How to get psychologists talking about precognition: a review of the Bem experiments and the reaction to them

Mainstream researchers remain uninterested in the findings of parapsychology, perhaps because of the anomalous nature of such findings, with no generally accepted theoretical framework to make sense of the data, or due to a natural suspicion that the methods used are so unlike those used in conventional areas and leave scope for ‘ESP’ effects to represent (methodological) ‘Error Some Place’. This can be overcome by encouraging mainstream researchers to attempt replications for themselves, so that they can scrutinize the methodology directly. Indeed, Daryl Bem has asserted that the ‘holy grail’ for psi researchers would consist of “a straightforward, transparent laboratory demonstration of psi that could be replicated by any competent experimenter” (2003, p. 7), and he has concentrated his efforts on identifying such demonstrations. A second feature of Bem’s approach, which also serves to address concerns about possible artifacts, is his strategy to take well-established and uncontroversial psychological effects and turn them into tests of precognition by reversing the temporal order of the elements to see if some of the effect survives. For example, it is generally accepted (cf. Klauer & Musch, 2003) that if participants are tasked with responding as quickly as possible to a stimulus to register whether it is positive or negative, then their reaction times will be affected by brief (often subliminal) exposure to a related stimulus called the ‘prime’; if the prime and stimulus are congruent then reaction times are faster, if they are incongruent then they are slower. In Bem’s precognitive version, participants are exposed to the prime after responding to the stimulus. Given that the only change to the methodology is to change the order of the elements, this approach is both familiar (and perhaps therefore unthreatening) to mainstream psychologists and also difficult to criticize on methodological grounds without undermining the already-accepted mainstream work. Bem has been remarkably successful in this venture. In 2011 he published in the prestigious Journal of Personality and Social Psychology a summary of nine discrete experiments that described time-reversed effects. Eight of the nine studies gave rise to a significant overall outcomes, with the other suggestive at p = .096 (effect sizes, d, ranging from 0.14 to 0.42). There are also suggestions of meaningful internal effects: 7 of 8 comparisons between those who scored high on a Stimulus-Seeking Scale and those who scored low produced significant differences. Taken together, these studies make a formidable case for the occurrence of precognition. There has been a skeptical response to these findings, particularly in raising suspicions about selective reporting and uncorrected multiple analyses, or inappropriate statistical methods (e.g., Wagenmakers et al., 2011, but see also Bem, Utts & Johnson, 2011). In this talk I will review that reaction and discuss some of the replication attempts by well-known sceptics such as Richard Wsieman and Chhris French as well as our own replication attempts at Northampton