Personality and Individual Differences 103 (2016) 209–219

Contents lists available at ScienceDirect

Personality and Individual Differences

journal homepage: www.elsevier.com/locate/paid

Hans Eysenck: A research evaluation

Michael W. Eysenck
Department of Psychology, Roehampton University, Whitelands College, Holybourne Avenue, London SW15 4JD, UK

a r t i c l e i n f o a b s t r a c t

Article history: Hans Eysenck made outstanding contributions to the description of human personality with his identification of
Received 21 December 2015 three orthogonal personality dimensions although his approach was less exhaustive than that of subsequent re-
Accepted 6 April 2016 searchers. He also proposed an ambitious agenda for developing comprehensive theoretical explanations based
on the experimental approach and the biological underpinnings of major personality dimensions. Subsequent
Keywords:
theories have followed his blueprint. Hans Eysenck's higher-level theoretical assumptions have stood the test
Hans Eysenck
Personality
of time better than his lower-level ones. However, a general limitation was his de-emphasis of cognitive process-
Intelligence es and structures. He was less successful at implementation and interpretation than theory generation. This oc-
Behavior therapy curred in part because of his preference for a lawyer-like approach to research rather than a more scientific and
Psychopathology objective one.
© 2016 Published by Elsevier Ltd.

1. Introduction cause in principle it should come closer to capturing the richness of

Hans Eysenck (my father) had a very long and phenomenally pro-
ductive career. It would be absurd of me to pretend to have a detailed
familiarity with all this work. In any case, it would be impossible to dis-
cuss it all in a single article. Instead, I have focused on his major contri-
butions as well as his approach to research. My relationship with my
father is discussed in M. Eysenck (2013).
1.1. Dimensions of personality: historical context
It is important to consider Hans Eysenck's taxonomic approach to in-
dividual differences in personality within the relevant historical context.
American psychologists typically assumed the optimal approach was to
identify a fairly large number of correlated or primary factors. For exam-
ple, Guilford (1939) argued that there were probably approximately 20
important primary traits.
In practice, Guildford identified 11 personality traits or factors. In
similar fashion, Cattell developed his Sixteen Personality Factor (16PF)
Questionnaire that identified 16 traits or factors, many dissimilar to
those identified by Guildford.
This approach can be contrasted with Hans Eysenck's emphasis on
identifying a small number of orthogonal or second-order factors. It
was plausible a priori to prefer the former, primary-factor approach be-
E-mail address: michael.eysenck@roehampton.ac.uk.
human personality. In the words of Cattell, Eber and Tatsouka (1970,
pp. 111-112):
“The primary factors give one most information, and we would ad-
vocate higher-strata [second-order] contributors only as supple-
mentary concepts.”
In practice, this putative advantage of the primary-factor approach
was not manifest for two main reasons. First, it proved very difficult to
replicate the primary factors identified with any given personality test.
For example, Barrett and Kline (1982) conducted several different factor
analyses on the 16PF. They found between seven and nine factors in
these analyses, and these factors were tenuously linked to Cattell's 16
factors. Second, it proved impossible to achieve consensus on the num-
ber and nature of the primary factors of personality.
It was also typically assumed (sometimes implicitly) that individual
differences in these primary factors depended almost entirely on envi-
ronmental influences. This assumption derived in part from the exces-
sively environmentalist behaviorist approach prevalent at the time,
and was subscribed to by later social learning theorists such as Rotter
and Bandura. Intriguingly, Guilford (1934, p. 337) argued that, “most
writers have regarded I–E [introversion–extraversion] as primarily a
matter of heredity”, however, he failed to test this hypothesis.
Many American personality researchers in the 1930s and 1940s re-
lied heavily on factor analysis in their studies of individual differences
in personality. This technique provides suggestive evidence concerning
the number and nature of personality factors. However, the personality

http://dx.doi.org/10.1016/j.paid.2016.04.039
0191-8869/© 2016 Published by Elsevier Ltd.
210 M.W. Eysenck / Personality and Individual Differences 103 (2016) 209–219
factors identified by factor analysis are not necessarily the most impor-
tant ones. Much additional research was required (but rarely carried
out) to establish their importance and validity and to provide explana-
tory accounts.
2. Dimensions of personality
Hans Eysenck's approach to personality was distinctively different
from previous ones. He advocated focusing on large, second-order fac-
tors. His early contention that extraversion and neuroticism are the
most important and readily replicable personality factors has stood
the test of time. Similar factors had previously been identified on both
sides of the Atlantic Ocean. However, nearly all American personality
theorists (and Freud) assumed extraversion was negatively correlated
with neuroticism/anxiety and so often failed to distinguish clearly be-
tween the two dimensions. Of interest, Jung (1923) argued that the
two dimensions were independent of each other. In fact, there is a mod-
erate negative correlation between extraversion and trait anxiety (Gray,
1981) but the neuroticism dimension was designed to be orthogonal to
extraversion.
In England, Burt (1915) identified a factor of general emotionality
resembling neuroticism. Burt (1940) replicated his earlier findings and
also identified “a bipolar factor making for aggressive behavior when
positive and for inhibited or introverted behavior when negative”
(p. 374). However, Burt's research was limited in scope and
methodology.
