Content uploaded by Marco Zdrenka
Author content
All content in this area was uploaded by Marco Zdrenka on Jan 04, 2019
Content may be subject to copyright.
Journal of Parapsychology, 81(1), 9-32
INDIVIDUAL DIFFERENCE CORRELATES
OF PSI PERFORMANCE IN FORCED-CHOICE PRECOGNITION
EXPERIMENTS: A META-ANALYSIS (1945–2016)1
By Marco Zdrenka and Marc Stewart Wilson
ABSTRACT: Previous research in parapsychology has not been particularly persuasive, in large part due to
a lack of replicability of signicant ndings. To address these concerns and better understand which factors
may be associated with stronger and more consistent effect sizes, all forced-choice precognition experiments
analysing individual differences (e.g., personality traits) were aggregated to determine which factors might
reliably predict psi performance. Overall, 55 studies published between 1945 and 2016, including 35 individual
difference measures, were subject to meta-analysis. Six individual difference measures, namely, luck belief
(the belief that luck is primarily controllable), perceptual defensiveness, openness to experience, belief in psi,
extraversion, and time belief as dynamic, were found to signicantly correlate with psi performance. Given
the particularly straightforward nature of forced-choice precognition experiments, a promising future avenue
would be to explore these factors in conrmatory studies. It is hoped that researchers can model their future
experiments off these ndings in conjunction with preregistration techniques, to ultimately create a more sys-
tematic and robust database.
Keywords: meta-analysis, psi, precognition, personality, individual differences, forced-choice
Statistician Jessica Utts (1991) made the statement that a “promising direction for future pro-
cess-oriented research [in parapsychology] is to examine the causes of individual differences in psychic
functioning” (p. 377). It seems that such an approach is not only reasonable but also necessary, given that
the evidence for psi (a general term used to describe anomalistic communication or interaction with the en-
vironment) is often inconsistent and elusive (Kennedy, 2001). If psi is to be taken seriously by the scientic
community, its nature needs to be observable under prespecied conditions (Alcock, 2003; Hyman, 2010).
Individual difference factors—such as specic personality traits (e.g., extraversion) or beliefs (e.g., belief in
psi)—have been extensively analysed and thus represent a promising avenue in this regard. However, many
researchers ignore individual difference factors (potentially missing important sources of between-individ-
ual variation in psi performance) or look at a multitude of varied factors that are difcult to sort through. An
actual effect may also be masked if an individual difference factor is systematically related to psi perfor-
mance. For example, participants who score high in a trait may overperform while participants who score
low in that trait may underperform, effectively cancelling each other out. Therefore, this meta-analysis was
intended to synthesise the relevant research to better understand the factors that may lead to a successful (or
unsuccessful) and consistent demonstration of psi in the laboratory.
This meta-analysis focuses specically on forced-choice experiments that have tested for precogni-
tion (i.e., the foreknowledge of an event without any known explanation). Forced-choice experiments give
participants several options to choose from for their response, whereas free-choice experiments allow partic-
ipants to make an unrestricted response. As free-choice experiments—such as the ganzfeld—have received
a lot of attention in recent literature (see Bem & Honorton, 1994; Milton & Wiseman, 1999; Storm, 2006),
this meta-analysis focuses exclusively on forced-choice experiments. It also focuses on precognition rather
than telepathy (anomalous communication between people) or clairvoyance (perception without using nor-
mal sensory modalities), as precognition experiments are less susceptible to sensory leakage (Steinkamp,
2005). For example, in some telepathy experiments, participants may potentially make decisions based on
the sender’s or experimenter’s facial cues, but this is not possible in precognition experiments in which the
1 An earlier version of this paper was presented at the 10th Biennial European Conference of the Society for Scientic
Exploration in Sigtuna, Sweden, 2016.
The Journal of Parapsychology
10
target cannot be known (even by the experimenter) until after the participant has already made their choice.
Previously, there have been two meta-analyses conducted on individual differences in psi laborato-
ry research: one looking at extraversion (Honorton, Ferrari, & Bem, 1998) and the other looking at belief in
psi, or what is known as the sheep-goat effect (Lawrence, 1993). Both meta-analyses found a relationship
with psi performance (r = .09 and r = .03, respectively), indicating a small yet robustly signicant overall
effect size. However, many other individual differences have been examined in individual studies and it
would be useful to summarise those studies here, and to compare them all with one another. Furthermore,
many meta-analyses combine studies from multiple domains, making it difcult to unpack exactly what
factors constitute a replicable psi experiment. Thus, the purpose of this meta-analysis is to (a) provide
a comprehensive and updated review of all forced-choice precognition experiments that have included
individual difference measures, and (b) estimate the overall magnitude of the relationship between each
individual difference measure and psi performance, with the overall goal to provide researchers with the
necessary information needed to design conrmatory studies.
Meta-Analysis
Although some meta-analyses have focused on only one research paradigm, such as the ganzfeld
(Hyman, 1985) or biological systems (Schmidt, Schneider, Utts, & Walach, 2004), this is the rst meta-anal-
ysis to combine both an experimental paradigm and individual differences. Forced-choice precognition is
the chosen paradigm, as it is the most efcient method available for replication; the experiments are often
automated (less potential for interference from both participants and experimenters); and exact probabilities
of hits/misses can be objectively calculated. Although free-choice experiments can also be quantied, this
requires an additional step, as participant responses need to be converted to target responses. This is avoided
in forced-choice experiments altogether.
Although forced-choice precognition experiments might seem too narrow a subset to analyse,
Steinkamp (2005) reviewed all forced-choice extrasensory perception (ESP) experiments—including telep-
athy, clairvoyance, and precognition—and found it difcult to come to any conclusions due to conicting
outcomes and wide variations in study designs. Furthermore, whereas some studies do not show any dif-
ferences in effect sizes between precognition and other domains (see Steinkamp, Milton, & Morris, 1998),
other studies have found a difference between clairvoyance, precognition, and telepathy effect sizes (Storm,
Tressoldi, & Di Risio, 2012; Tart, 1983). Therefore, in dening the inclusion criteria narrowly, we sought
to overcome the heterogeneity of studies in Steinkamp’s (2005) review.
For the purposes of this meta-analysis, the effect size of interest is the correlation coefcient be-
tween the individual difference measure and psi performance—not psi performance specically—with the
participant as the unit of analysis.
Method
Retrieval of Studies
Only studies in the published literature are included in the meta-analysis, since parapsychology is
a relatively small eld and it is unlikely that there are many unpublished dissertations or theses (Honorton
& Ferrari, 1989). Sourcing of relevant studies included the bibliography of two meta-analyses (Honorton &
Ferrari, 1989; Storm et al., 2012), a database search (described below), along with an inspection of all En-
glish-language parapsychological journals, namely, the Journal of Parapsychology, Journal of the American
Society for Psychical Research, Journal of the Society for Psychical Research, Research in Parapsychology,
Australian Journal of Parapsychology, European Journal of Parapsychology (including the Research Letter
of the Utrecht University Parapsychology Laboratory), and the Journal of Scientic Exploration.
An exhaustive search was conducted of research databases including PsycINFO, Google Scholar,
WorldCat, and LexScien, using the keywords “individual differences,” “precognition,” “parapsychology,”
“forced-choice,” “retrocausation,” “retrocausality,” “psi,” “ESP,” and “extrasensory perception.” Most of
Individual Difference Correlates: A Meta-Analysis 11
these searches located studies that were already found in the journals listed above.
The search period was intended to capture all experimental psi research published from 1945
through 2016 that included individual difference measures.
The search strategy revealed 35 individual difference variables including more common measures
such as extraversion and belief in psi, along with less widely used measures such as temporal lobe dysfunc-
tion and latent inhibition.
Selection Criteria
Studies were included from 1945 until 2016 if they met the following criteria:
1. Forced-choice design
2. Precognition design
3. Included individual difference measure(s)
4. A minimum of two human participants
Studies that did not include relevant information were excluded. For example, the results reported
in Wiseman and Greening (2002) could not be included as their precognition and clairvoyance data were
combined when reporting individual difference measures (e.g., the sheep-goat effect), and results reported
by Steinkamp (1998) could not be used as the number of participants was not reported.
The identication, screening, and eligibility of the studies followed the Preferred Reporting Items
for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Liberati et al., 2009). Figure 1 provides
a detailed summary of the database search and screening process.
Figure 1. A ow diagram illustrating the database search and screening stages involved in this meta-analysis.2
2 In both the screening stage and the eligibility stage, a number of articles (73 and 64, respectively) were excluded as
they did not measure individual differences and were thus not relevant to this meta-analysis. The only difference is that
the articles identied for exclusion in the screening stage were deemed irrelevant after only screening the titles, ab-
stracts, and keywords, whereas the others required a more thorough reading of the article to make that determination.
Records identified through
database searching
(n = 242)
Screening
Included
Eligibility
Identification
Additional records identified
through other sources
(n = 4)
Records after duplicates removed
(n = 218)
Records screened
(n = 218)
Records excluded
(n = 73)
Full-text articles assessed
for eligibility
(n = 145)
Studies included in
meta-analysis
(n = 45)
Full-text articles excluded,
with reasons (n = 100):
• Individual differences not
measured (n = 64)
• Case study (n = 19)
• Animal subjects (n = 11)
• Relevant information not
included (n = 6)
The Journal of Parapsychology
12
Denitions
Independent investigator. For the purposes of this meta-analysis, an “investigator” refers to the
lead author of a study. An investigator is considered independent of another investigator if both investiga-
tors have never worked on a paper together using the database being analysed (and/or worked with the other
investigator’s co-authors). The number of independent investigators can be helpful in determining how
replicable an effect might be; the fewer independent investigators there are—even if the studies have been
repeated multiple times—the less certain we can be that the results are replicable. As Hyman (1977) notes,
it is not enough to simply repeat the same results, for whatever errors or biases may have occurred in the
rst instance might also be part of the subsequent repetitions of the experiment.
