability of the data given a distribution of plausible alternate hypotheses. Almost without exception, the probability of the data given the null exceeded that for the alternate distribution. A Bayesian t test for the average effect size across all studies (N= 1,071) yielded a Bayes factor of 9, which can be taken as clear evidence supporting the null hypothesis; that is, a period of distraction had no noticeable improving effect on the range of decision-making tasks in our sample.

AB - Claims that a period of distraction-designed to promote unconscious thought-improves decisions relative to a period of conscious deliberation are as multifarious as they are controversial. We reviewed 16 experimental studies from two labs, across a range of tasks (multi-attribute choice, creativity, moral dilemmas), only one of which found any significant advantages for unconscious thought. The results of each study were analyzed using Bayesian t tests. Unlike traditional significance tests, these tests allow an assessment of the evidence for the null hypothesis\-in this case, no difference between conscious and unconscious thought. This is done by computing the likelihood ratio (or Bayes factor), which compares the probability of the data given the null against the probability of the data given a distribution of plausible alternate hypotheses. Almost without exception, the probability of the data given the null exceeded that for the alternate distribution. A Bayesian t test for the average effect size across all studies (N= 1,071) yielded a Bayes factor of 9, which can be taken as clear evidence supporting the null hypothesis; that is, a period of distraction had no noticeable improving effect on the range of decision-making tasks in our sample.

UR - http://www.scopus.com/inward/record.url?scp=82955194618&partnerID=8YFLogxK

U2 - 10.1521/soco.2011.29.6.711

DO - 10.1521/soco.2011.29.6.711

M3 - Article

AN - SCOPUS:82955194618

SN - 0278-016X

VL - 29

SP - 711

EP - 726

JO - SOCIAL COGNITION

JF - SOCIAL COGNITION

IS - 6

ER -