Evidence supporting the notion that extraversion and neuroticism
are of special importance has come from re-analyzing data from person-
ality questionnaires not explicitly assessing these dimensions
(e.g., Saville & Blinkhorn, 1976, 1981). There is a thorough discussion
of this research in Eysenck and M. Eysenck (1985) and it would be oti-
ose to repeat that discussion. Of particular importance is subsequent re-
search based on the fundamental lexical hypothesis (the notion that all
salient individual differences in personality are represented by single
words in language). Researchers (e.g., Goldberg, 1990) using this hy-
pothesis to discover the main personality factors have consistently iden-
tified factors resembling extraversion and neuroticism even though
their approach differs radically from Hans Eysenck's.
Support for the special importance of extraversion and neuroticism
has come from studies comparing the structure of personality and of
mood. Meyer and Shack (1989) compared the two-dimensional model
of mood (positive affect and negative affect) with the two-
dimensional model of personality (extraversion and neuroticism). The
two models shared a structural identity: extraversion aligned with pos-
itive affect and neuroticism aligned with negative affect.
There is much less consensus that Hans Eysenck's third orthogonal
factor, psychoticism, is of major importance (e.g., van Kampen, 2009).
The dominant approach to personality is the five-factor model (e.g.,
Costa & McCrae, 1992; McCrae & Costa, 1985). The Big Five factors iden-
tified within this model do not include psychoticism. Two of the five fac-
tors (agreeableness and conscientiousness) correlate negatively with
psychoticism although the correlation between conscientiousness and
psychoticism is relatively modest.
The above findings suggest agreeableness and conscientiousness
might be primary factors associated with the higher-order factor of
psychoticism (Eysenck, 1992a,b). However, there is no compelling sup-
port for this interpretation. It is noteworthy that essentially the same
five factors have been found using several very different methodologies
and across numerous cultures (McCrae & Costa, 1997).
Another reason for doubting the fundamental importance of
psychoticism (P) is the fact that its nature changed considerably over
the years: Claridge, Robinson and Birchall (1983) found the correlation
between the original P scale and that in the Eysenck Personality Ques-
tionnaire (EPQ) was only + .19. Finally, the mean scores on the 25-
item psychoticism scale on the EPQ were very low: 2.63 for females
and 3.78 for males. These low means (and a strong positive skew in
the distribution) imply that the P dimension is not directly relevant to
much of the healthy population.
In spite of the above doubts about whether psychoticism qualifies as
one of the main personality dimensions, it is clearly important in many
contexts. For example, Corr (2010) discussed much evidence of com-
monality between psychoticism and psychopathy. More specifically,
he argues persuasively that psychoticism and psychopathy have similar
deficits involving the behavioral approach system, the behavioral inhi-
bition system, and the fight–flight–freeze system.
3. Explanations of personality
Hans Eysenck adopted a more ambitious and systematic approach to
human personality than any previous theorist. The essence of his ap-
proach (Eysenck, 1994, pp. 7–8) was as follows:
“In order to qualify as a major dimension of personality, any concept
should be based on a general theory which has its antecedents in
DNA (genetic analysis), and advances through biological intermedi-
aries (psychophysiological, hormonal, etc.) to the dimensional and
factorial analysis of behavior and trait patterns. It should from there
proceed to the experimental testing of the general theory, making
deductions from the theory as to how people at various positions
on the relevant personality dimensions would behave in carefully
designed experimental situations. And finally one would expect that
the theory, and the knowledge acquired through experimental
study, should enable us to make predictions as to the general social
behavior of people in carefully defined situations.”
Below some of these aspects of Hans Eysenck's approach are
discussed. For reasons of space, it is not feasible to address all of them.
3.1. Heritability of personality
One of Hans Eysenck's key contributions to personality theory was
his emphasis on the major role played by genetic factors in accounting
for individual differences in personality. He argued (Eysenck, 1979,
p. 525), “Using measures of the major personality dimensions P, E, and
N, and calculating heritabilities along the lines of modern biometrical
genetical analysis, we get figures in the band from 60% to 80%, when
test unreliability has been allowed for.”
Vukasović and Bratko (2015) recently reported a comprehensive
meta-analysis of twin, family, and adoption studies of personality.
They assessed the percentage of individual differences in each
Eysenckian dimension due to genetic factors: for extraversion, it was
39%, for neuroticism it was 42%, and for psychoticism it was 30%.
The above findings indicate genetic factors are less important than
assumed by Hans Eysenck. However, an early study by Eysenck and
Prell (1951) is inconsistent with that conclusion. They reported a very
high heritability estimate of 81% in a twin study on neuroticism assessed
by a gallimaufry of unreliable tests including the body-sway and Ror-
schach tests. An attempted replication of Eysenck and Prell's study by
Blewett (1953) was notably unsuccessful: several tests failed to inter-
correlate as predicted, there was no clear neuroticism factor, and there
was scant evidence of any hereditary determination (Shields, 1954).
3.2. Biological underpinnings
As mentioned earlier, most early American personality researchers
focused on using factor analysis applied to questionnaire data to provide
descriptive accounts of human personality. This approach is intrinsically
limited. Hans Eysenck pointed out that explanatory accounts of person-
ality could be developed by relating individual differences in personality
to their biological underpinnings. This approach was a substantial ad-
vance, and has had a major impact (e.g., Cloninger, Svrakic, &
Przybeck, 1993; Gray, 1981; Pickering, Cooper, Smillie, & Corr, 2013).
