Schizotypy, self-referential thinking and the Barnum effect
Oliver J. Mason a,*, Katie Budge b
a
Department of Clinical Educational and Health Psychology, University College London, 1-19 Torrington Place, London WC1N 6BT, United Kingdom
b
Department of Experimental Psychology, University of Oxford, South Parks Road, Oxford, OX1 3UD, United Kingdom
a r t i c l e i n f o
Article history:
Received 24 June 2010
Received in revised form
5 November 2010
Accepted 10 November 2010
Keywords:
Schizotypy
Referential thinking
Barnum effect
O-LIFE
a b s t r a c t
Background and Objectives: The tendency for people to endorse, as an accurate description of themselves,
personality descriptions that are essentially bogus is well-attested. The study tested whether the socalled
‘Barnum’ effect could be predicted by individual differences in self-referential thinking, and
beyond this, schizotypy more generally.
Methods: 130 Participants completed four different measures of the Barnum effect followed by measures
of schizotypy and self-referential thinking.
Results: Both self-referential thinking and positive schizotypy independently predicted the degree of
agreement with several Barnum measures including both favorable and unfavorable personality
descriptions, as well as computer-generated and horoscope-based readings.
Limitations: The sample is heavily represented by students and is not representative of the general
population. Testing at a single point in time may have reduced differences between different indices of
the Barnum effect.
Conclusions: Self-referential thinking and schizotypy more generally are key contributors to the Barnum
effect across a wide range of indices.
Ó 2010 Elsevier Ltd. All rights reserved.
The term ‘Barnum Effect’ (also Forer effect) refers to the tendency
for individuals to endorse, as uniquely their own, personality
descriptions that are such general statements that as P.T. Barnum
observed, they have ‘something for everyone’. A now highly replicated
experiment by Forer (1949) had subjects ﬁrst complete the
Diagnostic Interest Blank, and then receive their ‘own’ test results
(actually an identical rather ‘positive’ personality description). In
this as in subsequent studies most thought the ‘bogus’ proﬁle highly
accurate. Subsequent studies have suggested that the relative
proportion of positive to negative traits in the description and its
perceived source seem important, as well as when they believe that
it has been obtained uniquely for themselves (eg. Snyder, Larson, &
Bloom, 1976).
Only one study to date (Claridge, Clark, Powney, & Hassan, 2008)
has investigated whether individual differences in schizotypy relate
to the size of the tendency, though other studies have reported that
believers in astrology and paranormal phenomena are more
susceptible (Tabacyk, Milford, Springer, & Tabacyk, 1988). Claridge
et al. (2008) found the positive symptom-like experiences of
schizotypy to predict the size of the Barnum effect as measured by
a ‘random’ positive statement. Positive schizotypy is the tendency to
report anomalous, perceptual, cognitive and emotional experiences
and to hold unusual beliefs, often based on these experiences
(sometimes termed ‘magical ideation’). Positive schizotypy, and
magical ideation in particular, has been associated with an increased
tendency to generate (false) hypotheses about random and illusory
contingencies (Brugger et al.,1993; Fyfe, Williams, Pickup, & Mason,
2008), and it may be that the Barnum effect also beneﬁts from this
tendency. Claridge et al. (2008) used a computerized ‘Brainworks’
program that pretends to generate a personality proﬁle from twenty
‘cognitive style-type’ questions (Synergistic Learning Incorporated,
2008). It is unknown whether this mode of delivery or its positive
content had an inﬂuence on the relationship found. The present
study sought to explore the relationship further by including
measures that differed both in terms of how personalized they
appeared, as well in how favorable a description was given. In
addition to the computer-generated proﬁle used by Claridge et al.,
we used both a ‘horoscope-style’ proﬁle (different depending on
date of birth) as well as positive and negative personality proﬁles
(universal). Based on Claridge et al. (2008) we hypothesized positive
schizotypy to predict the size of agreement with Barnum
statements.
Another related phenomenon is the tendency to experience
events as directly referring to one’s self, often termed referential
thinking e itself a part of positive schizotypy. This tendency seems
* Corresponding author. Tel.: þ44 207 76798230; fax: þ44 207 9161989.
E-mail address: o.mason@ucl.ac.uk (O.J. Mason).
Contents lists available at ScienceDirect
Journal of Behavior Therapy and
Experimental Psychiatry
journal homepage: www.elsevier.com/locate/jbtep
0005-7916/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jbtep.2010.11.003
J. Behav. Ther. & Exp. Psychiat. 42 (2011) 145e148
particularly associated with schizotypal personality attributes
(Meyer & Lenzenweger, 2009) and might be seen as a speciﬁc
social-cognitive feature that schizotypes tend to possess more often
than others. It is important to determine whether the majority of
the positive schizotypy effect reported by others is in fact due to
referential thinking, or whether individual differences in schizotypy
can contribute to explaining differences in the Barnum effect
above and beyond referential thinking style.