Individual differences. For the purposes of this research, an individual difference is dened as
anything that an individual may psychologically vary on, whether it is personality, beliefs, intelligence, or
aptitude (Nazimuddin, 2015). However, this meta-analysis makes a distinction between individual differ-
ences that are relatively constant regardless of situational factors (e.g., trait-level individual differences)
and those that are more temporary and can be affected by the experimental situation (e.g., state-level indi-
vidual differences such as participants’ mood in the experiment); the latter (e.g., classication of a partici-
pant high in trait-anxiety but low in state-anxiety in a particular experimental setting) will not be included
in this paper to reduce confusion while also limiting the number of variables analysed.
Individual differences were further categorised, combining similar measures (or subcomponents)
into families of similar constructs. Where multiple measures of a single individual difference were used in
a study, only the most appropriate measure was used. For example, some studies included both a sheep-goat
measure and an interest in psi measure—in these cases, only the sheep-goat measure was included, as it is
has historically been more consistently used as a measure of psi belief (Lawrence, 1993).
Procedural Features
As there are a multitude of individual difference measures included (35 individual difference con-
structs were identied for analysis), a separate meta-analysis was performed on each category of individual
difference. The majority of these meta-analyses contain less than ve studies in total, so it was not practical
to code for procedural features.
Quality coding of the studies was not implemented for four reasons. Firstly, up until 1976, the
founder of experimental parapsychology, J. B. Rhine, encouraged less detail in publications for nonsig-
nicant parapsychological ndings than signicant ndings (Steinkamp, 2005). Therefore, quality coding
would inevitably favour newer studies, since more information is available for post-1976 studies (which
may not correlate with the actual quality of the experiment). Secondly, precognition experiments have
less potential for procedural defects compared with parapsychological research more generally, which is
reected by Honorton and Ferrari (1989) having only 6 quality criteria for precognition experiments rather
than Rhine, Pratt, Stuart, Smith, and Greenwood’s (1940) 34 quality criteria for ESP experiments. Thirdly,
Honorton and Ferrari (1989) did not nd a relationship between forced-choice precognition experiments
and their quality ratings. Lastly, the majority of these meta-analyses had too few studies in total to mean-
ingfully differentiate them on quality.
However, year of publication was coded, as it allowed us to examine whether effect sizes have
increased over time, stayed the same, or even decreased. Honorton and Ferrari (1989) suggest that if effect
sizes do not increase over time—as they found in their meta-analysis—it might mean that researchers lack
an understanding of the underlying factors of psi performance (since they could not reliably increase its
magnitude over time). Alternatively, an increase in effect size over time would be more promising, as was
reported in Storm et al.’s (2012) meta-analysis.
Meta-Analysis of Correlation Coefcients
All indices of association between an individual difference measure and psi performance were con-
verted to correlation coefcients using Comprehensive Meta-Analysis (CMA) software version 2 (Boren-
Individual Difference Correlates: A Meta-Analysis 13
stein, Hedges, Higgins, & Rothstein, 2005) or manually. For example, t tests were converted to point-bise-
rial correlations and phi coefcients were computed from 2 x 2 contingency tables. Some studies gave only
trial-based data such as the critical ratio (z; e.g, Buzby 1967; Freeman & Nielsen, 1964). In these instances,
correlations were estimated using a method for estimating effect sizes from critical ratios described by Mc-
Carthy and Schechter (1986), providing an estimate of Cohen’s d—this was then converted to the r metric.
Unreported correlations were estimated using the provided p values, whereas studies that reported only
nonsignicance had their correlation set to .00,3 a practice consistent with the approach adopted by Honor-
ton et al. (1998) in their meta-analysis of extraversion and ESP performance. Where necessary, correlation
signs were adjusted to reect the appropriate relationship between the individual difference measure and psi
performance. Finally, CMA weighted each study—using a random effects model incorporating both sam-
ple size and between-study variance—giving an overall outcome metric (r) for each individual difference
measure in the database. A random effects model was used rather than a xed effects model, as most stud-
ies were not exact replications of each other and this model takes into account such variation (Borenstein,
Hedges, Higgins, & Rothstein, 2010). All p values are two-tailed.
Heterogeneity tests using Cochran’s Q were also conducted on each meta-analysis to determine
whether results from the included studies were representative of a single homogenous effect.4 For those
meta-analyses showing heterogeneity, moderator analyses were conducted using the year of publication
as a proxy for methodological quality. The I2 index was also reported, to give an idea of the degree of het-
erogeneity present. Finally, Rosenthal’s (1979) fail safe N, or the le drawer estimate, was calculated for
all meta-analyses that showed statistical signicance to determine how many unreported studies averaging
null results would need to exist for the effect to be reduced to overall nonsignicance. If the number is high,
then there is less likelihood for publication bias, that is, studies being reported only if they show statistical
signicance (Honorton & Ferrari, 1989). Because unreported nonsignicant studies may have an average
effect size below zero (Ferguson & Heene, 2012), an alternative method for examining publication bias,
namely Egger’s regression method, was also included for these studies (Egger, Smith, Schneider, & Minder,
1997). Egger’s test aims to quantify potential asymmetrical distributions of studies around the mean effect
size (Rothstein, 2008).
Results
Descriptives
Overall, this meta-analysis is comprised of 55 individual studies, which were reported in 45 papers
and conducted by 17 independent investigators. The studies span a total of 71 years, between 1945 and
2016. There were a total of 17,584 participants analysed, with sample sizes ranging from 13 to 13,941. In
the majority of these studies, students were the sample population.
Separate meta-analyses of the relationship between psi performance and each category of individ-
ual differences are reported below. It is ordered in terms of those variables that have the most exemplar
studies (from belief in ESP, the Big Five, various operationalisations of luck, which have the most studies)
through to variables for which there are only two or three studies (e.g., religiosity, emotional reactivity,
intelligence).
For a summary of the total number of studies, independent investigators, and participants, see Table 2.
Major Individual Difference Measures
3 However, this is an estimate, as the mean of the distribution of all possible nonsignicant outcomes is likely to be less
than zero after removing the outcomes that give signicant results by chance (as it effectively removes or truncates the
right tail of the distribution). See the Appendix, Table A1, for all studies that this applies to.
4 Note that these tests were conducted for all meta-analyses, even when there were only two or three studies, as
it would be hard to justify an arbitrary cut-off point. However, it does not imply that all of these tests should be
given equal weight. The heterogeneity analyses conducted on a limited number of studies should not be considered
denitive.
The Journal of Parapsychology
14
Belief in ESP. Overall, belief in ESP was the most studied potential individual difference correlate
in forced-choice precognition experiments, having been reported in 22 studies by 12 independent investi-
gators based on a total of 2,200 participants. The most common measurement questionnaire was a variant
of Schmeidler’s (1943) sheep-goat criterion, such as Thalbourne and Delin’s (1993) Australian Sheep-Goat
Scale or Bahdra’s (1966) Sheep-Goat Questionnaire. In general, participants who score high on these scales
are classied as “sheep” (or believers in ESP) and those who score low are classied as “goats” (or disbe-
lievers in ESP).
Figure 2 shows a forest plot of the correlation coefcients, with the correlations ranging from
-.17 to .72. The overall mean weighted effect size (r) is .13 (p = .002), with a 95% condence interval
between .05 and .20. This suggests that there is a small but signicant relationship between psi belief and
performance on a psi task, such that people who believe in psi tend to perform better than those who do not
believe in psi. This effect size is slightly larger than the effect size reported by Lawrence (1993) in his me-
ta-analysis on the sheep-goat effect (r = .03), but that also included telepathy and clairvoyance experiments.
Study na me Stati sti cs f or each st udy Correlation and 95% CI
Lo wer Upper
Correlation limit limit Total
Schmei dler , 1945 0.04 -0.19 0.26 75
Buzby, 1967 0.30 0.17 0.42 202
Ryzl, 1968a 0.29 0.04 0.50 63
Ryzl, 1968b 0.31 -0.09 0.63 25
Johnson, 1969 0.62 0.33 0.81 28
Nielsen, 1970 0.72 0.28 0.91 13
Honorton, 1972 0.00 -0.37 0.37 28
Haraldss on, 1975 0.00 -0.13 0.13 223
Haraldss on, 1980 0.11 0.02 0.20 449
Thalbourne et al., 1982 (Exp 1) 0.00 -0.20 0.20 101
Thalbourne et al., 1982 (Exp 2) 0.30 0.09 0.48 86
Thalbourne, 1996 -0.06 -0.25 0.14 99
Haraldss on et al ., 2002 0.06 -0.22 0.33 50
Storm, 2006 0.10 -0.04 0.24 199
Wil son & Hamlin, 2007 (Exp 1) 0.02 -0.24 0.27 61
Wil son & Hamlin, 2007 (Exp 2) 0.00 -0.30 0.30 44
Storm, 2008 -0.17 -0.32 -0.01 149
Luke et al., 2008 0.24 0.04 0.41 100
Luke & Mori n, 2009 0.49 0.21 0.69 41
Palmer, 2009 0.00 -0.25 0.25 64
Hitchman et al., 2012 -0.03 -0.31 0.25 50
Hitchman et al., 2015 (Combi ned) 0.01 -0.27 0.28 50
0.13 0.05 0.20 2200
-1.00 -0.50 0.00 0.50 1.00
Figure 2. Meta-analysis of the relationship between belief in ESP and psi performance in forced-choice
precognition experiments.
However, a test of heterogeneity was signicant (Q = 61.16, p < .001) which suggests that variation
in results may be due to factors other than the relationship between psi belief and performance (for example,
error, or the inuence of a moderator). The I2 was 66%. Consequently, a mixed effects model (method of
moments) meta-regression was conducted, which found year of publication to be a signicant moderator,
QR = 6.71, p = .01. Figure 3 shows effect sizes to decrease as year of publication increases. This means that
some of the heterogeneity that caused the signicant Q can be attributed to the year of publication.
Finally, the fail safe N, or the number of unreported studies averaging null results that would be
needed to bring the p value to nonsignicance, is 141. Egger’s test was not found to be signicant, t(20) =
1.21, p = .24.
Individual Difference Correlates: A Meta-Analysis 15
Figure 3. Meta-regression on the relationship between belief in ESP and psi performance in forced-choice
precognition experiments, using publication date as the moderator.