 M.W. Eysenck / Personality and Individual Differences 103 (2016) 209–219
3.2.1. Extraversion
Eysenck (1967) explained individual differences in extraversion by
assuming that the threshold for arousal in the ascending reticular
activating system (ARAS) was lower for introverts than for extraverts.
This theoretical approach (owing much to Corcoran, 1965, and Gray,
1964) proved extremely productive and led to a substantial amount of
research.
In spite of many anomalous findings, the evidence overall indicates
that there is typically a negative relationship between extraversion
and arousal (Eysenck, 1982; Gale, 1983). The positive findings include
numerous studies in which manipulations such as sleep deprivation,
noise, and caffeine were used to increase or decrease arousal levels.
Very strong support for the arousal model of extraversion was pro-
vided by Howarth and Eysenck (1968). Their study was based on a the-
ory of memory consolidation assuming that high arousal produces a
strong consolidation process interfering with short-term recall but facil-
itating long-term recall. Kleinsmith and Kaplan (1963) used galvanic
skin responses (GSRs) to assess the arousal associated with each item.
There was a large cross-over interaction between item arousal and re-
tention interval in memory performance: low-arousal items were well
recalled at short retention intervals but poorly at long ones; in contrast,
high-arousal items showed reminiscence (memory improved over
time).
Howarth and Eysenck (1968) predicted a similar cross-over interac-
tion should be obtained comparing low- and high-arousal individuals
(extraverts and introverts, respectively). There was no overall effect
of retention interval (between 0 min and 24 h). However, there
was a dramatic cross-over interaction between personality and
retention interval with a significance level considerably greater than
p b .0000000000000000001. In this interaction (one of the most signif-
icant in the whole of memory research), the memory performance of in-
troverts was approximately 55% greater after 24 h than 0 min.
There are three issues with this research on consolidation theory.
First, there was almost certainly a confounding between item arousal
and intra-list position in the study by Kleinsmith and Kaplan (1963).
Schürer-Necker (1990) found high-arousal items came predominantly
from the start of the list whereas low-arousal items came mostly from
the end of the list. When Schürer-Necker unconfounded these variables,
there was no interaction in memory performance between item arousal
and retention interval.
Second, there have been several failures to replicate Howarth and
Eysenck's (1968) findings even when using a very similar methodology
(e.g., Fuller, 1978, Knezović, 1985; Knezović & Bauer, 2003; McLaughlin,
1968; Schneller & Garske, 1976). No subsequent study has obtained
findings remotely as strong as the original ones of Howarth and
Eysenck.
Third, small reminiscence effects have occasionally been reported in
the literature. However, as Keppel (2014, p. 159) concluded in a review
of the research: “Reminiscence for high-arousal pairs and for high-
arousal subjects has been reported, but not by later researchers who
attempted to replicate the impressive findings.”
In sum, it has often been found that introverted individuals have
higher levels of arousal than extraverted ones. A crucial (but unre-
solved) issue concerns the centrality of high arousal to introversion.
The ARAS is responsible for alertness (Young & Pigott, 1999) and high
ARAS activation is associated with feelings such as ‘energized’, ‘coura-
geous’, and ‘alerted’, whereas low ARAS activation is associated with
feelings such as ‘tired’, ‘weary’, ‘drowsy’, and ‘powerless’ (Siebel,
Winkler, Seitz-Bernhard, & Noosomatik, 1990). However, it is implausi-
ble that you need to be very alert, courageous and energized to be an in-
trovert or tired, powerless, and weary to be an extravert. It is more likely
that high arousal occurs as a consequence (rather than cause) of
introversion.
Another theoretical issue concerns the common finding that arousal
manipulations have more influence on the performance of extraverts
than of introverts. M. Eysenck (1982) speculated that this occurs
211
because introverts have a strongly functioning higher-level control
mechanism that reduces the impact of arousal changes on performance.
Even more speculatively, the greater ARAS activation often found in in-
troverts may in part reflect the functioning of this control mechanism.
3.2.2. Neuroticism
Eysenck's (1967) explanation of neuroticism in terms of individual
differences in activation within the visceral system (hippocampus,
amygdala, cingulum, septum and hypothalamus) was also influential.
However, it de-emphasized the importance of brain networks associat-
ed with cognitive functioning. For example, the dorsolateral prefrontal
cortex and ventrolateral prefrontal cortex play major roles in emotion
regulation via downregulation of the amygdala (Kohn et al., 2014).
Such downregulation is most effective in individuals having the stron-
gest links between prefrontal areas and the amygdala (Lee, Suchday, &
Wylie-Rosett, 2012). Individuals high in neuroticism have much weaker
functional connections between brain areas than low scorers (Servaas
et al., 2015).
According to attentional control theory (M. Eysenck, Derakshan,
Santos, & Calvo, 2007; M. Eysenck, Moser, Derakshan, & Allen,
submitted for publication), individuals high in trait anxiety have im-
paired top-down attentional control combined with greater use of com-
pensatory strategies to minimize the adverse effects of impaired
control. As predicted, high- and low-anxious individuals differ in the
functioning of brain networks of direct relevance to the processes iden-
tified within attentional control theory (Sylvester et al., 2012). These
brain networks include the fronto-parietal network, the ventral atten-
tion network, the default mode network, and the cingulo-opercular
network.