2. Method
2.1. Participants
64 Male and 66 female participants (N ¼ 130) aged from 18 to 28
were recruited using an e-mailshot reaching the UCL university
student population and their associates. While most were current
students, a number were their friends and associates. At their
attendance for testing they were told that the study involved an
investigation of their personality by a range of methods. They
completed the four Barnum effect measures in a random order
(different each time) followed by the O-LIFE and REF scales. All were
native English speakers and received no incentive for participation.
2.2. Barnum effect measures
Four different ‘personality’ proﬁles were rated by participants
using a 5-point Likert scale in each case to indicate the extent to
which the proﬁle validly depicted their personality (1-“completely
unlike me” to 5-“completely like me”).
2.2.1. Brain works
This ‘self-assessment’ program was downloaded from the
Synergistic Learning Incorporated (2008) website. Participants
answer 20 questions by choosing one from several options, and are
then presented with a personality proﬁle supposedly based on how
their brain functions. This was administered according to the
method of Claridge et al. (2008).
2.2.2. Horoscope proﬁle
Participants received a pack of horoscope proﬁles, with one
proﬁle corresponding to each of the twelve signs of the zodiac,
extracted from an astrological website (Horoscope Proﬁles, 2008).
Participants were instructed to identify their corresponding proﬁle,
to read and then rate it.
2.2.3. Personality proﬁles
In addition, participants received both a positive (favorable) and
a negative (unfavorable) personality proﬁle, each generated for the
purposes of the present study (available from the authors). Both
proﬁles contained a series of Barnum statements with either
a positive (eg. ‘people enjoy you because you have a good sense of
humour’) or negative (eg. ‘although assertive, you have a tendency
to be stubborn’) orientation.
2.3. Self-report schizotypy measures
2.3.1. OxfordeLiverpool inventory of feelings and experiences
(O-LIFE)
The O-LIFE contains four subscales: Unusual Experiences,
Cognitive Disorganisation, Introvertive Anhedonia and Impulsive
Nonconformity. High levels of internal consistency have been
established, with reported alpha coefﬁcients in the region of
0.72e0.89, (Mason, Claridge, & Jackson, 1995) in addition to
numerous validation studies (eg. Steel, Hemsley, & Pickering, 2007).
2.3.2. Referential thinking scale (REF)
Lenzenweger, Bennett, and Lilenfeld’s (1997) 34-item REF scale
assesses a wide range of referential thoughts and experiences,
including both ‘simple’ and ‘guilty’ ideas of reference. The total
numbers of ‘true’ items were summed in order to provide a comparable
degree of schizotypic referential thinking with other participants.
Evidence suggests that the REF possesses good internal
consistencyreliability, testeretestreliability(Lenzenwegeretal.,1997)
and criterionvalidity (Meyer & Lenzenweger, 2009), as it discriminates
those with delusional ideas of reference (Startup et al., 2010).
2.4. Procedure
Following informed consent participants completed basic
demographic information and the seven self-report measures, the
order of which was counter-balanced in a Latin-square design.
Debrieﬁng information was supplied at the conclusion explaining
the purpose of the experiment. Relevant university research
protocols were observed.
Table 1
Sample descriptives (N ¼ 130).
Measure Mean S.D.
Age 21.82 2.05
Unusual experiences 10.82 4.98
Cognitive disorganisation 9.82 3.97
Referential thinking scale 18.62 8.41
Brain works 3.40 1.21
Horoscope 3.28 1.22
Positive personality 3.31 1.29
Negative personality 3.41 1.25
Table 2
Hierarchical regression analyses predicting each Barnum measure.
DV Adj. R2
ES (f2
) F IVs Beta, T, p value
Horoscope 0.44 0.79 35.0 (p < .001) REF CogDis
ImpNon
0.30, 3.52, <0.001
0.30, 3.78, <0.001
0.23, 2.88, <0.01
Brainworks 0.47 0.89 37.5 (p < .001) REF CogDis
UnEx
0.37, 4.36, <0.00
0.22, 2.79, <0.01
0.22, 2.61, <0.01
Positive personality 0.41 0.69 31.3 (p < .001) REF
ImpNon
UnEx
0.21, 2.23, <0.05
0.35, 4.33, <0.001
0.25, 3.00, <0.01
Negative personality 0.56 1.27 57.1 (p < .001) REF
ImpNon
UnEx
0.28, 3.37, <0.001
0.33, 4.77, <0.001
0.32, 4.35, <0.001
REF e referential thinking, UnEx e unusual experiences.