The Big Five
The next set of meta-analyses cover personality indicators consistent with the Big Five personality
traits (McCrae & Costa, 1987): extraversion (how outgoing and social a person is), neuroticism (a long-
term tendency to be in a negative emotional state, such as anxious or frustrated), openness to experience
(intellectually curious, willing to try new things, and imaginative), agreeableness (how kind and sympathet-
ic a person is), and conscientiousness (organised and diligent).
Extraversion. Extraversion was the second most studied individual difference measure in forced-
choice precognition experiments, having been reported in 14 studies by seven independent investigators
with a total of 1,206 participants. Extraversion was typically measured within larger personality question-
naires such as the 16PF (Cattell & Mead, 2008), but subcomponents such as Bem’s (2011) Sensation Seek-
ing Scale were also included. High scorers on these measures are generally considered to be extraverted
and low scorers introverted.
Figure 4 shows a forest plot of the correlation coef cients, with the correlations ranging from -.28
to .35. The overall mean weighted effect size (r) is .08 (p = .02), with a 95% con dence interval between
.01 and .15. This suggests that there is a small but signi cant relationship between extraversion and psi
performance, such that people who are extraverted tend to perform better than those who are more intro-
verted. This result is consistent with previous studies that have also found a positive relationship between
extraversion and psi performance (Mangan, 1958; Palmer, 1978; Honorton et al., 1998). Furthermore, a test
of heterogeneity was not signi cant (Q = 17.23, p = .19), with an I2 index of 25%.
Finally, the fail safe N, or the number of unreported studies averaging null results that would be
needed to bring the p value to nonsigni cance, is nine, whereas Egger’s test was not signi cant , t(12) =
0.56, p = .59.
Neuroticism. Neuroticism was measured using a variety of different questionnaires encompassing
anxiety, affect, and mood, and was included in nine studies by seven independent investigators and a total
of 528 study participants. The correlation coef cients range from -.38 (Humphrey, 1945) to .60 (Freeman
& Nielsen, 1964). The overall mean weighted effect size (r) is .05 (p = .43), with a 95% con dence interval
between -.08 and .19. The results are inconclusive about whether an actual effect occurs, falling slightly
short of Steinkamp’s (2005) suggestion that neuroticism was a promising predictor of ESP forced-choice
experiments.
The Journal of Parapsychology
16
Study name Statistics for each study Correlati on and 95% CI
Lowe r Upper
Correlati on limit limit Total
Humphrey, 1945 0.35 -0.12 0.69 19
Nielsen, 1970 0.00 -0.55 0.55 13
Thalbourne et al., 1982 (Exp I) 0.00 -0.20 0.20 101
Thalbourne et al., 1982 (Exp II) 0.00 -0.21 0.21 86
Storm & Thalbourne, 1989-99 0.28 0.08 0.46 93
Storm & Thalbourne, 2001 0.00 -0.19 0.19 107
Haraldsson et al., 2002 -0.03 -0.31 0.25 50
Storm, 2002 -0.03 -0.33 0.27 43
Wilson & Hamlin, 2007 -0.28 -0.53 0.02 44
Luke et al., 2008 0.16 -0.03 0.35 100
Bem, 2011 (Exp 1) 0.18 -0.02 0.36 100
Bem, 2011 (Exp 2) 0.17 0.01 0.32 150
Wagenmakers et al., 2012 0.13 -0.07 0.32 100
Thalbourne & Storm, 2014 0.04 -0.10 0.18 200
0.08 0.01 0.15 1206
-1.00 -0.50 0. 00 0.50 1.00
Figure 4. Meta-analysis of the relationship between extraversion and psi performance in forced-choice precog-
nition experiments.
However, a test of heterogeneity was signicant (Q = 17.29, p = .03) which suggests that there were
potential moderating factors in this database. The I2 was 54%. A mixed effects model (method of moments)
meta-regression did not nd year of publication to be a signicant moderator, QR = 0.30, p = .58. Due to
meta-regression analyses not being recommended for meta-analyses with less than 10 studies (Borenstein,
Hedges, Higgins, & Rothstein, 2009), this nding should be treated with caution.
Openness to experience. Openness to experience was reported in nine studies of 522 participants, by
ve independent investigators. The most common measurement questionnaire was the Openness to Experience
Scale (Goldberg, 1999). Figure 5 shows a forest plot of the correlation coefcients, with the correlations rang-
ing from -.08 to .46. The overall mean weighted effect size (r) is .12 (p = .02), with a 95% condence interval
between .02 and .22, indicating a small but signicant relationship between openness to experience and psi per-
formance, such that people who prefer new experiences tend to perform better than those who prefer familiar
routines. Furthermore, a test of heterogeneity was not signicant (Q = 11.56, p = .24), with an I2 index of 22%.
Study name Statistics for each study Correlati on and 95% CI
Lowe r Upper
Correlati on limit limit Total
Storm & Thalbourne, 1998-99 0. 00 -0. 20 0.20 93
Haraldsson et al., 2002 0. 16 -0. 12 0.42 50
Wilson & Hamlin, 2007 0.18 -0.12 0. 45 44
Luke et al . , 2008 0.46 0.13 0.70 32
Luke & Morin, 2009 -0.08 -0.38 0. 23 41
Palmer, 2009 0.27 0.02 0. 49 61
Hitchman et al., 2012 0. 29 0. 01 0. 53 50
Hitchman et al., 2015a 0. 01 -0. 27 0.29 49
Hitchman et al. , 2015b (Combined) -0.01 -0. 29 0. 27 50
Hitchman et al., 2016 0. 06 -0. 22 0.33 52
0.12 0. 02 0.22 522
-1.00 -0.50 0.00 0.50 1.00
Figure 5. Meta-analysis of the relationship between openness to experience and psi performance in forced-
choice precognition experiments.
Individual Difference Correlates: A Meta-Analysis 17
In this case, the fail safe N, or the number of unreported studies averaging null results that would be
needed to bring the p value to nonsignicance, is 12. Egger’s test was not signicant, t(8) = 1.30, p = .23.
Agreeableness. Agreeableness was reported in seven studies by four independent investigators
with a total of 556 participants. The most common measurement questionnaire was the Independence factor
of the 16PF (Cattell, 1996; Cattell & Mead, 2008). Correlation coefcients range from -.36 (Humphrey,
1945) to .23 (Storm & Thalbourne, 1998-1999), with an overall mean weighted effect size (r) of .02 (p =
.71) and a 95% condence interval between -.09 and .13. Although the results are inconclusive, the data
suggest that the true effect size is below .13 and could be zero. There was no indication of signicant het-
erogeneity (Q = 9.34, p = .16), with an I2 index of 36%.
Conscientiousness. Conscientiousness was the least studied Big Five personality trait in forced-
choice precognition experiments, having been reported in only three studies by three independent investiga-
tors with a total of 187 participants. Conscientiousness was measured as a component of larger personality
questionnaires such as the NEO Five-Factor Inventory (McCrae & Costa, 1987). Correlations range from
.00 (Storm & Thalbourne, 1998-1999) to .23 (Wilson & Hamlin, 2007). The overall mean weighted effect
size (r) is .06 (p = .45), with a 95% condence interval between -.09 and .20, which is inconclusive but
suggests that the effect size may lie below .20 and could be zero. Furthermore, a test of heterogeneity was
not signicant (Q = 1.69, p = .43). The I2 is 0%.
Beliefs About Luck
Although luck has been explored in multiple studies, Smith (1998) showed that “luck” can mean
different things to different people. Therefore, various measurements of luck and luck beliefs have been
reported in forced-choice precognition experiments that include not only perceived luckiness, but also con-
trollable luck belief, chance belief, providence belief, and fortune belief. The most common tool of mea-
surement used in these experiments was Luke, Delanoy, and Sherwood’s (2003) Questionnaire of Beliefs
about Luck, which incorporates all of these subcomponents together. They will now be discussed in turn.
Perceived luckiness. Perceived luckiness has (prior to 2008) been the standard measurement used
to explore luck in psi experiments and refers to how lucky one perceives oneself to be. For forced-choice
precognition experiments, perceived luckiness was reported in four studies by two independent investiga-
tors with a total of 231 participants. Correlations range from -.20 (Hitchman, Row, & Sherwood, 2012) to
.26 (Luke, Delanoy, & Sherwood, 2008), with an overall mean weighted effect size (r) of .08 (p = .49); 95%
CI [-.14, .28]. These results are inconclusive but suggest that the effect size is below .28 (and could be zero).
A test of heterogeneity was not signicant (Q = 7.09, p = .07), with an I2 index of 58%.
Luck belief. Luck belief refers to the belief that luck is primarily controllable, and participants who
score high in this belief also view luck as internal, stable, and nonrandom (Luke et al., 2003). Luck belief
was reported in ve studies by one independent investigator with a total of 248 participants. Figure 6 shows
a forest plot of the correlation coefcients, ranging from -.09 to .26. The overall mean weighted effect size
(r) is .13 (p = .048), with a 95% condence interval between .001 and .26, indicating a small but reliable
relationship between luck belief and psi performance, such that people who believe luck to be controllable
tend to perform better than those who see luck as uncontrollable. Furthermore, a test of heterogeneity was
not signicant (Q = 4.11, p = .39). The I2 is 3%. Finally, the fail safe N, or the number of unreported studies
averaging null results that would be needed to bring the p value to nonsignicance, is less than 1. However,
Egger’s test is not signicant, t(3) = 1.34, p = .27.
Chance belief. Chance belief refers to the belief that luck is random, unpredictable, unstable, and
inert (Luke et al., 2003). Chance belief was reported in ve studies by one independent investigator with
a total of 248 participants. Correlations range from -.16 (Luke et al., 2008) to .48 (Luke, Roe, & Davison,
2008). The overall mean weighted effect size (r) is .14 (p = .23), with a 95% condence interval between
-.09 and .36.
The Journal of Parapsychology
18
Study name Statistics for each study Correlati on and 95% CI
Lowe r Upper
Correlati on limit limit Total
Luke et al., 2008a 0.26 0.07 0.44 100
Luke et al., 2008b (Study 1) 0.14 -0.27 0.51 25
Luke et al., 2008b (Study 2) 0.12 -0.24 0.45 32
Luke & Morin, 2009 -0.09 -0.39 0.22 41
Hitchman et al. , 2012 0.04 -0.24 0.31 50
0.13 0.00 0.26 248
-1.00 -0. 50 0.00 0. 50 1.00
Figure 6. Meta-analysis of the relationship between luck belief and psi performance in forced-choice pre-
cognition experiments.
