Am I Luckier Than You? Individual Differences in Precognition Performance

Abstract

The results found in parapsychological research have not been particularly persuasive, in large part due to a lack of replicability of those studies purporting evidence in support of the existence of psi phenomena. I propose that a more promising avenue of research into psi focuses on the potential correlates of psi performance (i.e. factors that correlate with above- or below-chance performance on luck-based tasks without any known causal mechanism). Specifically, individual differences, such as belief in ESP or extraversion, have been shown to correlate with psi performance in forced- choice precognition experiments (where participants predict a future chance- event by choosing one option from a limited number of options). Thus, the goal of this thesis was to synthesise existing literature on predictors of psi performance, to identify the best predictors of psi performance, and to test these predictors using the latest experimental paradigms while also examining its external validity in the real world.

My first study (Study I) was a never-before-conducted meta-analysis of all research that has looked at individual differences and psi performance in forced-choice precognition experiments. Overall, 57 studies published between 1945 and 2016, including 35 individual difference measures, were subject to meta-analysis (Zdrenka & Wilson, 2017). Six individual difference measures, namely, luck belief (the belief that luck is primarily controllable), perceptual defensiveness, openness to experience, belief in ESP, extraversion, and time belief as dynamic, were found to significantly correlate with psi performance.

Study II attempted to examine these promising individual difference variables using one of Bem’s (2011a) previously successful experimental paradigms, while also using Bem’s own materials and resources. Thus, Study II is a replication of the first of his experiments, Precognitive Detection of Erotic Stimuli, which was shown to be the most robust of the studies reported by Bem (Bem, Tressoldi, Rabeyron, & Duggan, 2015). In summary, Study II explored precognition using this most robust experimental paradigm, in combination with the most promising individual difference correlates identified in Study I. Belief in ESP was found to significantly predict psi performance (i.e. a ‘sheep- goat’ effect; Schmeidler, 1943, 1945) whereas the other individual differences did not.

The aims of Studies III(A) and III(B) were to investigate the utility of paranormal belief as a predictor of performance in two real-life lottery studies. In the first Study III(A), a large sample of New Zealanders completed a measure of paranormal belief and provided their preferred lotto numbers. Paranormal belief (and a range of other lottery behaviours) failed to predict lottery success. Finally, a pilot study is presented (Study III(B)) incorporating the same individual difference measures from Study I and II, but applied in a real-world setting. We examined participants’ predictions of numbers drawn in the New Zealand lottery over eight weeks, as the lottery presents a real-world example of forced-choice precognition (in that the future event is ostensibly unpredictable and there are a limited number of choices to pick from, i.e. the numbers 1-40). However, there was a low rate of participation that was caused, in part, by the time-commitment required, and therefore the study was underpowered and the results inconclusive. Limitations were also identified by potential preference biases distorting the results (due to there being only a single target set per draw, i.e. the winning lotto number set).

Overall, this thesis attempted to answer two main research questions: (1) which individual difference measures are the most predictive of psi performance in the forced-choice precognition domain? (Studies I) and, (2) are these individual difference predictors consistent and reliable across a range of settings? (Studies II through III(B)). I found partial support for a significant predictor, in that Belief in ESP significantly correlated with psi performance in a forced-choice precognition replicatory experiment, but further research is necessary to confirm whether this finding—based on experimental laboratory- based studies—translates to real-world psi, and to address the limitations found in Studies II and III. An alternative interpretation is that the results are due to random variation—given that they are weak and inconsistent—and psi does not exist. Pre-registered studies and prospective meta-analyses are likely to determine which interpretation prevails in the end.