A Preregistered Direct Replication of Bem’s Retroactive Facilitation of Recall Effect
Also known as: Bem replication, Ritchie–Wiseman–French replication, Preregistered replication of retroactive facilitation of recall
This dossier is a research synthesis sourced using AI, not documentary evidence. Use the reference leads to check important claims.
This subject is a 2012 United Kingdom study generally recalled as a preregistered direct replication of Daryl Bem’s retroactive facilitation of recall experiment. It belongs to the modern experimental-psi controversy rather than to a traditional haunting, vision, or spontaneous paranormal report. The central question was whether a person’s present ability to remember words could be improved by practice or restudy that occurred only after that person had already attempted recall. If the predicted pattern were reliably obtained while the future practice items were selected at random after the recall test, it would appear to challenge the ordinary temporal direction assumed in learning and memory research. The study is therefore important chiefly as a tightly framed test of an extraordinary cognition claim, not as evidence that precognition was established. The recalled paradigm is simple in outline. Participants first encountered a set of ordinary verbal stimuli, then attempted to recall as many items as possible, and only afterwards received additional exposure or practice on a randomly determined subset. The critical comparison was between words later selected for practice and words not selected for practice. Bem’s earlier reports had suggested that people recalled more of the words that would subsequently be practiced, a result described as retroactive facilitation of recall. On a conventional account of memory, later practice can improve later retention but cannot alter a recall response already recorded. The anomalous interpretation thus depends on the temporal order, the integrity of the random selection, and the claim that the comparison was specified and evaluated appropriately. Stuart J. Ritchie, Richard Wiseman, and Chris French are recalled as the authors of the replication. Their work is notable because it was preregistered: design and analysis commitments were made before the outcome was known. Preregistration does not make an experiment infallible, and it cannot by itself exclude programming errors, unanticipated ambiguities, or weaknesses inherited from a target protocol. It does, however, reduce latitude to choose a favorable stopping point, outcome transformation, exclusion rule, or statistical analysis after inspecting the data. In a field where small effects and repeated testing can make chance patterns look persuasive, that restriction is a substantive methodological feature rather than a ceremonial label. The recalled result was a failure to reproduce the predicted retroactive-recall advantage in the planned test. That outcome is best stated narrowly. It reported no persuasive support for the specific effect under the replication’s materials, participant pool, randomization and prespecified analysis; it did not demonstrate that every claim labeled psi is false, nor did it independently adjudicate every one of Bem’s experiments. Likewise, a null or near-null outcome does not prove that no tiny effect could ever be detected under any protocol. Its force comes from directly testing a prominent, operationalized claim with controls intended to make retrospective analytic flexibility less influential. The setting matters for interpretation but should not be turned into a causal explanation. The work is associated with Edinburgh and Hertfordshire in the United Kingdom, whereas the controversy that motivated it was strongly associated with North American social-personality and psi research, including Bem’s Cornell-based work. A direct replication attempts to preserve the essential task while necessarily occurring with different researchers, laboratories, participants, software, instructions, local recruitment practices, and expectations. Such differences can reveal a fragile contextual dependency, a harmless implementation mismatch, or the absence of the original effect. They do not themselves provide evidence for a paranormal mechanism or for misconduct. The observable events in this case were ordinary laboratory events: visually presented words, attention and encoding, typed or written recall responses, a post-recall study phase, and a numerical comparison of recall rates. There is no necessary report of unusual sensations, apparitions, trance, altered consciousness, or a participant knowingly receiving information from the future. The alleged anomaly is statistical and counterfactual: an aggregate difference between items categorized by a later random event. This distinction is crucial for cross-case comparison, because it places the claim closer to experimental design, probability, and memory measurement than to anecdotal paranormal experience. Several mundane explanations compete with a retrocausal interpretation. Sampling variation can generate an apparent advantage in a small or noisy dataset. Selective reporting, multiple analytic decisions, optional stopping, and choices about exclusions or outcome definitions can inflate apparent evidential strength even without deliberate deception. Imperfect replication fidelity can also matter: small changes in timing, randomization implementation, stimulus lists, participant engagement, or scoring may alter a conventional memory effect and may be invoked by proponents as reasons a direct replication failed. Conversely, differences in laboratories and personnel do not explain why a genuine, robust, protocol-level precognition effect should disappear; the degree of protocol similarity and statistical power must be checked from documentation before assigning weight to either position. The study circulated within a broader replication