However, a test of heterogeneity was signicant (Q = 11.40, p = .02) which suggests that there were
potential moderating factors in this database. The I2 is 65%. A mixed effects model (method of moments)
meta-regression was conducted, but it did not nd year of publication to be a signicant moderator (QR =
1.75, p = .19).
Providence belief. Providence belief refers to the belief that luck is something that is managed
by external forces or higher beings (Luke et al., 2003). Providence belief was reported in ve studies by
one independent investigator with a total of 248 participants. Correlations range from -.09 (Hitchman,
Roe, & Sherwood, 2012) to .39 (Luke et al., 2008). The overall mean weighted effect size (r) is .12 (p
= .11; 95% CI [-.03, .27]). Furthermore, a test of heterogeneity was not signicant (Q = 5.34, p = .25).
The I2 is 25%.
Fortune belief. Fortune belief refers to the belief that luck is meant as a metaphor for life’s su
c-
cesses rather than as a literal event (Luke et al., 2003) and was reported in ve studies by one independent
investigator based on 248 participants. Correlations range from -.08 (Luke & Morin, 2009) to .15 (Luke et
al., 2008). The overall mean weighted effect size (r) is .03 (p = .62; 95% CI [-.10, .16]). A test of heteroge-
neity was not signicant (Q = 1.98, p = .74), and the I2 index is 0%.
Uncategorised Individual Difference Measures
Creativity. Creativity was reported in nine studies by three independent investigators with a total
of 506 participants. The most common measurement questionnaires were the short version of the Per-
sonal-Social Motivational Inventory (Torrance, 1963) and the Creative Cognition Inventory (Holt, 2002).
Correlations range from -.17 (Schmeidler, 1964c) to .20 (Luke et al., 2008), with an overall mean weighted
effect size (r) of .05 (p = .46; 95% CI [-.08, .17]). Furthermore, a test of heterogeneity was not signicant
(Q = 5.90, p = .21), with an I2 index of 32%.
Perceptual defensiveness. Perceptual defensiveness refers to psychological defence mechanisms
and is related to subliminal perception and preconscious processing. All reported studies administered the
Defense Mechanism Test (Kragh & Smith, 1970), with a total of six studies conducted by one independent
investigator and a total of 272 participants. The test incorporates a tachistoscopic technique using peripheral
stimuli to trigger subliminal anxiety and thereby defensive reactions. Figure 7 shows a forest plot of the cor-
relation coefcients, with the correlations ranging from -.04 to .30. The overall mean weighted effect size
(r) is .12 (p = .049; 95% CI [.001, .24]), suggesting a small but signicant relationship between perceptual
defensiveness and psi performance, such that people who exhibit high preconscious defensiveness tend to
perform better than those who do not. There was no evidence of heterogeneity (Q = 5.13, p = .40), with an
I2 index of 3%.
Transliminality. Transliminality is dened as “the hypothesised tendency for psychological ma-
terial to cross thresholds into or out of consciousness” (Thalbourne & Delin, 1994, p. 31), and was used in
ve studies by one independent investigator based on a total of 542 participants, with the most common
measurement questionnaire being the Transliminality Scale (Thalbourne, 1998). Correlations range from
Individual Difference Correlates: A Meta-Analysis 19
-.13 (Thalbourne, 1996) to .27 (Storm & Thalbourne, 1998-1999). The overall mean weighted effect size (r)
is .01 (p = .91), with a 95% condence interval between -.13 and .15. Although these results are inconclu-
sive, a test of heterogeneity was signicant (Q = 9.81, p = .04), but year of publication was not a signicant
moderator (QR = 1.29, p = .26). Due to the small number of studies, this nding should be treated with
caution (Borenstein et. al., 2009). The I2 is 59%.
Study na me Stati sti cs f or each s tudy Correlation and 95% CI
Lo wer Upper
Correlation limit limit Total
Haralds son, 1977 0.10 -0.23 0.41 37
Haralds son & Johnson, 1979 0.00 -0.31 0.31 41
Johnson & Haraldss on, 1984 (Study IV) 0.29 0.02 0.52 54
Johnson & Haraldss on, 1984 (Study V) 0.30 0.01 0.54 46
Haralds son & Johnson, 1986 0.04 -0.26 0.33 44
Haralds son et al., 2002 -0.04 -0.31 0.24 50
0.12 0.00 0.24 272
-1.00 -0.50 0.00 0.50 1.00
Figure 7. Meta-analysis of the relationship between perceptual defensiveness and psi performance in
forced-choice precognition experiments.
Pro attitude. Pro attitude refers to “an attitude that is favourably directed towards an outcome”
(Storm, 2002, p. 47) and three studies with 393 participants were conducted by a single investigator. All
used Thalbourne and Storm’s (2014) Pro Attitude Scale. Correlations range from -.16 (Storm, 2002) to .02
(Storm, 2008), with an overall mean weighted effect size (r) of -.02 (p = .70; CI 95% [-.12, .08]). A test of
heterogeneity was not signicant (Q = 1.04, p = .60), with an I2 of 0%.
Dream recall. Dream recall (specically, whether an individual recalls their dreams or not) was
reported in four studies by three independent investigators with a total of 799 participants. It was typical-
ly measured using a one-item questionnaire, which asked participants how frequently they recalled their
own dreams. Correlations range from .03 (Thalbourne, 1996) to .43 (Honorton, 1972). The overall mean
weighted effect size (r) is .07 (p = .23; 95% CI [-.04, .18]), and a test of heterogeneity was not signicant
(Q = 5.84, p = .12). The I2 is 49%.
Reports of unusual spontaneous experiences. Unusual spontaneous experiences can be described
as seemingly paranormal experiences in everyday life (as opposed to the experimental laboratory) and were
reported in three studies by two independent investigators based on a total of 695 participants. It was typi-
cally measured using a single question, which asked participants if they have had any precognitive dreams
(i.e., dreams that they thought predicted the future). Correlations range from -.21 (Schmeidler, 1964d) to
.00 (Haraldsson 1975, 1980), with an overall mean weighted effect size (r) of -.01 (p = .87; 95% CI [-.08,
.07]). A test of heterogeneity was not signicant (Q = 0.85, p = .65), with an I2 of 0%.
Religiosity. Religiosity was reported in two studies by two independent investigators with a total of
149 participants. In both studies, religiosity was measured using the Religiosity Scale (Haraldsson, 1993),
with the correlations ranging from -.13 (Thalbourne, 1996) to .08 (Haraldsson, Houtkooper, Schneider, &
Bäckström, 2002). The overall mean weighted effect size (r) is -.05 (p = .59; 95% CI [-.24 and .14]). A test
of heterogeneity was not signicant (Q = 1.34, p = .25). The I2 is 26%.
Emotional reactivity. Emotional reactivity is a measure of one’s emotional reaction to violent,
scary, or gruesome content in photographs, movies, and videos. All studies used the Emotional Reactivity
Scale (Bem, 2003). A total of three studies by one investigator looked at emotional reactivity, with 151
participants being included in the experiments. Correlations range from -.27 (Hitchman, Sherwood, & Roe,
2015) to .29 (Hitchman, Pfeuffer, Roe, & Sherwood, 2016), with an overall mean weighted effect size (r)
of .06 (p = .71; 95% CI [-.27, .38]). A test of heterogeneity was signicant (Q = 8.53, p = .01), with the year
of publication being a signicant moderator (QR = 3.96, p = .046). Figure 8 shows effect sizes to increase
as year of publication increases. The I2 is 77%.
The Journal of Parapsychology
20
Figure 8. Meta-regression on the relationship between emotional reactivity and psi performance in forced-
choice precognition experiments, using publication date as the moderator.
Temporal lobe dysfunction. Temporal lobe dysfunction measures symptoms of temporal lobe
damage such as disturbances of perception, selective attention of auditory input, and impaired organisation
of verbal material. A total of two studies by two independent investigators looked at temporal lobe dysfunc-
tion, using either the 13-item LIMBEX Scale or the Complex Partial Epileptic Signs cluster of the Personal
Philosophy Inventory (Persinger & Makarec, 1987). There were a total of 114 participants in all of the
experiments. Correlations range from -.01 (Hitchman, Roe, & Sherwood, 2015) to .00 (Palmer, 2009), with
an overall mean weighted effect size (r) of -.004 (p = .96; 95% CI [-.19, .18]). A test of heterogeneity was
not signi cant (Q = 0.003, p = .96), with an I2 of 0%.
Study name Statistics for each study Correlati on and 95% CI
Lowe r Upper
Correlati on limit limit Total
Schmeidler, 1945 (Combined) -0.04 -0.27 0.19 75
Schwartz & De Matt ei, 1983 -0.02 -0.03 -0.00 13941
-0.02 -0.03 -0.00 14016
-1.00 -0.50 0.00 0.50 1.00
Figure 9. Meta-analysis of the relationship between time belief as dynamic and psi performance in forced-
choice precognition experiments.
Time belief as dynamic. “Time belief as dynamic” was reported in two studies by two indepen-
dent investigators with a total of 14,016 participants and refers to how strongly one sees time as being in a
constant ow. In both studies, Knapp and Garbutt’s (1958) Time Metaphor Test was administered, which
measures the belief that time is metaphorically “a dashing waterfall” or “a fast moving shuttle,” for exam-
ple. Figure 9 shows a forest plot of the correlation coef cients, with the correlations ranging from -.04 to
-.02. The overall mean weighted effect size (r) is -.02 (p = .04; (95% CI [-.03, -.001]), suggesting a small
but signi cant negative relationship between time belief as dynamic and psi performance, such that people
who do not view time as dynamic tend to perform better than those who see time as dynamic and hasty.
Individual Difference Correlates: A Meta-Analysis 21
Furthermore, a test of heterogeneity was not signicant (Q = 0.05, p = .82), and the I2 is 0%. The fail safe N
was not calculated, as it is impractical to do so with less than three studies.
