Are Bem's "Feeling the Future" results real or statistical artifacts?
Quick Summary
Daryl Bem's 2011 publication of nine experiments appearing to show retroactive influence on cognition — precognition tested with standard social-psychology paradigms — ignited a firestorm about statistical methodology, replication standards, and publication bias.
It became a flashpoint in the broader replication crisis.
Current Consensus
This is the defining controversy of modern psi research and a landmark case in the broader replication crisis. The pro-psi position rests on the 2015 meta-analysis, which aggregates 90 experiments and claims strong Bayesian support. Critics counter that the meta-analysis inherits the problems of its constituent studies (flexible stopping rules, optional analyses, publication bias). Direct pre-registered replications have generally failed. The debate has shifted the conversation in mainstream psychology toward pre-registration and Bayesian methods, making Bem's work historically significant regardless of its parapsychological conclusions. A bibliometric analysis of 162 citing texts (2011-2015) found that psychology journals engaged through replication and Bayesian critique lenses while non-psychology sources (parapsychology, physics) engaged the empirical content — quantifying the disciplinary fracture in reception Bem's 'Feeling the Future' ....
Evidence Breakdown
Based on 24 papersSupporting Evidence
Feeling the Future: Experimental Evidence for Anomalous Retroactive Influences on Cognition and Affect
Bem (2011) -- Nine experiments, eight yielding statistically significant evidence of retroactive influence on cognition and affect
Feeling the Future: A Meta-Analysis of 90 Experiments on the Anomalous Anticipation of Random Future Events
Bem, Tressoldi, Rabeyron & Duggan (2015) -- Meta-analysis of 90 experiments from 33 laboratories in 14 countries yields overall d = 0.09 with Bayesian support exceeding 10^9 to 1
Predictive Physiological Anticipation Preceding Seemingly Unpredictable Stimuli: A Meta-Analysis
Mossbridge, Tressoldi & Utts (2012) -- Independent meta-analysis of predictive anticipatory activity corroborates the general class of effects Bem reported
Meta-Analysis of Free-Response Studies 2009-2018: Assessing the Noise-Reduction Model Ten Years On
Storm, Tressoldi & Di Risio (2020) -- Updated meta-analysis of forced-choice psi experiments supports small but consistent effects
"Future Telling": A Meta-Analysis of Forced-Choice Precognition Experiments, 1935-1987
Honorton & Ferrari (1989) -- Meta-analysis of 309 forced-choice precognition experiments (1935-1987) finds combined z = 11.41 (p = 6.3 x 10^-31) across 62 investigators; fail-safe N of 14,268; no r...
Must Psychologists Change the Way They Analyze Their Data?
Bem, Utts & Johnson (2011) -- Bayesian reanalysis using a knowledge-based normal prior (90th percentile |d| at 0.5) yields combined BF = 13,669 across all nine experiments, with posterior P(H0) = 7...
Feeling the Future Again: Retroactive Avoidance of Negative Stimuli
Maier et al. (2014) -- Independent European replication of Bem's Exp 2 retroactive avoidance paradigm across 7 experiments (N≈2,970); 4/7 significant, meta-analytic ES=0.07, z=3.79, p<.0001, combin...
Critical Evidence
Why Psychologists Must Change the Way They Analyze Their Data: The Case of Psi
Wagenmakers et al. (2011) -- Bayesian reanalysis of Bem's 10 critical tests using default Bayesian t-test: only 1 of 10 yields 'substantial' evidence for psi (BF01 = 0.17); 3 yield 'substantial' ev...
Correcting the Past: Failures to Replicate Psi
Galak et al. (2012) -- Seven direct replications of Bem's experiments all yield null results; argues the original findings reflect flexible data analysis
Failing the Future: Three Unsuccessful Attempts to Replicate Bem's 'Retroactive Facilitation of Recall' Effect
Ritchie, Wiseman & French (2012) -- Three independent labs fail to replicate Bem's retroactive facilitation of recall
Results from a Confirmatory Replication Study of Bem (2011): Precognitive Detection of Erotic Stimuli?