3.2.3. Gray's biological model
Jeffrey Gray (e.g., 1981; Pickering & Gray, 1999) provided an im-
proved biological model based on the notion that trait anxiety
(consisting mainly of neuroticism but also including some introversion)
involves susceptibility to punishment and is linked to the behavioral ap-
proach system (BAS). As predicted, individual differences in neuroti-
cism are associated with volume variations in brain areas involved in
threat, punishment, and negative affect (DeYoung et al., 2010) [this is
Jeremy Gray rather than Jeffrey Gray].
Gray also assumed there are individual differences in susceptibility
to reward linked to the BAS and are plausibly associated with extraver-
sion and impulsivity. As predicted, there is a positive association be-
tween extraversion and volume of medial orbitofrontal cortex, a brain
area involved in coding the reward value of stimuli (DeYoung et al.,
2010). Other research has shown that brain areas associated with re-
ward processing are more activated in extraverts than introverts (e.g.,
Canli et al., 2001).
Gray's (1981) biological approach based on independent behavioral
inhibition system (BIS) and the BAS linked to trait anxiety/neuroticism
and impulsivity/extraversion, respectively, accounts for the existence
of the two orthogonal factors of extraversion (impulsivity) and neurot-
icism (trait anxiety). Research of relevance to Gray's approach has been
obtained in studies focusing on mood state. Two major orthogonal di-
mensions of mood state (positive affect and negative affect) have been
identified (Watson & Tellegen, 1985). There is a moderately strong rela-
tionship between extraversion and positive affect and between neurot-
icism and negative affect (Clark & Watson, 1999; Hepburn & M. Eysenck,
1989; Meyer & Shack, 1989).
Other related research has indicated that subjective well-being is as-
sociated with high extraversion and low neuroticism (Steel, Schmidt, &
Shultz, 2008). It is typically assumed that such findings reflect a direct
influence of personality on subjective well-being. However, recent evi-
dence indicates there are reciprocal influences of personality and well-
being on each other (Soto, 2015).
212 M.W. Eysenck / Personality and Individual Differences 103 (2016) 209–219
3.2.4. Conclusions
In spite of previous theorizing on the biological substrate of person-
ality by Russian psychologists (e.g., Pavlov and Teplov), Hans Eysenck
made a substantial contribution with his biological approach. The gen-
eral biological approach has become increasingly influential over the
several decades since Hans Eysenck first advocated it. His arousal theory
of extraversion has proved very successful based on the findings from
empirical research. At the theoretical level, however, the causal rela-
tionships between extraversion and arousal remain somewhat unclear.
More generally, the extent to which we can understand the differences
between introverts and extraverts in terms of arousal differences has
not been established.
Gray's biological model is an advance on Hans Eysenck's. Of most im-
portance, his model provides a coherent theoretical rationale for the
dominance of orthogonal dimensions of extraversion and neuroticism
in the personality domain. These two dimensions relate to independent
brain mechanisms associated with susceptibility to reward and to pun-
ishment, respectively. In addition, there are two orthogonal mood di-
mensions (positive and negative affect) that are closely related to
extraversion and neuroticism.
4. Continuity between personality and psychopathology
The assumption of continuity between personality and psychopa-
thology has been advocated by numerous theorists (including Freud,
Jung and Kretschmer). However, Hans Eysenck made more use of the
continuity assumption than previous researchers. At the risk of over-
simplification, his key assumptions were as follows:
(1) The various neuroses or anxiety disorders are all related.
(2) High scorers on neuroticism are more susceptible than low
scorers to develop neuroses.
(3) There is a continuum from low levels of neuroticism through to
neurosis with no precise dividing line between normality and ab-
normality.
(4) “All functional psychoses…are related and do not form indepen-
dent categories” (Eysenck, 1992a, p. 670)
(5) High scorers on psychoticism within the healthy population are
more susceptible than low scorers to develop psychoses.
(6) There is a continuum from low levels of psychoticism through to
psychosis with no precise dividing line between normality and
abnormality.
4.1. Evidence and evaluation
Space limitations preclude a detailed discussion of the evidence re-
lating to all six theoretical assumptions. However, a few points can be
made. First, there are similar genetic risk factors across the anxiety dis-
orders. Tambs et al. (2009) conducted a twin study across the anxiety
disorders and concluded, “The latent liability to all anxiety disorders
was substantially more heritable (54%) than the individual disorders
(23% to 40%)” (p. 301). However, note that these findings were mostly
based on individuals possessing sub-clinical levels of the disorders.
Other research suggests that the picture is more complex. Hettema,
Prescott, Myers, Neale and Kendler (2005) found in a twin study that
two genetic factors were shared across the anxiety disorders. One factor
was most strongly associated with generalized anxiety disorder, panic
disorder, and agoraphobia, whereas the other factor was mostly associ-
ated with specific phobia. Social phobia was moderately associated with
both factors.
Second, there is convincing evidence that high neuroticism is associ-
ated with all the major anxiety disorders and with depression (Kotov,
Gamez, Schmidt, & Watson, 2010). Indeed, neuroticism was more
strongly associated with these disorders than were any other personal-
ity trait considered (extraversion, disinhibition, conscientiousness,
agreeableness and openness). However, this research was limited by
being cross-sectional.