Future-orientation. Future-orientation refers to being more attentive towards future events than
past events and was reported in two studies by two independent investigators based on a total of 118 par-
ticipants. Both studies used the Attitude Toward the Future Questionnaire (Vaughan & Houck, 1993). Cor-
relations range from -.04 (Haraldsson et al., 2002) to .12 (Vaughan & Houck, 1993), and the overall mean
weighted effect size (r) is .05 (p = .57; 95% CI [-.13, .23]). A test of heterogeneity was not signicant (Q =
.65, p = .42), with an I2 of 0%.
Intelligence. Intelligence was measured in two studies by two independent investigators with a
total of 80 participants. Both studies included only child participants, ranging in age from 4 to 14. The
measurement tools used were the Peabody Picture Vocabulary Test (Dunn & Hottel, 1961) and a formal
mathematical ability test using basic addition, subtraction, multiplication, and division problems. Correla-
tions range from -.23 (Winkelman, 1981) to .04 (Drucker, Drewes, & Rubin, 1977), with an overall mean
weighted effect size (r) of -.07 (p = .62; 95% CI [-.31, .19]). A test of heterogeneity was not signicant (Q
= 1.2, p = .27), and the I2 is 19%.
Single Studies
There were also a number of individual difference measures that were analysed only for a single
study. As single studies cannot be meta-analysed, they will be presented individually in Table 1 below, in
order of effect size.
Table 1
Summary Statistics of All Individual Difference Measures Included in Only One Study
Individual Difference Measure Study Author(s) Effect Size (r)Sample Size
Memory Stanford, 1970 .36*30
Sensitivity to Punishment Hitchman et al., 2016 -.21 51
Optimism Haraldsson et al., 2002 .20*50
Fantasy Proneness Thalbourne, 1996 -.13 99
Latent Inhibition Hitchman , Sherwood, &
Roe, 2015
-.11 48
Belief in the Occult Haraldsson et al., 2002 .10 50
Time Belief as Naturalistic Schmeidler, 1964b -.08 75
Psychotism Haraldsson et al., 2002 -.08 50
Hypersensitivity Thalbourne, 1996 .04 99
Time Belief as Humanistic Schmeidler, 1964b .01 75
Cerebral Lateralization Palmer, 2009 .00 64
Sensitivity to Reward Hitchman et al., 2016 .00 51
*p <.05
Summary of Results
Below (Table 2) is a summary of all of the individual difference measures that were meta-analysed
for comparison. The individual difference measures are ordered by overall effect size.
The Journal of Parapsychology
22
Column 1 is the individual difference measure; Column 2 is the 95% condence interval,5 Column
3 is the overall effect size; Column 4 is the number of individual studies included in the meta-analysis;
Column 5 is the number of independent investigators in the meta-analysis; Column 6 is total number of
participants for all of the studies in the meta-analysis combined.
Table 2
Summary Statistics of All Individual Difference Measures That Were Meta-Analysed
Individual
Difference
Measure
95%
Condence
Intervals
Overall
Effect
Size (r)
Number
of
Studies
Number of
Independent
Investigators
Total
Number of
Participants
Chance Belief -.09, .36 .141 5 1 248
Luck Belief .001, .26 .131*5 1 248
Belief in Psi .05, .20 .125* 22 12 2,200
Perceptual Defensiveness .001, .24 .125*6 1 272
Providence Belief -.03, .27 .125 5 1 248
Openness to Experience .02, .22 .124*9 5 522
Extraversion .01, .15 .080* 14 7 1,206
Perceived Luckiness -.14, .28 .076 4 2 231
Dream Recall -.04, .18 .070 4 3 799
Intelligence -.31, .19 -.065 2 2 80
Emotional Reactivity -.27, .38 .064 3 1 151
Conscientiousness -.09, .20 .056 3 3 187
Neuroticism -.08, .19 .054 9 7 528
Religiosity -.24, .14 -.054 2 2 149
Future-Oriented -.13, .23 .053 2 2 118
Creativity -.08, .17 .047 9 3 506
Fortune Belief -.10, .16 .032 5 1 248
Agreeableness -.09, .13 .021 7 4 556
Pro Attitude -.12, .08 -.019 3 1 393
Time Belief as Dynamic -.03, -.001 -.017*2 2 14,016
Transliminality -.13, .15 .008 5 1 542
Spontaneous Experences -.08, .07 -.006 3 2 695
Temporal Lobe
Dysfunction
-.19, .18 -.004 2 2 114
*p < .05
Discussion
As we can see from the summary of results, the majority of these individual difference measures
have not been extensively investigated, with the exception of belief in psi, extraversion, and neuroticism. It
5 As effect size estimates based on previous research are inherently uncertain, conrmatory studies based on lower
condence intervals are less likely to overestimate the true effect size (Kennedy, 2016). Therefore 80% and 68% con-
dence intervals are provided in the Appendix, Table A2.
Individual Difference Correlates: A Meta-Analysis 23
might therefore be argued that such a meta-analysis is unnecessary. However, without a meta-analysis, re-
searchers will likely impose their own synthesis of the data, and a meta-analysis can provide greater clarity
in this regard—even if it only incorporates two or three studies, as Valentine, Pigott, and Rothstein (2010)
argue that all other synthesis techniques are less transparent and/or less likely to be valid. At the same time,
it is not intended to stop researchers from exploring individual difference measures that may not yet seem
promising, especially those that have been tested only a handful of times; it is merely given as a benchmark
of past results.
With that being said, the results suggest that there may be only a small pool of individual difference
measures that are robustly correlated with performance on a forced-choice precognition task. This is also
consistent with Steinkamp’s (2005) review of forced-choice ESP experiments, where she found that “there
are few variables which have correlated clearly with success . . . most variables tested provided little evi-
dence either way as being ultimately psi-conducive and there were relatively few variables that appeared to
be encouraging” (p. 155). However, notable exceptions in this meta-analysis include extraversion and belief
in psi, which show more consistent results across a larger number of studies. It should also be noted that
with the number of meta-analyses conducted in this paper, there is an increased risk of family-wise error,
and that one or more of these signicant ndings might be the result of multiple analyses (e.g., represent a
false positive). Further, while forced-choice ESP tests produce normal distributions (unless the number of
trials is very small and/or the number of response alternatives is very large), it may be useful to incorporate
nonparametric statistics in the case of any highly skewed distributions. Researchers should bear this in mind
when setting their expectations for future experiments.
Nevertheless, the ndings suggest a small but signicant relationship between the following indi-
vidual difference measures and psi performance: luck belief (specically, the belief that luck is primarily
controllable), perceptual defensiveness, openness to experience, belief in psi, extraversion, and time be-
lief as dynamic. Perhaps what underlies these individual differences is a mechanism derived from being
open-minded, curious, social, and intuitive—all of which might lead people to discuss, think about, and
explore the “paranormal.” Consequently, these same people may act on information or intuitions that others
may ignore or supress, leading them to make better decisions about the future than we would expect by
chance alone. This may mean that a relationship exists between these variables and performance on a psi
task, where such traits may either facilitate or innately allow demonstrations of psi.
However, given that even the strongest predictor (extraversion) in this meta-analysis accounts only
for approximately 2% of the variance on its own, perhaps these predictors are not related and are instead ad-
ditive, and provide more power when analysed together. Therefore, the best strategy for future researchers
may be to combine individual difference factors, not just for the additive benets but also to examine poten-
tial interactions (see Baron & Kenny, 1986) between the factors that may predict precognitive performance.
Alternatively, these results may be due to statistical anomalies, having arisen from the large amount
of analyses being conducted on individual differences in psi research, or even due to methodological aws.
The other possibility is that the results reect an actual relationship between certain variables combined
with imperfect research designs. Taking into consideration the fact that many of these ndings, including
the nonsignicant results, were based on only a limited amount of studies, it is difcult to come to any
strong conclusions. If one also considers the possibility of experimenter psi (i.e., where the experiment-
er inuences the nal results of an experiment due to his or her own psi abilities), it becomes extremely
difcult to disentangle the data, especially in the case of a meta-analysis with only a single independent
investigator. This potential explanation has previously been offered for the relationship between perceptual
defensiveness and psi performance (see Haraldsson et al., 2002).
Retrospective meta-analyses also have several limitations, so it is not appropriate to make any de-
nitive statements about the results without rst conducting conrmatory studies. One such limitation of
retrospective meta-analysis is that included studies are often affected by publication bias or the le-drawer
effect, whereby only signicant results are reported or published. Although parapsychological journals
generally publish more nonsignicant results than most mainstream scientic journals (Mousseau, 2003),
no eld is entirely immune, especially when there may be tens or even hundreds of secondary analyses con-
The Journal of Parapsychology
24
ducted (e.g., various individual difference measures). Indeed, the low fail safe N numbers found in several
of these meta-analyses (e.g., nine for extraversion) suggest that publication bias is a possibility. At the same
time, there is no indication to argue strongly that publication bias is a problem when taking into account
Egger’s test results, which should be reassuring for parapsychologists given Mousseau’s (2003) ndings.
Secondly, there will always be subjectivity involved in meta-analytical procedures and interpretation, such
as dening and judging exclusion criteria, using search strategies, or coding the studies (Murray, 2011).
Biases will come into play—whether conscious or unconscious—that inuence procedural decisions, es-
pecially since researchers will already be aware of the results of the individual studies (Watt & Kennedy,
2016). This subjectivity allows psi proponents and critics to conduct meta-analyses whose conclusions
often support their own prior beliefs, but never manage to convince the other side (Palmer, 2003).
Yet meta-analyses are still useful in that they can suggest the conditions under which replication
is most likely to occur, assuming an effect exists at all. An overall effect size also gives future researchers
the ability to calculate how many participants they would need to include in their experiment for it to be
adequately powered. With these key pieces of information, prospective meta-analyses (which dene the
exclusion criteria and other details beforehand) can then be conducted using only future studies that are to
be preregistered. A prospective meta-analysis therefore avoids all the potential issues of publication bias
and subjectivity that are evident in a retrospective meta-analysis, while also addressing methodological
issues such as optional stopping. In fact, Watt (2016) has recently set up a registration-based prospective
meta-analysis of one of the most thorough yet controversial paradigms in parapsychology, the ganzfeld—
Watt’s (2016) meta-analysis protocol specically includes only preregistered individual studies that pro-
spectively t their criteria. This is a positive direction for parapsychology, as it brings structure and focus
to the eld.