Wagenmakers et al. (2012) -- Pre-registered confirmatory replication of Bem's (2011) Experiment 1 with 100 female participants; all six pre-specified Bayes factor tests favored the null hypothesis ...
A Bayes Factor Meta-Analysis of Bem's ESP Claim
Rouder & Morey (2011) -- Meta-analytic JZS Bayes factor finds BF = 38.7 for emotional stimuli but only 3.23 (erotic) and 1.57 (neutral); argues even the strongest value is "orders of magnitude lowe...
A Bayes Factor Meta-Analysis of Recent Extrasensory Perception Experiments: Comment on Storm, Tressoldi, and Di Risio (2010)
Rouder & Morey (2013) -- Bayesian reanalysis of Bem's data finds the evidence does not favor psi when using appropriate default priors
Too Good to Be True: Publication Bias in Two Prominent Studies from Experimental Psychology
Francis (2012) -- Publication bias analysis suggests Bem's nine experiments show an implausible rate of success given their statistical power
The Garden of Forking Paths: Why Multiple Comparisons Can Be a Problem, Even When There Is No "Fishing Expedition" or "P-Hacking" and the Research Hypothesis Was Posited Ahead of Time
Gelman & Loken (2013) -- "Garden of forking paths" analysis shows how researcher degrees of freedom can generate false-positive results even without deliberate p-hacking
Why Most Published Research Findings Are False
Ioannidis (2005) -- The foundational mathematical framework for understanding why Bem's results are likely false positives; shows that with small effects (d=0.09), multiple paradigms tested, and fl...
A Preregistered Multi-Lab Replication of Maier et al. (2014, Exp. 4) Testing Retroactive Avoidance
Maier et al. (2020) -- Preregistered multi-lab replication (N=2,004) of Maier et al. (2014, Exp. 4) retroactive avoidance with quantum RNG; Bayesian analysis yields BF01 = 4.38 (moderate evidence f...
Retro-priming, priming, and double testing: psi and replication in a test–retest design
Rabeyron (2014) -- Test-retest retro-priming study: 28 high scorers from initial study (N=162) failed to replicate on retest (es=-0.25, ns), with a significant negative cross-session correlation (r...
Editors' Introduction to the Special Section on Replicability in Psychological Science: A Crisis of Confidence?
Pashler & Wagenmakers (2012) -- Editors' introduction to the Perspectives on Psychological Science replicability special section explicitly frames Bem's (2011) precognition findings as a catalyst f...
False-Positive Psychology: Undisclosed Flexibility in Data Collection and Analysis Allows Presenting Anything as Significant
Simmons, Nelson & Simonsohn (2011) -- Monte Carlo simulations demonstrate that four common researcher degrees of freedom (flexible DVs, optional stopping, covariate selection, condition dropping) i...
Fearing the Future of Empirical Psychology: Bem's (2011) Evidence of Psi as a Case Study of Deficiencies in Modal Research Practice
LeBel & Peters (2011) -- Methodological commentary arguing that three systemic deficiencies in modal research practice (MRP) — overemphasis on conceptual replication, failure to verify measurement ...
Back from the Future: Parapsychology and the Bem Affair
Alcock (2011) -- Experiment-by-experiment critique identifying mid-study procedure changes, uncorrected multiple t-tests (Bonferroni renders p=.01 nonsignificant at adjusted p≈.06), ad hoc 2-item s...
Searching for the Impossible: Parapsychology's Elusive Quest
Reber & Alcock (2019) -- Broad critique arguing psi is impossible a priori based on violations of causality, thermodynamics, and time-asymmetry; cites Bem's 2017 admission that his experiments were...
Raising the value of research studies in psychological science by increasing the credibility of research reports: the transparent Psi project
Kekecs et al. (2023) -- Pre-registered multi-lab replication (10 labs, 9 countries, N=2,115, 37,836 trials) of Bem's Exp 1, co-designed by consensus panel of proponents and skeptics; hit rate 49.89...