CogDis e cognitive disorganisation, ImpNon e impulsive nonconformity.
O.J. Mason, K. Budge / J. Behav. Ther. & Exp. Psychiat. 42 (2011) 145e148146
3. Results
Table 1 presents the means, standard deviations and ranges for
scores yielded from each of the self-report measures. The REF scale
correlated moderately with several schizotypy scales (UnEx r ¼.60,
CogDis r ¼.52; p < 0.01, ImpNon r ¼ 0.54, p, 0.01). Interestingly, the
size and distribution of the four Barnum effect indices were highly
similar, with means falling between 3.28 and 3.41 on a ﬁve point
scale. There were no signiﬁcant differences in extent of agreement
across the indices. The four Barnum indices were also moderately
to highly inter-correlated with all falling in the range of 0.50e0.66
(all p < .01). Age and gender were unrelated to Barnum indices,
though the age range here is highly restricted.
The hypotheses were tested using hierarchical linear regression
with each of the Barnum effect indices as the dependent variable in
four separate analyses. The REF scale was entered in the ﬁrst block,
followed by a second block of schizotypy scales selected stepwise
using the relatively stringent criterion of F to enter at p < .05 (see
Table 2). According to Cohen’s (1992) conventions all effect sizes
were large suggesting positive schizotypy makes a substantial
contribution to predicting responses to Barnum statements
regardless of the format or degree of positivity of statements. In
every case the REF scale was a signiﬁcant predictor in the ﬁrst block.
However, the schizotypy scales of Unusual Experiences, Cognitive
Disorganisation, and Impulsive Nonconformity explained additional
variance in every case.
4. Discussion
Although it is perhaps unsurprising that referential thinking is
highly relevant to the Barnum effect, we found repeated convincing
evidence that this is so across a variety of methods. At the very least,
this certainly serves to help validate the construct and scale. It is of
note that those proﬁles (from Brainworks and horoscope readings)
that contain a method by which the ‘reading’ is more speciﬁcally
directed to the respondent had the strongest relationships with the
REF scale. As much of the REF scale content concerns negative selfreferential
perception and thinking (so can discriminate often paranoid
delusional patients well, see Startup, Sakrouge, & Mason, 2010),
one might also expect greater agreement with negative than positive
self-statements, and indeed the better prediction was for the negative
proﬁle, though the differences were modest with positive
descriptions also producing signiﬁcant results. Given that grandiosity
and paranoia co-occur in the clinic, this may not be so surprising.
If we consider referential thinking a key element of positive
schizotypy though previous studies have not measured this, the
results are very largely in line with those of Claridge et al. (2008) in
that various combinations of the indices of positive schizotypy (REF
scale, Unusual Experiences and Cognitive Disorganisation) predicted
the size of the Barnum effect to averysubstantial extent. Importantly,
this pattern of results was very largely repeated across a range of
Barnum indices, strongly suggesting that it is not dependent on
a particular Barnum methodology. Our results demonstrate clearly
that schizotypy isofrelevance totheBarnumeffectabove andbeyond
the contribution of a self-referential thinking style.
Entering the REF scale prior to considering other schizotypy
scales is likely to have diminished the variance available for those
scales that have an overlapping relationship with both the REF scale
and the Barnum effect. Nevertheless, Unusual Experiences was
a signiﬁcant additional predictor in three of the four regression
analysis suggesting that it has a broad relevance beyond the one
aspect of referential thinking. In other contexts, Unusual Experiences
has been associated with seeing illusory contingencies
(Brugger et al., 1993; Fyfe et al., 2008), and the present results are
clearly supportive of this occurring here. Greater salience
regardless of the relevance or otherwise of personality descriptors
in the statements seems to have been operating relatively indiscriminately
as it was seen for both positive and negative statements,
and when statements were not presented as personalized to
the individual (‘based’ on personality questions or astrological
sign). Seeing illusory contigencies in this way may well form part of
a stable cognitive style of ‘aberrant salience’ recently proposed by
Kristin and Roiser (2010) based on the clinical concept of the same
name (Kapur, 2003). Psychosis-prone individuals, as well as those
on the spectrum, seen in clinic are therefore much more likely to
‘read into’ statements, including by clinicians, in ways that were not
intended. Our results suggest that this may be the case not simply
for those who already exhibit ideas of reference, but are prone to
positive symptoms more generally. If Kristin and Roiser are correct,
they may do so, not because of a relatively superﬁcial ‘yea-saying’
or other social psychological reason, but because of fundamental
cognitive processes allocating personal signiﬁcance and meaning to
a wide range of stimuli.
