Here we have a guest blog from Michael Duggan who is one of the authors of the meta-analysis of replications of Daryl Bem’s controversial retroactive facilitation experimentshttp://papers.ssrn.com/sol3/papers.cfm?abstract_id=2423692, and has a long term interest in serious psi research. He is classically trained in the sciences (Ph.D Catalytic chemistry), but it is psi research that truly excites him. Here, he blogs about the controversy surrounding the publication of Daryl Bem’s seminal Feeling The Future Paper in 2011, and the experience of trying to publish a meta analysis of 90 experiments supporting Bem’s earlier findings.
“Feeling the Future” of Bem’s findings
The publication of the paper “Feeling the future: experimental evidence for anomalous retroactive influences on cognition and affect” in the mainstream and highly esteemed Journal of Personality and Social Psychology in 2011 by Daryl Bem (Bem, 2011), professor emeritus at Cornell University, was the academic equivalent of a Big Bang for many scientists and lay persons alike because the heretical message was “precognition is scientifically supported”.
But what did Daryl Bem find really? In a series of nine experiments, he found that in four different tasks, the performance of the participants was related to what happened after their decision was made. For example, he found that participants were able to detect behind which door hid an erotic image approximately 3% above chance level or avoid a negative image nearly 2% above the expected chance level. In another task, defined as retroactive facilitation of recall, participants recalled approximately 2 to 4% more words if they were later repeated several times. In another study, subjects were exposed to a prime after categorizing a picture as pleasant or unpleasant. Bem showed that when the future prime was congruent, i.e., pleasant prime – pleasant picture, the reaction times were faster. It is as if the future can percolate into the present to influence participant’s behavioural responses in these standard but time-reversed psychological tests.
To have an idea of the impact of this paper, it is instructive to take note of the number of times this paper was cited in other scientific journals: 342 times as of 03th July 2015. Furthermore if you search “Feeling the future” with Google you can find approximately two million pages that includes a dedicated page on Wikipedia. Additionally Bem was interviewed by many radio and TV stations, from the Colbert Report to the David Letterman show and Through The Wormhole with Morgan Freeman. Such widespread and generally positive media coverage is unprecendented for any kind of parapsychological type finding. Bem deserves plaudits for breaking new ground in this way.
It is also extremely important to bear in mind that evidence for precognition is not limited to the behavioural studies of Bem and others but extends to presentiment effects in human physiology (Mossbridge, et al, 2012) and forced choice guessing (Honorton and Ferrari, 1989).
This event is quite interesting to observe both from the scientific and sociological point of view. From a scientific point of view, it is curious to note that most other papers related to “parapsychological” stuff published in mainstream journals (e.g. Psychological Bulletin, British Journal of Psychology) did not receive similar “attention” even if they usually referred to telepathy or distant mental interaction. Very probably, precognition is deemed not so “woo” as other parapsychological phenomena. This more favourable attitude to precognition might have something to do with recent serious attention to the idea of retrocausation in quantum mechanics through the work of physicists such as John Cramer and Daniel Sheehan. Indeed, several AAAS (The American Association for the Advancement of Science) symposia have been dedicated to exploring retrocausality from both a physics and parapsychological perspective, inviting contributors from both fields.