debate that mixed technical discussion with unusually public disagreement. Bem’s findings had become a symbol for concerns about conventional significance testing and publication practices, so the replication was read both as a targeted empirical result and as an example in arguments over preregistration, Bayesian inference, meta-analysis, and the publication of null findings. That symbolic role can distort reception. Skeptical accounts may overstate the replication as a universal disproof of psi, while affirmative accounts may treat any departure from the original protocol as sufficient to set the null result aside. A careful dossier keeps the claim at its actual scale: a preregistered direct test of one recalled experimental effect. Commercial and institutional incentives are relevant as contextual influences, not accusations of improper motive. Publicity surrounding a counterintuitive result can attract media attention, book and lecture interest, journal readership, professional visibility, and funding conversations; dramatic refutations can be similarly marketable. Academic incentives to publish novel positive results and structural disincentives for null reports are especially pertinent to replication debates. There is no basis in the recalled lead alone to assign financial motives to the authors or to infer that publicity determined their findings. The useful question is instead whether the design, preregistration, data handling, publication process, and later commentary can be independently examined. This dossier is an unverified recalled synthesis. The supplied lead identifies the paper and its broad significance, but it is discovery context rather than documentary evidence. Exact sample size, preregistration wording, effect estimates, registration location, software details, and publication metadata should be checked against the study report and associated materials before use in a formal review. The reference leads below are deliberately suggestions for later retrieval, not sources consulted for this account.
- Words
- 2,781
- Observations
- 12
- Reference leads
- 5
- Validation score
- 100/100
Chronology
The relevant prehistory is Bem’s earlier report of retroactive facilitation of recall, which supplied the protocol and anomalous prediction that the later study sought to test. The original claim was already controversial because it recast an ordinary word-memory procedure as a possible test of information or influence across time.
In 2012, Ritchie, Wiseman, and French reportedly conducted and published a preregistered direct replication associated with United Kingdom research settings. Their central planned comparison concerned recall of words later assigned to a practice condition versus words not assigned to it.
The recalled planned result did not reproduce a statistically persuasive advantage for the subsequently practiced words. It became a frequently cited episode in the methodological debate over whether preregistered replications should receive special evidential weight when they fail to reproduce an attention-grabbing original result.
Later discussion broadened from this one task to questions about replication fidelity, Bayesian and frequentist standards, meta-analytic claims, publication bias, and the evidential status of experimental psi as a research program.
People, organisations, and setting
Stuart J. Ritchie, Richard Wiseman, and Chris French are the principal researchers associated in recalled accounts with this replication. Daryl J. Bem is the originator of the earlier experimental program that supplied the specific retroactive-recall claim under examination.
The reported setting spans Edinburgh and Hertfordshire in the United Kingdom, with the underlying controversy linked to North American academic psychology and psi research. These place labels identify institutional and research contexts, not a location where an anomalous event was independently observed.
Relevant organisations include the researchers’ university environments, the scholarly journal venue recalled for the replication, and the wider networks of psychology, parapsychology, and open-science commentators. Their respective roles should be distinguished: authors conduct and report a study, journals disseminate it, and later advocates or critics may interpret it without having conducted the experiment.
Reported phenomena
The target phenomenon was an alleged excess of correct recall for verbal items that participants would only later practice or restudy. The claimed temporal sequence is essential: encoding and initial recall occurred before the later practice selection and practice exposure.
At the sensory level, the procedure involved ordinary visual word presentation and the normal perceptual demands of reading a screen or task materials. At the behavioural level, participants attempted retrieval, produced recall responses, and later encountered a subset of items again; the anomaly was inferred from pooled performance data rather than from a participant’s subjective experience.
No recalled account requires participants to report foreknowledge, uncanny feelings, voices, visions, or a consciously perceived future event. A participant could therefore complete the task sincerely and experience it as a routine memory experiment even if an aggregate statistical difference had appeared.
The replication reportedly found no meaningful version of the predicted later-practice advantage in its prespecified analysis. This is a reported outcome of the task, not verification that the future cannot influence cognition under every conceivable condition.
Investigation and method
The investigation was designed as a direct replication rather than an open-ended hunt for anomalous experiences. Its testable prediction was that recall already recorded would differ according to a future random assignment to practice, with the direction specified by Bem’s earlier finding.