As for Claridge et al. (2008), Cognitive Disorganisation contributed
signiﬁcant additional variance. As this schizotypal factor
explicitly contains both social anxiety and cognitive errors, it is
plausible that this ﬁnding relates to other psychosis-relevant
reasoning styles e ‘jumping to conclusions’ and poor ‘tolerance of
ambiguity’ (eg. Colbert & Peters, 2002). The relative inability to
tolerate ambiguity and thus to tend to reach a conclusion on insufﬁcient
evidence has been shown to contribute to the formation of
delusions (Broome et al., 2007). Finally, although Impulsive
Nonconformity is relatively rarely studied it contributed to three out
of the four regression analysis. This factor is sometimes considered
as being of greater relevance to bipolar disorder: a disorder not
immune from excessive self-focus and referential thinking.
Limitations include that the sample was a highly educated one,
and is not representative of the general population, and that the
serial presentations of Barnum measures may have led to ‘crosscontamination’
by one another. Though too preliminary for strong
clinical implications to be drawn, the results nevertheless suggest
that psychosis-proneness measures are highly relevant to the
perception of self-relevant information in the social world as tapped
by the Barnum effect.
References
Broome, M. R., Johns, L., Valli, I., Woolley, J. B., Tabraham, P., Brett, C., et al. (2007).
Delusion formation and reasoning biases in those at clinical high risk for
psychosis. British Journal of Psychiatry, 191, 38e42.
Brugger, P., Regard, M., Landis, T., Cook, C., Krebs, D., & Niederberger, J. (1993).
Meaningful patterns in visual noise: effects of lateral stimulation and the
observer’s belief in ESP. Psychopathology, 26, 261e265.
Claridge, G., Clark, K. H., Powney, E., & Hassan, E. (2008). Schizotypy and the Barnum
effect. Personality and Individual Differences, 44, 436e444.
Cohen, J. (1992). A primer of power. Psychological Bulletin, 112, 155e159.
Colbert, S. M., & Peters, E. R. (2002). Need for closure and jumping-to-conclusions in
delusion-prone individuals. Journal of Nervous and Mental Disease, 190, 27e31.
Forer, B. R. (1949). The fallacy of personal validation: a classroom demonstration of
gullibility. Journal of Abnormal and Social Psychology, 44, 118e123.
Fyfe, S., Williams, C., Pickup, G., & Mason, O. (2008). Apophenia, theory of mind and
schizotypy: perceiving meaning and intentionality in randomness. Cortex, 44,
1316e1325.
Horoscope Proﬁles. (2008). Retrieved March 2, 2008 from. http://www.jyotishvani.
com/sunsigns.
Kapur, S. (2003). Psychosis as a state of aberrant salience: a framework linking
biology, phenomenology, and pharmacology in schizophrenia. American Journal
of Psychiatry, 160, 13e23.
Kristin, S., & Roiser, J. P. (2010). Assessing the construct validity of aberrant salience.
Frontiers in Behavioral Neuroscience, 4, 1e12.
Lenzenweger, M. F., Bennett, M. E., & Lilenfeld, L. R. (1997). The referential thinking
scale as a measure of schizotypy: scale development and initial construct
validation. Psychological Assessment, 9(4), 452e463.
Mason, O., Claridge, G., & Jackson, M. (1995). New scales for the assessment of
schizotypy. Personality and Individual Differences, 18, 7e13.
Meyer, E., & Lenzenweger, M. (2009). The speciﬁcity of referential thinking:
a comparison of schizotypy and social anxiety. Psychiatry Research, 165, 78e87.
O.J. Mason, K. Budge / J. Behav. Ther. & Exp. Psychiat. 42 (2011) 145e148 147
Snyder, C. R., Larson, D., & Bloom, L. J. (1976). Acceptance of personality interpretations
prior to and after receiving diagnostic feedback supposedly based on
psychological, graphological and astrological assessment procedures. Journal of
Clinical Psychology, 32, 258e265.
Startup, M., Sakrouge, R., & Mason, O. J. (2010). The criterion and discriminant
validity of the referential thinking (REF) scale. Psychological Assessment, 22,
65e69.
Steel, C., Hemsley, D. R., & Pickering, A. D. (2007). Associations between schizotypal
personality traits and the facilitation and inhibition of the speed of contextually
cued responses. Psychiatry Research, 150, 131e140.
Synergistic Learning Incorporated. (2008). ‘Brain works’. Retrieved March 2, 2008
from. www.planetpsych.com/zinteractive/brain.htm.
Tabacyk, J., Milford, G., Springer, T., & Tabacyk, Z. (1988). Paranormal beliefs and the
Barnum effect. Journal of Personality Assessment, 52, 737e739.
O.J. Mason, K. Budge / J. Behav. Ther. & Exp. Psychiat. 42 (2011) 145e148148