Third, other evidence reported by Kotov et al. (2010) is less consis-
tent with Hans Eysenck's approach. Of the six personality traits consid-
ered, conscientiousness was the second most strongly associated with
anxiety disorders and depression, always in the negative direction. Ex-
traversion was the third most strongly associated with anxiety disor-
ders and depression, again always in the negative direction.
Fourth, there is evidence (e.g., Hettema, Neale, Myers, Prescott, &
Kendler, 2006) that genetic factors shared with neuroticism account
for up to 50% of the genetic risk associated with anxiety and depressive
disorders. However, much of the overlap between neuroticism and the
various anxiety disorders may simply reflect ‘criterion overlap’ (Jang,
Wolf, & Larstone, 2006): the personality factor and the mental disorders
are based on rather similar descriptive categories.
Another complication is the finding (Spinhoven, Penelo, de Rooij,
Penninx, & Ormel, 2014) that changes in the symptoms of anxiety disor-
ders or depression were followed by changes in neuroticism. Thus, there
are reciprocal effects of personality and psychopathology rather than
the typically assumed unidirectional impact of personality on
psychopathology.
Fifth, the links between psychoticism and psychosis are typically
weaker than those between psychoticism and criminality and antisocial
behavior. Chapman, Chapman and Kwapil (1994) found in a 10-year
longitudinal study that individuals initially scoring high in psychoticism
were no more likely to develop a psychosis than those scoring lower on
psychoticism. Lonnqvist et al. (2009) found in a longitudinal study that
the onset of schizophrenia was predicted by high neuroticism and low
extraversion.
Sixth, the associations between personality and psychopathology
are often more complex than implied by the above assumptions. For ex-
ample, Andersen and Bienvenu (2011) identified five possible explana-
tions for links between personality and psychopathology: (1) the
vulnerability hypothesis (personality predisposes to mental disorders);
(2) the scar hypothesis (illness adversely affects personality); (3) the
pathoplasty hypothesis (no direct causal relationship between person-
ality and mental disorder; (4) the common cause hypothesis (shared
environmental or genetic causes); (5) criterion overlap hypothesis
(similar descriptive categories for personality and mental disorder).
There is some empirical support for all these explanations and Hans
Eysenck often did not distinguish clearly among them.
5. Personality and mortality
It is very important to show that individual differences in personality
have strong predictive power for major real-world outcomes. The most
dramatic such evidence emerged from Hans Eysenck's collaborative re-
search with Grossarth-Maticek. Grossarth-Maticek, Eysenck and Vetter
(1988) argued that individuals with Type 1 personalities (hopelessness
and depression) are susceptible to cancer, Type 2 personalities (arousal
and aggression) are susceptible to coronary heart disease, and Type 4
personalities (personal autonomy or control) are the healthiest. The rel-
ative risk of Type 1s dying of cancer compared to Type 4s was 55.6 and
for Type 2s the relative risk of dying of heart disease was 58.8 (Lee,
1991).
The above relative risks are almost 20 times greater than those of
smoking. The mystery deepens when we consider questionnaire mea-
sures of Type 1 and Type 2 personalities correlate very highly (+.81)
(Amelang & Schmidt-Rathjens, 1992). This makes it hard to understand
how Type 1 personality could be strongly related to death from cancer
but not coronary heart disease whereas the opposite pattern was
found for Type 2s.
Grossarth-Maticek and Eysenck (1995) assessed mortality over a
15-year period in individuals initially between 45 and 68 years of age.
The correlation between scores on the Self-Regulation Inventory (on
which high scorers have low neuroticism) and mortality was −.79
 M.W. Eysenck / Personality and Individual Differences 103 (2016) 209–219
suggesting self-regulation accounts for 64% of the variance in mortality.
If account is taken of the unreliability of the questionnaire, the correla-
tion becomes −.84 (accounting for 70% of the variance). Even that is
an underestimate given the wide age range (23 years) within the
sample.
5.1. Explanations
How can we explain these various findings (which are hugely more
impressive than those reported by other researchers)? Perhaps the
questionnaires used were markedly superior. However, consider their
items. The following representative 84-word item does not adhere to
the universal recommendation to use short, single-barreled, unambigu-
ous items”.
“Do you change your behavior according to consequences of previ-
ous behavior, i.e., do you repeat ways of acting which have in the
past led to positive results, such as contentment, well-being, self-
reliance, etc., and to stop acting in ways which lead to negative con-
sequences, i.e., to feelings of anxiety, hopelessness, depression, ex-
citement, annoyance, etc.? In other words, have you learned to
give up ways of acting which have negative consequences, and to re-
ly more and more on ways of acting which have positive conse-
quences?”
Attempts to interpret the unique findings of Hans Eysenck and
Grossarth-Maticek were make harder by their reluctance (or refusal)
to provide full information about their methodology and question-
naires. For example, Grossarth-Maticek, Eysenck and Barrett (1993) re-
ported that a large-scale study of theirs starting in 1973 had shown that
a personality questionnaire predicted future cancer and coronary heart
disease accurately only when the interviewer established trust with the
participants and explained the questions. This crucial aspect of the pro-
cedure was not mentioned in any of their previous research publica-
tions. In any case, the question is so complex and convoluted that it
would seem impossible to provide simple explanations of them that
captured all the numerous nuances of meaning they contain.