As expected, Bem’s paper was analyzed and criticized, in particular by scientists, starting from theoretical grounds progressing into qualms about methodology and statistics, and finally questioning Bem’s interpretation of results. The main theoretical criticisms stand on the premise that “precognition” does not exist and hence all evidence supporting it is surely based on flawed procedures or interpretations. It is clear that this attitude is based on a prejudiced old Newtonian ideology rather than an examination of the state of the debate. As mentioned above, serious scientists are exploring retrocausation with obvious implications for the possibility of precognition. More generally, quantum mechanics with its basis in probability theory, non-locality, and entanglement effects is so strange and counter-intuitive, that outlawing parapsychological phenomena because they contradict the so-called laws-of-physics seems somewhat premature. More interesting, are the criticisms related to methodology and statistics. Initially criticisms of the reviewers and the editor that accepted the paper for publication evolved into an attack on the type of statistics used, a situation inflamed by the fact that 8 out of 9 experiments yielded statistically significant results that were “too good to be true”. Eventually these discussions lead to a more general critique –in some quarters – of how ALL psychology experiments are conducted leading to constructive advice moving forward, for example; specifying a priori which hypotheses are exploratory and which ones are confirmatory, pre-registering the studies in order to reduce so-called questionable research practices (i.e. optional stopping in the data collection; post-hoc selection of only statistically significant comparisons; arbitrary elimination of outliers, etc.), and ultimately including the abandonment of frequentist statistics in favour of Bayesian ones. This last criticism is particularly interesting because the debate, supported by scientific papers, yielded very different results using the same data. For example, Eric-Jan Wagenmakers of the Department of Psychology University of Amsterdam and his colleagues (Wagenmakers, et al.,2011), found weak to nonexistent support when analyzing Bem’s data, whereas a rebuttal by Bem, with two experts in Bayesian statistics (Bem et al., 2011), found strong support. Another independent Bayesian analysis carried out by Jeff Rouder of the Department of Psychological Sciences University of Missouri and Richard Morey of the Rijksuniversiteit in Groningen(Rouder and Morey, 2011), concluded as follows:“There is some evidence, however, for the hypothesis that people can feel the future with emotionally valenced nonerotic stimuli, with a Bayes factor of about 40. Although this value is certainly noteworthy, we believe it is orders of magnitude lower than what is required to overcome appropriate skepticism of ESP”. A Bayes factor of 40 in this case means that Bems’ data are 40 times more likely to have occurred under the hypothesis that “feeling the future” is true than under the hypothesis that it is not true. It is curious to note that for the supporters of the Bayesian approach, a Bayes factor of 30 is considered as strong support to the hypothesis.
This continued enhancement of the criteria of scientific evidence to accept parapsychological phenomena, is quite interesting. For example, as for all other lines of research, a given phenomenon is considered more probable if it is observed by independent replications. In a recent meta-analysis of all experiment related to “feeling the future”, we found 90 studies overall, conducted by Bem and thirty others, that elicited very strong support all of Bem’s original experiments with the exception of the retroactive recall facilitation tasks. However, yet again, this evidence was deemed insufficient by skeptics, in particular by Wagenmakers (Wagenmakers, 2014), in part because most of these experiments were not pre-registered and hence it cannot be excluded that the experiments are contaminated by flawed procedures and statistical analyses. Also, recently, Daniel Lakens has provided a critique based on our assumptions of the file drawer:http://daniellakens.blogspot.nl/2015/04/why-meta-analysis-of-90-precognition.html. We are in the process of providing a robust argument to his conclusions, but it is worth noting that if we were to apply such a stringent approach to existing meta-analyses related to psychological investigations, most would be viewed as suspect.
Final reflections: Though we do not believe there is a file-drawer problem, we are in the process of setting up collaborative pre-registered replications so this criticism will not apply to future experiments.
Bem, D. J. (2011). Feeling the future: experimental evidence for anomalous retroactive influences on cognition and affect. Journal of Personality and Social Psychology, 100(3), 407-425.
Bem, D., Utts, J., Johnson, W.O. (2011). Must psychologists change the way they analyze their data?Journal of Personality and Social Psychology, 101(4), 716-719.
Honorton, C., and Ferrari., D. (1989) “Future Telling” A meta-analysis of forced choice precognition experiments, 1935 – 1987.
Journal of Parapsychology, Vol. 53. December 1989.
Mossbridge, J., Tressoldi, P., and Utts, J. (2012) Predictive physiological anticipation preceding seemingly unpredictable stimuli: a meta-analysis.
Frontiers in psychology. 17th October 2012.
Rouder, J. N., & Morey, R. D. (2011). A Bayes factor meta-analysis of Bem’s ESP claim. Psychonomic Bulletin & Review, 18(4), 682-689.
Wagenmakers, E. J., Wetzels, R., Borsboom, D., & van der Maas, H.(2011). Why psychologists must change the way they analyze their data:The case of psi: Comment on Bem (2011). Journal of Personality and Social Psychology, 100, 426–432.