Preregistration is the investigation’s defining procedural safeguard. By committing important design and analysis decisions before observing results, the researchers aimed to make it harder for ordinary flexibility in data analysis to generate an apparently anomalous finding after the fact.
Appropriate later review should inspect the exact randomization sequence, whether future practice assignment was inaccessible at the time of recall, participant exclusions, scoring rules, deviations from the protocol, stopping rule, statistical model, and treatment of missing data. Those details determine how direct the replication was and how strongly its result bears on the target claim.
The study was not an investigation of fraud, a survey of belief, or an instrument-based search for physical anomalies. Its evidential object was a predicted relationship in a controlled cognitive task.
Disputes and alternative explanations
A central disagreement concerns evidential scope. Skeptical readers may see a preregistered direct null replication as strong evidence that the original result was a chance or analytic artifact, whereas proponents may argue that subtle procedural differences, experimenter effects, or a genuinely unstable anomaly prevented replication.
A second disagreement concerns statistical standards. The case is often connected to criticism that a conventional threshold can make weak evidence appear decisive when a hypothesis has an exceptionally low prior plausibility, while others emphasize cumulative evidence and different statistical frameworks. Neither framing removes the need to inspect the actual registered analysis and results.
Mundane explanations include sampling error, regression toward the mean, publication bias favoring initially striking outcomes, chance imbalances, task noise, participant inattention, scoring differences, and unrecognized analytic flexibility in the original literature. Replication mismatch is also a mundane possibility, but it should be demonstrated with protocol comparison rather than used as an automatic exemption from a failed direct test.
There is no need to infer dishonesty from disagreement or from nonreplication. The case concerns the reliability and interpretation of a reported effect, and competing explanations should be evaluated through transparent methods and independent repetition.
Transmission, reception, and commercial context
The replication entered public circulation through scholarly publication and then through methodological discussion, blogs, teaching examples, talks, and popular accounts of the replication crisis. Its intelligible narrative—a test of whether future study can improve past memory—made it readily transmissible beyond specialist audiences.
Transmission often strips away the qualification that this was one paradigm within a larger disputed literature. Retellings may describe it either as the definitive defeat of precognition or as merely an unsuccessful attempt with no bearing on the original evidence; both formulations overstate what a single study can establish.
The controversy also had attention-economy effects. Extraordinary positive claims, decisive-sounding rebuttals, and arguments about scientific reform can each create professional visibility and media interest, while null findings historically face weaker publication incentives. These structural influences are relevant to what becomes prominent, but the recalled information does not establish any participant’s commercial or financial motivation.
For later research, the most reliable transmission route is comparison of the original report, preregistration record, replication report, data and code where available, and independent methodological commentary. Secondary summaries should be treated as interpretive layers rather than substitutes for those records.
Cross-case connections
The strongest cross-case motif is retroactive information or influence: an outcome measured now is alleged to correspond to a random event selected later. This connects the case to presentiment, precognition, and other laboratory paradigms that infer anomaly from time-ordered statistical associations rather than from narrative testimony.
A second motif is preregistration as an evidential boundary. Cases in which a dramatic effect is tested under prespecified methods can be compared for randomization security, blinding, sample-size rationale, registered primary outcome, and the difference between planned and exploratory analyses.
A third motif is memory as a fallible measurement system. Unlike witness accounts of a strange event, this task operationalizes performance through item-level recall, which invites comparison with ordinary mechanisms such as encoding variability, retrieval cues, scoring decisions, and attention.
A fourth motif is replication controversy. The case can be compared with other high-profile findings where a null replication, a disputed directness claim, and later meta-analytic arguments coexist without producing immediate consensus.
Limits of the recalled record
This dossier does not treat the supplied recalled lead as documentary proof. It should not be used to report exact numerical outcomes, sample size, effect size, confidence intervals, preregistration language, or publication details without retrieving and checking the primary materials.
The direct replication addressed retroactive facilitation of recall, not every experiment in Bem’s work and not all claims about anomalous cognition. General conclusions must remain proportional to the tested design and to the quality of subsequent independent evidence.
The term direct replication is itself a methodological claim that should be evaluated against instructions, stimuli, timing, randomization, and scoring rather than accepted as a binary label. A later reviewer should record both declared similarities and verified departures from the original protocol.