No unequivocal interpretation of Eysenck and Grossarth-Maticek's
findings is possible. However, an article by van der Ploeg (1991), if accu-
rate, may clarify some issues.
6. Personality research: overall assessment
What is most impressive about Hans Eysenck's entire approach to
human personality is its sheer ambitiousness. Below I evaluate his ap-
proach with respect to the research topics discussed above.
6.1. Personality factors
A potential advantage of Hans Eysenck's ambitious approach with
respect to identifying personality factors is that it is arguably the opti-
mal method to ensure that those identified are of substantial impor-
tance. In contrast, it is possible that some of the personality factors
identified by use of the fundamental lexical hypothesis are only of
minor importance.
Hans Eysenck typically argued that his entire approach was
theoretically-driven but this seems an exaggeration. For example, as
Goldberg pointed out in a letter to Hans Eysenck (February 6, 1995;
quoted in van Kampen, 2009, p. 13): “It is not clear to the world at
large how your PEN model is not well described as an example of the
‘heuristic’ school, given that you explicitly adopted ‘some psychiatric
system of classification’ for P and N, and you used your own ‘notion of
what traits might be important’ to select E [extraversion].”
Hans Eysenck's approach may fail to provide a truly comprehensive
description of human personality. It is arguable that the emphasis on
finding personality traits or factors meaningfully related to categories
213
of mental disorders is in danger of omitting important traits having little
direct relevance to mental disorder (e.g., agreeableness and openness:
Kotov et al., 2010). As discussed above, the preponderance of very low
scores on psychoticism suggests this dimension may lack major rele-
vance within healthy populations.
In contrast, it has been argued that the fundamental lexical hypoth-
esis provides a principled basis for describing all major aspects of
human personality. However, this hypothesis has been criticized on sev-
eral grounds (e.g., Uher, 2013) and it is debatable whether all aspects of
personality are expressible in one word. Despite that, findings obtained
using the lexical hypothesis have typically produced factors overlapping
substantially with those from the five-factor model (e.g., Goldberg,
1990). The loadings of personality terms on the Big Five were excellent
predictors of those terms' relational impact (rated importance of each
personality term when responding to another person) (Wood, 2015) in-
dicating the social importance of all Big Five factors.
6.2. Biological approach
A key assumption of Hans Eysenck's biological approach to personal-
ity was that individual differences in personality depend to a large ex-
tent (60%–80%) on genetic factors. This assumption has received
partial support. As discussed earlier, a recent meta-analysis (Vukasović
& Bratko, 2015) indicated that the mean heritability of the three
Eysenckian personality factors was 37%, which is substantially lower
than assumed by Hans Eysenck.
Of course, genetic factors are important even if they account for
under 40% of individual differences in a given personality factor. How-
ever, the fact that individual differences in personality depend more
on environmental than genetic factors means it is of great importance
to study in detail environmental factors influencing personality
development.
6.3. Personality and cognition
In spite of the very broad and comprehensive nature of Hans
Eysenck's theorizing about personality, he failed to give proper consid-
eration to individual differences in cognitive processes and associated
brain networks. For space reasons, I will discuss these issues primarily
with respect to neuroticism.
Consider the typical finding that individuals high in trait anxiety re-
port spending far more time in the cognitive activity of worrying than
do those low in trait anxiety (M. Eysenck & van Berkum, 1992). Anxious
members of other species probably spend little time worrying about
possible future negative events.
Other research indicates an important role for cognitive factors in in-
dividual differences in neuroticism or trait anxiety. There is consider-
able evidence (M. Eysenck, 1992; M. Eysenck, 1997) indicating that
only those high in neuroticism or trait anxiety exhibit several cognitive
biases (e.g., attentional bias: excessive attention to threat-related stim-
uli; and interpretive bias: a tendency to interpret ambiguous stimuli and
situations in a threatening fashion). Cognitive biases are causally related
to individual differences in anxiety. Cognitive bias modification tech-
niques designed to reduce or eliminate these biases are effective in re-
ducing anxiety (MacLeod & Clarke, 2013).
6.4. Personality and psychopathology
Another major feature of Hans Eysenck's ambitious research pro-
gram was his systematic attempt to integrate theory and research on
personality and psychopathology. This proved only partially successful.
The findings with respect to neuroticism and the anxiety disorders have
mostly supported his predictions: high neuroticism is associated with
all anxiety disorders and with depression; there are common genetic
risk factors found across anxiety disorders; there are commonalities in
the genetic factors underlying high neuroticism and the anxiety
214 M.W. Eysenck / Personality and Individual Differences 103 (2016) 209–219
disorders. However, there are some concerns about criterion overlap
between neuroticism and the anxiety disorders. The most important
unresolved issue concerns causality: cross-sectional studies cannot clar-
ify the extent to which any overlap between high neuroticism and the
anxiety disorders reflects an influence of the former on the latter.
6.5. Validation of personality factors
The value of a personality measure can be assessed via concurrent or
predictive validity. All Hans Eysenck's personality scales demonstrated
reasonable or good validity in these senses. However, his research
with Grossarth-Maticek produced (to my knowledge) the strongest
predictive validity in the entire history of personality research. Their re-
search findings have not been replicated (Amelang, Schmidt-Rathjens,
& Matthews, 1996). Amelang and Schmidt-Rathjens (2003, p. 12)
reviewing findings on personality, cancer and coronary heart disease,
concluded:
“The literature on etiological factors of cancer and coronary heart
disease (CHD) indicates that personality factors are of little impor-
tance. Not more than 1–2% of the heart-disease variance is explained
by personality variables.”