Preregistration has been made even easier by the Koestler Parapsychology Unit (KPU) registry, an
initiative started at the University of Edinburgh in 2012 that allows researchers to prospectively register
their experiments in detail, publically, and is not afliated with a specic journal (Watt & Kennedy, 2015).
Not only are prospective meta-analyses the ideal way to test the replicability of psi phenomena, but they are
also the best way to conrm the null hypothesis should psi not exist. Alcock (2003) claims that the latter hy-
pothesis often does not get serious consideration by parapsychologists, so multiple prospective meta-analy-
ses showing nonsignicance may force parapsychologists to give the null hypothesis more deliberation than
a single study or retrospective meta-analysis would.
Another consideration for attempting replication is the researcher conducting the experiment. Al-
though some parapsychologists believe that the psi experimenter effect eliminates the possibility of true
replication, that is, that due to the nature of psi only experimenters who are proponents of psi will get
positive results in psi experiments whereas sceptics will not (Utts, 2015), most researchers would only be
satised that psi phenomena exist if it were to be consistently demonstrated by neutral scientists and not
just a select few who believe in psi (Alcock, 2003; Palmer, 2016). The current meta-analysis was conducted
with this goal in mind, as forced-choice precognition experiments are arguably the easiest to run and can be
automated using computer programmes. For example, Bem (2011) ran his Precognitive Detection of Erot-
ic Stimuli experiment using an automated computer programme. This allows researchers to collect large
amounts of data with relatively little effort, an important consideration if researchers are to try and replicate
the small effect sizes shown in this meta-analysis (Steinkamp, 2005).
Ultimately, it is hoped that this meta-analysis can be used as a springboard for future research,
allowing the ndings to be used in a productive way and perhaps aiding in the development of research
programmes that are specic and structured. As Watt (2005) comments, “Parapsychologists need to be far
more systematic in how they tackle these questions. . . . Systematic follow-up is an essential prerequisite
for demonstrating a replicable effect” (p. 222). With parapsychology being such a small eld, it is important
that researchers work together to build up a body of evidence that is considered respectable by both para-
psychologists and mainstream academics. With the recent failures to replicate many foundational studies in
both psychology and medicine (Open Science Collaboration, 2015), now is the perfect time to dene what
a replicable psi experiment really is and take advantage of the benets of preregistration. Only then will
Individual Difference Correlates: A Meta-Analysis 25
we be able to nally conrm or disconrm some of the major hypotheses in psi research. Depending on
whether you are extraverted or believe in psi, you may already know how it will turn out.
References
[*Paper in the meta-analysis]
Alcock, J. (2003). Give the null hypothesis a chance: Reasons to remain doubtful about the existence of psi. Journal
of Consciousness Studies, 10, 29–50.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research:
Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–
1182. doi:10.1037/0022-3514.51.6.1173
Bem, D. J. (2003). Precognitive habituation: Replicable evidence for a process of anomalous cognition. Proceedings
of Presented Papers: The Parapsychological Association 46th Annual Convention, 6–20.
*Bem, D. J. (2011). Feeling the future: Experimental evidence for anomalous retroactive inuences on cognition and
affect. Journal of Personality and Social Psychology, 100, 407–425. doi:10.1037/a0021524
Bem, D. J., & Honorton, C. (1994). Does psi exist? Replicable evidence for an anomalous process of information
transfer. Psychological Bulletin, 115, 4–18. doi:10.1037/0033-2909.115.1.4
Bhadra, B. H. (1966). The relationship of test scores in ESP. Journal of Parapsychology, 30, 1–17.
Borenstein, M., Hedges, L., Higgins, J., & Rothstein, H. (2005). Comprehensive meta-analysis (version 2) [Computer
software]. Englewood, NJ: Biostat.
Borenstein, M., Hedges, L., Higgins, J., & Rothstein, H. (2009). Introduction to Meta-Analysis. Chichester, UK: Wiley.
Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2010). A basic introduction to xed-effect and
random-effects models for meta-analysis. Research Synthesis Methods, 1, 97–111. doi:10.1002/jrsm.12
*Buzby, D. E. (1967). Subject attitude and score variance in ESP tests. Journal of Parapsychology, 31, 43–50.*
Cattell, H. E. P. (1996). The original big-ve: A historical perspective. European Review of Psychology, 46, 5–14.
Cattell, H. E., & Mead, A. D. (2008). The Sixteen Personality Factor Questionnaire (16PF). In G. Boyle, G. Matthews,
& D. Saklofske (Eds.), The SAGE handbook of personality theory and assessment (Vol. 2), (pp.135–178). Trow-
bridge, UK: Cromwell Press.
*Drucker, S. A., Drewes, A. A., & Rubin, L. (1977). ESP in relation to cognitive development and IQ in young chil-
dren. Journal of the American Society for Psychical Research, 71, 289–298.
Dunn, L. M., & Hottel, J. V. (1961). Peabody Picture Vocabulary Test performance of trainable mentally retarded
children. American Journal of Mental Deciency, 65, 448–452.
Egger, M., Davey Smith, G., Schneider, M., & Minder, C. (1997). Bias in meta-analysis detected by a simple, graph-
ical test. BMJ, 315, 629–634. doi: 10.1136/bmj.315.7109.629
Ferguson, C. J., & Heene, M. (2012). A vast graveyard of undead theories: Publication bias and psychological sci-
ence’s aversion to the null. Perspectives on Psychological Science, 7, 555–561. doi:10.1177/1745691612459059
*Freeman, J., & Nielsen, W. (1964). Precognition score deviations as related to anxiety levels. Journal of Parapsy-
chology, 28, 239–249.
Goldberg, L. R. (1999). A broad-bandwidth, public domain, personality inventory measuring the lower-level facets of
several ve-factor models. Personality Psychology in Europe, 7, 7–28.
*Haraldsson, E. (1975). Reported dream recall, precognitive dreams, and ESP [Abstract]. In J. D. Morris, W. G. Roll,
& R. L. Morris (Eds.), Research in parapsychology 1974 (pp. 47–48). Metuchen, NJ: Scarecrow Press.
*Haraldsson, E. (1980). Scoring in a precognition test as a function of the frequency of reading on psychical phenom-
ena and belief in ESP. Research Letter, 1, 1–8.
Haraldsson, E. (1993). Are religiosity and belief in an afterlife better predictors of ESP performance than belief in
psychic phenomena?. Journal of Parapsychology, 57, 259–273.
*Haraldsson, E., Houtkooper, J. M., Schneider, R., & Bäckström, M. (2002). Perceptual defensiveness and ESP per-
formance: Reconstructed DMT ratings and psychological correlates in the rst German DMT-ESP experiment.
Journal of Parapsychology, 66, 249–270.
*Hitchman, G. A., Pfeuffer, C. U., Roe, C. A., & Sherwood, S. J. (2016). The effects of experimenter-participant in-
teraction qualities in a goal-oriented nonintentional precognition task. Journal of Parapsychology, 80, 45–69.
*Hitchman, G. A., Roe, C. A., & Sherwood, S. J. (2012). A reexamination of nonintentional precognition with open-
ness to experience, creativity, psi beliefs, and luck beliefs as predictors of success. Journal of Parapsycholo-
gy, 76, 109–145.
*Hitchman, G. A., Roe, C. A., & Sherwood, S. J. (2015). The relationship between lability and performance at inten-
The Journal of Parapsychology
26
tional and nonintentional versions of an implicit PMIR-type psi task. Journal of Parapsychology, 79, 65–86.
*Hitchman, G. A., Sherwood, S. J., & Roe, C. A. (2015). The relationship between latent inhibition and performance
at a non-intentional precognition task. Explore: The Journal of Science and Healing, 11, 118–126. doi:10.1016/j.
explore.2014.12.004
Holt, N. J. (2002). Creativity and boundary permeability. Unpublished manuscript, University of Northampton, UK.
*Honorton, C. (1972). Reported frequency of dream recall and ESP. Journal of the American Society for Psychical
Research, 66, 369–374.
Honorton, C., & Ferrari, D. C. (1989). “Future telling”: A meta-analysis of forced-choice precognition experiments,
1935-1987. Journal of Parapsychology, 53, 281–308.
Honorton, C., Ferrari, D. C., & Bem, D. J. (1998). Extraversion and ESP performance: A meta-analysis and a new
conrmation. Journal of Parapsychology, 62, 255–276.
*Humphrey, B. M. (1945). An exploratory correlation study of personality measures and ESP scores. Journal of Para-
psychology, 9, 116–123.
Hyman, R. (1977). The case against parapsychology. The Humanist, 37, 37–49.
Hyman, R. (1985). The ganzfeld psi experiment: A critical appraisal. Journal of Parapsychology, 49, 3–49.
Hyman, R. (2010). Meta-analysis that conceals more than it reveals: Comment on Storm et al. (2010). Psychological
Bulletin, 136, 486–490. doi:10.1037/a0019676
Kennedy, J. E. (2001). Why is psi so elusive? A review and proposed model. Journal of Parapsychology, 65, 219–246.
Kennedy, J. E. (2016). Is the methodological revolution in psychology over or just beginning? Journal of Parapsy-
chology, 80, 156–168.
Knapp, R. H., & Garbutt, J. T. (1958). Time imagery and the achievement motive. Journal of Personality, 26, 426–
434. doi:10.1111/j.1467-6494.1958.tb01597
Kragh, U., & Smith, G. W. (Eds.) (1970). Percept genetic analysis. Lund, Sweden: Gleerup.
Lawrence, T. R. (1993). Gathering in the sheep and goats: A meta-analysis of forced-choice sheep-goat ESP stud-
ies, 1947-1993. Proceedings of Presented Papers: The Parapsychological Association 36th Annual Convention,
75–86.
Liberati, A., Altman, D. G., Tetzlaff, J., Mulrow, C., Götzsche, P. C., Ioannidis, J. P. A., & Moher, D. (2009). The
PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare inter-
ventions: Explanation and elaboration. BMJ, 339, b2700. doi:10.1136/bmj.b2700.
Luke, D. P., Delanoy, D., & Sherwood, S. J. (2003). Questionnaire of beliefs about luck. Unpublished instrument,
University of Northampton, UK.