No paranormal claim is verified here. The case is best retained as a documented-to-be-checked example of how an extraordinary experimental claim was subjected to a preregistered replication attempt and became part of a broader scientific dispute.
Chronology
Bem’s retroactive-recall claim provides the target paradigm.
Earlier experiments reportedly found better initial recall for words that participants would only later practice, creating the claim that a future event was associated with an already completed memory response.
reportedPreregistered United Kingdom replication is conducted.
Ritchie, Wiseman, and French reportedly tested the retroactive facilitation of recall effect in research settings associated with Edinburgh and Hertfordshire.
reportedReplication report is published and discussed.
The study was presented as a preregistered direct replication and reportedly did not support the predicted future-practice recall advantage in its planned analysis.
reportedThe case becomes part of the wider replication debate.
Commentators used the result to discuss preregistration, statistical inference, publication bias, replication fidelity, and the status of experimental psi claims.
reportedPeople and roles
Stuart J. Ritchie
Replication coauthor.He is recalled as one of the researchers who reported the preregistered direct replication.
Richard Wiseman
Replication coauthor.He is recalled as one of the researchers who reported the preregistered direct replication.
Chris French
Replication coauthor.He is recalled as one of the researchers who reported the preregistered direct replication.
Daryl J. Bem
Originator of the target experimental claim.His earlier retroactive facilitation of recall work supplied the specific effect that the replication sought to test.
University research settings in Edinburgh and Hertfordshire
Reported replication setting.These United Kingdom settings should be distinguished from the North American institutional context of the original controversy.
PLOS ONE
Recalled publication venue.The venue is a reference lead requiring verification and should not be treated here as independently checked publication evidence.
Connections to explore
Future event linked to present performance.
The case tests whether a later random practice event is associated with an earlier memory response, making it comparable to experimental precognition and presentiment paradigms.
Suggested search: experimental psi retroactive influence presentiment preregistered replication.Preregistration and evidential discipline.
Its stated preregistration makes it useful for comparing how planned analyses, stopping rules, and exclusions affect claims based on small statistical effects.
Suggested search: preregistration anomalous cognition replication analysis flexibility.Null replication versus effect-specific conclusions.
The case illustrates the distinction between failing to reproduce one operationalized result and disproving an entire contested research domain.
Suggested search: direct replication null result scope inference experimental psi.Memory measurement and ordinary cognitive confounds.
The paradigm depends on encoding, recall scoring, and item assignment, so it should be compared with conventional memory research as well as with paranormal-claim literature.
Suggested search: word recall task scoring random assignment retroactive facilitation.Unretrieved reference leads
A Preregistered Direct Replication of Bem’s Retroactive Facilitation of Recall Effect
Stuart J. Ritchie, Richard Wiseman, and Chris French. · Replication report.
This is the central suggested source for checking the protocol, preregistration, sample, analyses, and outcome.
Suggested search: Ritchie Wiseman French 2012 preregistered direct replication Bem retroactive facilitation recall.Feeling the Future: Experimental Evidence for Anomalous Retroactive Influences on Cognition and Affect
Daryl J. Bem. · Original experimental report.
This is the suggested source for checking the target paradigm, the original reported effect, and the meaning of a direct replication.
Suggested search: Daryl Bem Feeling the Future retroactive facilitation of recall experiment.Why Psychologists Must Change the Way They Analyze Their Data: The Case of Psi
Eric-Jan Wagenmakers and collaborators. · Methodological critique.
This is a suggested lead for understanding objections concerning statistical inference and extraordinary claims.
Suggested search: Wagenmakers Why Psychologists Must Change the Way They Analyze Their Data case of psi.Correcting the Past: Failures to Replicate Psi
Matthew Galak and collaborators. · Replication study or commentary.
This is a suggested lead for comparing other attempted replications of Bem-associated paradigms and their interpretation.
Suggested search: Galak Correcting the Past failures to replicate psi Bem.Preregistration materials associated with the Ritchie, Wiseman, and French study
Ritchie, Wiseman, and French. · Study materials or registration record.
These materials are the key suggested route for verifying what was fixed in advance and whether the reported analysis matched the plan.
Suggested search: Ritchie Wiseman French Bem replication preregistration materials protocol.