6.6. Lumpers vs. splitters
Charles Darwin in a letter to J. D. Hooke in 1857 introduced the dis-
tinction between lumpers and splitters. The former use broad categories
because they emphasize similarities and de-emphasize differences
whereas the latter use numerous relatively narrow categories because
they emphasize differences rather than similarities. Hans Eysenck was
obviously a lumper rather than a splitter: he identified only three per-
sonality factors, the term ‘neurosis’ was applied to all anxiety disorders,
the term ‘psychosis’ was applied to schizophrenia and bipolar disorder,
and so on.
Is it better for theorists to be lumpers or splitters? There is no
straightforward answer to that question. However, there are major ad-
vantages in being a lumper early in the development of a research area.
Lumpers can provide a comprehensive framework that is very produc-
tive in generating research.
The disadvantages of being a lumper are obvious. First, lumpers' the-
ories are inevitably oversimplified because entities that are clearly dif-
ferent are nevertheless categorized together. For example, the use of a
category such as ‘neurosis’ ignores the differences among the various
anxiety disorders, and the same is true of the category ‘psychosis’. In ad-
dition, it is improbable that the richness of human personality can be
captured by three orthogonal factors.
6.7. Broad vs. narrow theorizing
The distinction between broad versus narrow theories is related to
that between lumpers and splitters. Theorizing in psychology is typical-
ly most productive when it proceeds from the broad or general to the
narrow or specific as we can see by considering cognitive psychology.
Several classic studies in cognitive psychology involve oversimplified
dichotomies (M. Eysenck & Groome, 2015). Examples include the dis-
tinction between episodic and semantic memory; between the dorsal
and ventral streams in perceptual processing; and between analytical
and heuristic processing when making judgments. In all these cases
(and others), the postulation of an extremely broad dichotomy led
over time to richer and more complex theoretical accounts.
The importance of dichotomies within Hans Eysenck's theorizing
can easily be seen. Theorists may choose between several dichotomies:
(1) do individual differences in personality depend in part on genetic
factors?; (2) does factor analysis provide an adequate account of per-
sonality; (3) is it preferable to identify orthogonal or oblique personality
factors?; (4) are anxiety disorders independent of each other or relat-
ed?; (5) are psychoses independent of each other or related?; (6) is a di-
mensional or categorical approach preferable to an understanding of
mental disorder?
It is arguable that Hans Eysenck consistently selected the more ap-
propriate answer to these dichotomies. However, it is essential at
some point to move away from simple dichotomies in order to provide
more nuanced accounts of human personality.
7. Other contributions
Hans Eysenck produced numerous articles and books on topics not
directly associated with personality. Here I will consider a few of his
best-known contributions outside personality research.
7.1. Behavior therapy
Hans Eysenck provided an early comprehensive theoretical rationale
for behavior therapy (Eysenck, 1959) based squarely on Pavlovian
learning theory. For example, he assumed symptoms are maladaptive
conditioned responses, effective treatment involves extinguishing mal-
adaptive autonomic and skeletal conditioned responses and establish-
ing desirable conditioned responses.
Hans Eysenck was a very effective proselytizer for behavior therapy
in the late 1950s and early 1960s, in part through his founding of the
journal Behavior Research and Therapy. However, his influence waned
over the years because his views changed remarkably little. Here is an
example:
“The different types of neurotic illness arise through a process of
Pavlovian conditioning, and can be eliminated through a process of
Pavlovian extinction...all cures of neurotic disorders are based on
Pavlovian extinction.” (Eysenck, 1987, pp. 6-7)
It would be supererogatory to document all the counter-evidence to
these views. For example, the etiology of the anxiety disorders depends
on vicarious and informational pathways as well as Pavlovian condi-
tioning (Rachman, 2015). In addition, there is increased differentiation
of effective forms of therapy for the various anxiety disorders.
The dramatic impact of cognitive and cognitive-behavioral therapy
on the treatment of anxiety disorders has revealed the limitations of
an excessive emphasis on behavior therapy and the changing of behav-
ioral patterns. However, this neglect of cognitive processes and struc-
tures is entirely consistent with Hans Eysenck's neglect of the
cognitive system when accounting for individual differences in
personality.
7.2. Social attitudes
Eysenck (1954) identified two orthogonal dimensions of social atti-
tudes: radicalism and toughmindedness. Of major importance, a subse-
quent twin study by Eaves and Eysenck (1974) reported that individual
differences in social attitudes have a fairly strong genetic influence:
the heritability for conservatism was .65 and it was .54 for
toughmindedness. They also found the relationship between social atti-
tudes and personality was due primarily to latent shared genetic influ-
ences. Subsequent research has replicated this finding and indicated
an absence of causal influences of personality on social attitudes
(e.g., Hatemi & Verhulst, 2015).
These findings are of major importance because they identify strong
influences on social attitudes typically ignored by social psychologists.
For example, Tesser and Shaffer (1990) in a review of research on atti-
tudes reported no references at all to heritability in the social psycholo-
gy literature.