*Luke, D. P., Delanoy, D., & Sherwood, S. J. (2008). Psi may look like luck: Perceived luckiness and beliefs about
luck in relation to precognition. Journal of the Society for Psychical Research, 72, 193–207.
*Luke, D. P., & Morin, S. (2009). Luck beliefs, PMIR, psi and the sheep-goat effect: A replication. Abstracts of papers
presented at the 33rd International Conference of the Society for Psychical Research, University of Nottingham
(pp. 12–13).
*Luke, D. P., Roe, C. A., & Davison, J. (2008). Testing for forced-choice precognition using a hidden task: Two rep-
lications. Journal of Parapsychology, 72, 133–154.
Mangan, G. L. (1958). A review of published research on the relationship of some personality variables to ESP scoring
level. New York, NY: Parapsychology Foundation.
McCarthy, D., & Schechter, E. I. (1986). Estimating effect size from critical ratios [Abstract]. In D. H. Weiner & D. I.
Radin (Eds.), Research in parapsychology 1985 (pp. 95–96). Metuchen, NJ: Scarecrow Press.
McCrae, R. R., & Costa, P. T. (1987). Validation of the ve-factor model of personality across instruments and observ-
ers. Journal of personality and social psychology, 52, 81–90. doi:10.1037/0022-3514.52.1.81
Milton, J., & Wiseman, R. (1999). Does Psi exist? Lack of replication of an anomalous process of information trans-
fer. Psychological Bulletin, 125, 387–391. doi:10.1037/0033-2909.125.4.387
Mousseau, M. C. (2003). Parapsychology: Science or pseudo-science. Journal of Scientic Exploration, 17, 271–282.
Murray, A. L. (2011). The validity of the meta-analytic method in addressing the issue of psi replicability. Journal of
Parapsychology, 75, 261–277.
Nazimuddin, S. K. (2015). A study of individual differences in educational situations. International Journal of Scien-
tic Engineering and Research, 3, 180–184.
Open Science Collaboration. (2015). Estimating the reproducibility of psychological science. Science, 349, aaac4716-
1–aaac4716-8. doi:10.1126/science.aac4716
Palmer, J. (1978). Extrasensory perception: Research ndings. In S. Krippner (Ed,), Advances in parapsychological
research 2 (pp. 59–243). New York, NY: Plenum. doi:10.1007/978-1-4615-9092-7_3
Individual Difference Correlates: A Meta-Analysis 27
Palmer, J. (2003). ESP in the ganzfeld: Analysis of a debate. Journal of Consciousness Studies, 10, 51–68.
*Palmer, J. (2009). Decision augmentation in a computer guessing task. Journal of Parapsychology, 73, 119–136.
Palmer, J. (2016). Hansel’s ghost: Resurrection of the experimenter fraud hypothesis in parapsychology [Editorial].
Journal of Parapsychology, 80, 5–16.
Persinger, M. A., & Makarec, K. (1987). Temporal lobe epileptic signs and correlative behaviors displayed by normal
populations. Journal of General Psychology, 114, 179–195. doi:10.1080/00221309.1987.9711068
Rhine, J. B., Pratt, J. G., Stuart, C. E., Smith, B. M., & Greenwood, J. A. (1940). Extra-sensory perception after
sixty years: A critical appraisal of the research in extra-sensory perception. New York, NY: Henry Holt.
doi:10.1037/13598-000
Rosenthal, R. (1979). The “le drawer problem” and tolerance for null results. Psychological Bulletin, 86, 638–641.
doi:10.1037/0033-2909.86.3.638
Rothstein, H. R. (2008). Publication bias as a threat to the validity of meta-analytic results. Journal of Experimental
Criminology, 4, 61–81. doi:10.1007/s11292-007-9046-9
Schmeidler, G. R. (1943). Predicting good and bad scores in a clairvoyance experiment. A preliminary report. Journal
of the American Society for Psychical Research, 37, 103–110.
*Schmeidler, G. R. (1964a). An experiment on precognitive clairvoyance. Part I. The main results. Journal of Para-
psychology, 28, 1–14.
*Schmeidler, G. R. (1964b). An experiment on precognitive clairvoyance, Part III. Precognition scores related to the
subject’s way of viewing time. Journal of Parapsychology, 28, 94–101.
*Schmeidler, G. R. (1964c). An experiment on precognitive clairvoyance: Part IV. Precognition scores related to cre-
ativity. Journal of Parapsychology, 28, 102–108.
*Schmeidler, G. R. (1964d). An experiment on precognitive clairvoyance: Part V. Precognition scores related to feel-
ings of success. Journal of Parapsychology, 28, 109–125.
Schmidt, S., Schneider, R., Utts, J., & Walach, H. (2004). Distant intentionality and the feeling of being stared at: Two
meta-analyses. British Journal of Psychology, 95, 235–247. doi:10.1348/000712604773952449
Smith, M. D. (1998). Perception of one’s own luck: The formation, maintenance and consequences of perceived luck-
iness (Unpublished doctoral thesis). University of Hertfordshire, UK.
*Stanford, R. G. (1970). Extrasensory effects upon memory. Journal of the American Society for Psychical Re-
search, 64, 161–186.
Steinkamp, F. (1998). Psychology, psi, and the web: An exploratory study. Journal of the American Society for Psy-
chical Research, 92, 256–278.
Steinkamp, F. (2005). Forced-choice ESP experiments: Their past and their future. In M. A. Thalbourne & L. Storm
(Eds.), Parapsychology in the 21st century: Essays on the future of psychical research (pp. 124–163). Jefferson,
NC: McFarland.
Steinkamp, F., Milton, J., & Morris, R. L. (1998). A meta-analysis of forced-choice experiments comparing clairvoy-
ance and precognition. Journal of Parapsychology, 62, 193–218.
*Storm, L. (2002). A parapsychological investigation of the I Ching: Seeking psi in an ancient Chinese system of
divination. Australian Journal of Parapsychology, 2, 44–62.
Storm, L. (2006). Technical Paper No. 11: Meta-analysis in parapsychology: I. The ganzfeld domain. Australian Jour-
nal of Parapsychology, 6, 35–53.
*Storm, L. (2008). Investigations of the I Ching: I. Relationships between psi and time perspective, paranormal belief
and meaningfulness. Australian Journal of Parapsychology, 8, 103–127.
*Storm, L., & Thalbourne, M. A. (1998-1999). The transliminal connection between personality and paranormal ef-
fects in an experiment with the I Ching. European Journal of Parapsychology, 14, 100–124.
Storm, L., Tressoldi, P. E., & Di Risio, L. (2012). Meta-analysis of ESP studies, 1987-2010: Assessing the success of
the forced-choice design in parapsychology. Journal of Parapsychology, 76, 243–273.
Tart, C. T. (1983). Information acquisition rates in forced-choice ESP experiments: Precognition does not work as well
as present-time ESP. Journal of the American Society for Psychical Research, 77, 293–310.
*Thalbourne, M. (1996). An attempt to predict precognition scores using transliminality-relevant variables. Journal of
the Society of Psychical Research, 62, 129–140.
Thalbourne, M. A. (1998). Transliminality: Further correlates and a short measure. Journal of the American Society
for Psychical Research, 92, 402–419.
Thalbourne, M. A., & Delin, P. S. (1993). A new instrument for measuring the sheep-goat variable: Its psychometric
properties and factor structure. Journal of the Society for Psychical Research, 59, 172–186.
Thalbourne, M. A., & Delin, P. S (1994) A common thread underlying belief in the paranormal, mystical experience
The Journal of Parapsychology
28
and psychopathology. Journal of Parapsychology, 58, 3–38.
*Thalbourne, M. A., & Storm, L. (2014). A further study of psychopraxia using the “I Ching.” Australian Journal of
Parapsychology, 14, 115–142.
Torrance, E. P. (1963). Preliminary Manual for Personal-Social Motivation Inventory. Mineapolis, MN: University
of Minnesota.
Utts, J. (1991). Replication and meta-analysis in parapsychology. Statistical Science, 6, 363–378.
Utts, J. (2015). What constitutes replication in parapsychology?. In E. C. May, & S. B. Marwaha (Eds.), Extrasensory
perception: Support, skepticism, and science (Vol. 1). Santa Barbara, CA: ABC-CLIO.
Valentine, J. C., Pigott, T. D., & Rothstein, H. R. (2010). How many studies do you need? A primer on statistical power for
meta-analysis. Journal of Educational and Behavioral Statistics, 35, 215–247. doi:10.3102/1076998609346961
*Vaughn, A., & Houck, J. (1993). A “success” test of precognition and attitude toward the future. Journal of the Amer-
ican Society for Psychical Research, 59, 259–268.
Watt, C. (2005). Parapsychology’s contribution to psychology: A view from the front line. Journal of Parapsycholo-
gy, 69, 215–231.
Watt, C. (2016). A prospective meta-analysis of pre-registered ganzfeld ESP studies. KPU Study Registry. Retrieved
from http://www.koestler-parapsychology.psy.ed.ac.uk/Documents/KPU_Registry_1024.pdf
Watt, C., & Kennedy, J. E. (2015). Lessons from the rst two years of operating a study registry. Frontiers in psychol-
ogy, 6, 173. doi:10.3389/fpsyg.2015.00173
Watt, C., & Kennedy, J. E. (2016). Options for prospective meta-analysis and introduction of registration-based pro-
spective meta-analysis. Frontiers in Psychology, 7, 2030. doi:10.3389/fpsyg.2016.02030
*Winkelman, M. (1981). The effect of formal education on extrasensory abilities: The Ozolco study. Journal of Para-
psychology, 45, 321–336.
Wiseman, R., & Greening, E. (2002). The Mind Machine: A mass participation experiment into the possible existence
of extrasensory perception. British Journal of Psychology, 93, 487–499. doi:10.1348/000712602761381367
Meta-Analysis Papers Not Cited
Haraldsson, E., & Johnson, M. (1979). ESP and the Defense Mechanism Test (DMT) Icelandic study No. III: A case
of the experimenter effect? European Journal of Parapsychology, 3, 11–20.
Haraldsson, E. (1978). ESP and the defense mechanism test (DMT): A further validation. European Journal of Para-
psychology, 2, 104–114.
