Measuring precognitive effects using a fast implicit and fast explicit task

There has been a dramatic growth in the use of modified cognitive paradigms to test whether priming or practice in the future can influence performance in the present. This has led some to suggest that fast implicit type tasks are more effective as they rely less on conscious cognitive processes. However, this view is neither consistent nor clear. For instance, Bem (2011) reported more robust precognitive effects using a slower explicit recall task (i.e., Expts. 8 & 9). Hence, the aim of this study was to conduct two fast thinking tasks, one that relies predominantly on implicit processes and one that incorporates more explicit processes and examine which of these two tasks elicits the most robust precognitive effect. The fast thinking implicit task selected was a speeded version of the precognitive preference task using erotic images which involves presenting the participant with two hidden target locations on a screen, usually depicted by two curtained areas, and requiring them to select the location with the hidden target picture behind it. The fast thinking explicit task was an explicit precognitive recognition task. In essence, this is a standard explicit recognition task with an additional ‘precognitive’ twist. The traditional recognition task has a study phase, during which stimuli are initially presented, followed by a test phase when a selection of the original stimuli along with new unseen items are presented and the participant needs to recognise the ‘old’ (i.e., seen before) and ‘new’ (i.e., not seen before) items. The precognitive twist is that following on from the test phase there will be a post-test practise phase during which half of the ‘old’ items will be presented again with an emphasis on practise and re-processing. The study was pre-registered at the Koestler Unit (https://koestlerunit.wordpress.com/study-registry/) reference #1036.

A total 166 of participants completed the implicit preference task (20 Male and 145 Female, with 1 failing to provide information on gender), aged 18y to 48y (M:20.8y SD: 4.8y), and one hundred and fifty nine participants completed the explicit recognition task (21 Male and 137 Female, with 1 failing to provide information on gender), aged 18y to 46y (M:20.9y SD: 4.3y).

Response time data was initially cleaned by removing any outliers using a low cut-off of 200ms or above 2.5 Sd from the mean, in a non-recursive manner (see Van Selst & Jolicoeur, 1994). For response times and accuracy, the parametric assumption of normality was checked using both the Shapiro-Wilk test and values of skewness and kurtosis (e.g., DeCarlo 1997; Field, 2013; Razali & Wah, 2011). If all were violated non-parametric tests were used. Data from the implicit preference task and the explicit recognition task were analysed separately and all statistical tests were two-tailed.

For the implicit preference task the first confirmatory hypothesis tested whether participants would correctly identify the location of an Erotic image in less time than a Neutral image. Analysis using a Wilcoxon non-parametric test showed no difference in median response times between Erotic and Neutral images, Z=-0.517, p=0.61, r=-0.02. The second confirmatory hypothesis tested whether participants would be more accurate at identifying the location of Erotic images compared to chance (i.e., 50%). A one sample t test comparing accuracy of responses to the Erotic images to chance showed no significant difference, t(165)=0.363, p=0.717, 95% CI (-0.21, 0.31), d=0.02.

For the explicit recognition task the first confirmatory hypothesis tested whether participants would correctly recognise words that would be repeated later (i.e., Precognitively) in less time than those not repeated. Analysis using a repeated measures t test showed no difference in response times between Repeated and Not Repeated conditions (1073.8ms and 10859.9ms respectively), t(158)=1.212, p=0.227, 95% CI (-42.42, 10.15), d=0.06. The second confirmatory hypothesis tested whether participants would be more accurate (i.e., exhibit greater sensitivity) at recognising words which would be repeated later (i.e., Precognitively) compared to those not repeated. Analysis using a Wilcoxon non-parametric test showed no difference in median sensitivity levels between Repeated and Non-Repeated words, Z=-0.4561, p=0.65, r=-0.02.

Overall data from the implicit preference task and the explicit recognition task show no evidence of any precognitive effects. This raises the issue of whether there is no effect to find or whether this study simply failed to elicit them.