Haraldsson, E., & Johnson, M. (1986). The Defense Mechanism Test (DMT) as a predictor of ESP performance: Ice-
landic studies VI and VII [Abstract]. In D. H. Weiner & D. I. Radin (Eds.), Research in parapsychology 1985
(pp. 43–44). Metuchen, NJ: Scarecrow Press.
Honorton, C. (1967). Creativity and precognition scoring level. Journal of Parapsychology, 31, 29–42.
Johnson, M. (1969). Attitude and target differences in a group precognition test. Journal of Parapsychology, 33,
324–325.
Johnson, M., & Haraldsson, E. (1984). The Defense Mechanism Test as a predictor of ESP scores: Icelandic studies
IV and V. Journal of Parapsychology, 48, 185–200.
Nielsen, W. (1970). Relationships between precognition scoring level and mood. Journal of Parapsychology, 34,
93–116.
Ryzl, M. (1968a). Precognition scoring and attitude toward ESP. Journal of Parapsychology, 32, 1–8.
Ryzl, M. (1968b). Precognition scoring and attitude. Journal of Parapsychology, 32, 183–189.
Schwartz, S. A., & De Mattei, R. J. (1983). The Mobius Psi-Q Test: Preliminary ndings [Abstract]. In R. A. White &
J. Solfvin (Eds.), Research in parapsychology 1984 (pp. 103–105). Metuchen, NJ: Scarecrow Press.
Smith, M. D., Wiseman, R., Machin, D., Harris, P., & Joiner, R. (1997). Luckiness, competition and performance on a
psi task. Journal of Parapsychology, 61, 33–43.
Storm, L. (2006). A parapsychological investigation of the I Ching: The relationship between psi, intuition, and time
perspective. Journal of Parapsychology, 70, 121–141.
Storm, L., & Thalbourne, M. A. (2001). Studies of the I Ching: I. A replication. Journal of Parapsychology, 65,
105–124.
Thalbourne, M., Beloff, I., & Delanoy, D. (1982). A test for the “extraverted sheep versus introverted goats” hy-
pothesis [Abstract]. In W. G. Roll, R. L. Morris, & R. A. White (Eds.), Research in parapsychology 1981 (pp.
155–156). Metuchen, NJ: Scarecrow Press.
Wagenmakers, E. J., Wetzels, R., Borsboom, D., van der Maas, H. L., & Kievit, R. A. (2012). An agenda for purely
Individual Difference Correlates: A Meta-Analysis 29
conrmatory research. Perspectives on Psychological Science, 7, 632–638. doi:10.1177/1745691612463078
Wilson, S., & Hamlin, I. (2007). Implicit learning in a card prediction task. European Journal of Parapsychology, 22,
3–29.
Acknowledgements
The authors would like to thank Caroline Watt, Gergö Hadlaczky, and Jan Dalkvist for their comments on
an earlier draft of this paper. We would also like to thank John Palmer and the two anonymous referees for
their constructive feedback.
Victoria University of Wellington
Kelburn Parade
PO Box 600
Wellington 6012, New Zealand
Marco.Zdrenka@vuw.ac.nz
Appendix
Table A1
Number of Studies for Each Individual Difference Measure That Reported Nonsignicance
Without Providing Any Other Statistical Information
Individual Difference
Measure
Number of non-signicant studies
estimated as r = .00
Belief in Psi 5
Extraversion 3
Neuroticism 2
Agreeableness 2
Spontaneous Experiences 2
Openness to Experience 1
Dream Recall 1
Conscientiousness 1
Temporal Lobe Dysfunction 1
Cerebral Lateralization 1
The Journal of Parapsychology
30
Table A2
68% and 80% Condence Intervals for All Individual
Difference Measures Meta-Analysed
Individual Difference
Measure
68% Condence
Intervals
80% Condence
Intervals
Chance Belief .03, .25 -.01, .28
Luck Belief .07, .20 .05, .21
Belief in Psi .09, .16 .07, .18
Perceptual Defensiveness .06, .19 .04, .20
Providence Belief .05, .20 .02, .22
Openness to Experience .07, .17 .06, .19
Extraversion .05, .11 .04, .12
Perceived Luckiness -.03, .18 -.06, .21
Dream Recall .01, .13 -.004, .14
Intelligence -.19, .07 -.23, .10
Emotional Reactivity -.11, .23 -.16, .28
Conscientiousness -.02, .13 -.04, .15
Neuroticism -.01, .12 -.03, .14
Religiosity -.15, .05 -.18, .07
Future-Oriented -.04, .15 -.07, .17
Creativity -.02, .11 -.04, .13
Fortune Belief -.03, .10 -.05, .12
Agreeableness -.04, .08 -.05, .10
Pro Attitude -.07, .03 -.08, .05
Time Belief as Dynamic -.03, -.01 -.03, -.01
Transliminality -.06, .08 -.08, .10
Spontaneous Experiences -.04, .03 -.05, .04
Temporal Lobe Dysfunction -.10, .09 -.13, .12
Individual Difference Correlates: A Meta-Analysis 31
Abstracts in Other Languages
French
LES CORRELATS DES DIFFERENCES INTERINDIVIDUELLES DES PERFORMANCES PSI DANS
LES EXPERIMENTATIONS DE PRECOGNITION A CHOIX FORCE : UNE META-ANALYSE (1945-
2016)
RESUME : Les précédentes recherches en parapsychologie ne se sont pas révélées particulièrement con-
vaincantes, en grande partie du fait de manque de réplicabilité des résultats signicatifs. Pour traiter ce
problème et mieux comprendre les facteurs qui pourraient être associés avec des tailles d’effet plus fortes
et consistantes, toutes les expérimentations de précognition à choix forcé analysant des différences inter-
individuelles (par exemple, des traits de personnalité) ont été rassemblées pour déterminer quels facteurs
peuvent prédire de façon able les performances psi. Globalement, un ensemble de 55 études publiées entre
1945 et 2016, comprenant 35 mesures des différences interindividuelles, fut soumis à une méta-analyse.
Six mesures de différence interindividuelle se sont révélées être corrélées signicativement avec les per-
formances psi, à savoir : la croyance à la chance (croyance que la chance peut être directement contrôlée),
la défense perceptive, l’ouverture à l’expérience, la croyance au psi, l’extraversion et la croyance à un
temps dynamique. Etant donnée la nature particulièrement directe des expérimentations de précognition à
choix forcé, il semble prometteur de continuer à explorer ces facteurs dans des études conrmatoires. Nous
espérons que des chercheurs pourront modeler leurs futures expérimentations en conjonction avec des tech-
niques de pré-enregistrement, an de créer ultimement une base de données plus systématique et robuste.
German
KORRELATE INTERINDIVIDUELLER UNTERSCHIEDE VON PSI-LEISTUNGEN BEI PRÄKOGNI-
TIONSEXPERIMENTEN MIT BEGRENZTER WAHL: EINE META-ANALYSE (1945-2016)
ZUSAMMENFASSUNG: Die bisherige Forschung in der Parapsychologie war nicht besonders überze-
ugend, was größtenteils mit der mangelnden Replizierbarkeit signikanter Ergebnisse zusammenhängt.
Um diesem Mangel abzuhelfen und um besser zu verstehen, welche Faktoren zu stärkeren und konsis-
tenteren Effektstärken beitragen könnten, wurden alle Präkognitionsexperimente mit begrenzter Wahl,
die interindividuelle Unterschiede untersuchten, z. B. Persönlichkeitsmerkmale, zusammengefasst, um zu
bestimmen, welche Faktoren zuverlässig Psi-Leistungen vorhersagen könnten. Insgesamt wurden 55 Stu-
dien, die zwischen 1945 und 2016 veröffentlicht worden waren und 35 Messgrößen für interindividuelle
Unterschiede verwendeten, für eine Meta-Analyse zusammengefasst. Bei sechs Messgrößen für interin-
dividuelle Unterschiede, nämlich Glaube an Glück (der Glaube, dass Glück primär kontrollierbar sei),
Wahrnehmungsabwehr, Offenheit für Erfahrungen, der Glaube an Psi, Extraversion und der Glaube, dass
Zeit dynamisch sei, zeigte sich eine signikante Korrelation mit der Psi-Leistung. Angesichts der besonders
einfachen Natur dieser Art von Experimenten könnte eine erfolgversprechende Forschungsrichtung darin
bestehen, diese Faktoren in Bestätigungsexperimenten zu untersuchen. Es wäre zu hoffen, dass die Forscher
bei ihren zukünftigen Experimenten diese Ergebnisse berücksichtigen, um – zusammen mit den Techniken
der vorherigen Registrierung – zur Schaffung einer letztlich systematischeren und robusteren Datenbank
beizutragen.
Spanish
CORRELATOS DE DIFERENCIAS INDIVIDUALES EN EXPERIMENTOS DE PRECOGNICIÓN DE
ELECCIÓN FORZADA: UN META-ANÁLISIS (1945-2016)
The Journal of Parapsychology
32
RESUMEN: Las investigaciones previas en parapsicología no han sido particularmente convincentes, en
gran parte debido a la falta de replicabilidad de los hallazgos signicativos. Para resolver este problema
y comprender mejor qué factores pueden estar asociados con tamaños de efecto más fuertes y consisten-
tes, se agregaron todos los experimentos de precognición de elección forzada con análisis de diferencias
individuales (p. ej., rasgos de personalidad) para determinar qué factores podrían predecir ablemente el
rendimiento psi. En total, 55 estudios publicados entre 1945 y 2016, incluyendo 35 medidas de diferen-
cia individuales, fueron objeto de meta-análisis. Seis medidas de diferencias individuales correlacionaron
signicamente con la tarea psi: la creencia en la suerte (la creencia de que la suerte es básicamente contro-
lable), defensividad perceptual, apertura a la experiencia, creencia en psi, extraversión, y creencia en un
tiempo dinámico. Dada la naturaleza clara de los experimentos de precognición de elección forzada, una
vía futura prometedora sería explorar estos factores en estudios conrmatorios. Se espera que los inves-
tigadores puedan modelar sus futuros experimentos en estos hallazgos en conjunción con las técnicas de
prerregistro, para crear una base de datos más sistemática y robusta.