Hogg and Vaughan (2005) failed to mention genetic influences in a
chapter on attitudes in their major textbook on social psychology.
 M.W. Eysenck / Personality and Individual Differences 103 (2016) 209–219 215

Research subsequent to the study by Eaves and Eysenck (1974) indi- Striking disproof of the above assumptions came in a study by
cated a genetic influence on 26 social attitudes (Olson, Vernon, Harris, & Brooks-Gunn, Klebanov and Duncan (1996) on black and white
Jang, 2001). The exciting (and to many psychologists very surprising) American 5-year-olds. The mean IQ of the black children was 1 standard
implication is that individual differences in a very large variety of atti- deviation lower than that of the white children. However, “Adjustments
tudes and behavior may depend in part on genetic factors although pre- for economic and social differences in the lives of black and white chil-
cisely how these effects are mediated remains to be determined. dren all but eliminate differences in the IQ sores between the two
groups” (Brooks-Gunn et al., p. 396).
Further evidence disconfirming Hans Eysenck's assumptions was re-
7.3. Intelligence: event-related potentials (ERPs)
ported by Turkheimer, Haley, Waldron, D'Onofrio and Gottesman
(2003) in a study on American children. Within affluent families of
Hendrickson and Hendrickson (1980), working in Hans Eysenck's
high socioeconomic status, genetic factors accounted for 72% of individ-
Department, developed a ‘string’ measure based on event-related po-
ual differences in intelligence. In contrast, genetic factors accounted for
tentials (ERPs) which involved fitting a piece of string to the waveform
only 10% of individual differences in intelligence among impoverished
of each ERP record for the first 250 to 500 ms after stimulus presenta-
families of low socioeconomic status. The relevance of these findings
tion. Using this measure, Elaine Hendrickson (1982) reported a correla-
is that Afro-American families are disproportionately of low socioeco-
tion of −.72 between IQ and string length. This finding was interpreted
nomic status.
as supporting the hypothesis that more intelligent individuals make
In sum, many arguments adduced by Hans Eysenck in the book Race,
fewer errors in processing and process information more efficiently.
Intelligence and Education are fallacious. This seems especially irrespon-
Eysenck (in Eysenck & M. Eysenck, 1989, p. 112) was very confident
sible given the very socially sensitive nature of the issues discussed in it.
about the value of ERPs: “There seems to be no doubt that we can now
measure intelligence physiologically with an accuracy that compares fa-
8. The research enterprise: scientists vs. lawyers
vorably with the best IQ tests currently available.” The phrase “no
doubt” is perhaps debatable: Rust (1975) and Shagass, Roemer,
How should researchers conduct themselves? A consensual view
Straumanis and Josiassen (1981) had previously reported non-
was expressed by Cournand (1977). First, he argued scientists must pos-
significant findings; Vetterli and Furedy (1985) reported a correlation
sess intellectual integrity and objectivity, “They must prevent their de-
between the string measure and IQ in the wrong direction; Vogel,
sires and aversions from penetrating their observations of the
Kruger, Schalt, Schnobel and Hassling (1987) in the largest study in
phenomena that they study and their analyses of these observations.”
this area of research found that none of the numerous correlations
(p. 700).
they calculated between various ERP measures and IQ scores even
Second, scientists should have a “doubt of certitude” including a
approached statistical significance.
willingness to acknowledge and admit to errors. The same basic point
Barrett and Eysenck (1992a) reported 8 correlations between ERP
was made eloquently by Feynman (1985, p. 341): “If you're doing an ex-
string length and IQ ranging from −.20 to −.44 with every correlation
periment, you should report everything that you think might make it in-
in the opposite direction to prediction. These findings prompted
valid — not only what you think is right about it: other causes that could
Barrett and Eysenck (1992b) into a dramatic volte face from the total
possibly explain your results … if you make a theory, … then you must
confidence of Eysenck in Eysenck and M. Eysenck (1989): “With regard
also put down all the facts that disagree with it.”
to the various studies of AEP [average evoked potential] correlates of IQ,
Third, scientists should display unselfish engagement with respect
…for every study that claims a ‘significant’ result, there is another study
to the growth of scientific knowledge. Their focus should be on increas-
that claims the opposite … no confidence in correlations between AEP
ing understanding rather than on self-aggrandizement.
parameters and IQ can be expressed.”
The above idealized version of the scientist's approach can be
It is unclear theoretically why we would expect positive findings
contrasted with a caricatured version of the lawyer's approach (my
from tasks imposing practically no cognitive demands. A superior strat-
apologies to any lawyers who happen to read this article!). In essence,
egy is to focus on brain areas (especially within the prefrontal cortex)
the stereotypical view is that lawyers present information very selec-
most closely associated with complex cognition. For example, Basten,
tively (perfectly acceptable within our adversarial legal system).
Hilger and Fiebach (2015) reported a meta-analysis of functional
When defending a client, they emphasize all the evidence pointing in
brain-imaging studies in which participants performed cognitive tasks.
the direction of his/her innocence while ignoring, de-emphasizing,
The findings supported the parieto-frontal integration theory of intelli-
distorting, or dismissing evidence suggestive of guilt.
gence: there were generally positive associations between intelligence
and activation levels in frontal and parietal brain regions. There was lit-
8.1. Hans Eysenck: scientist or lawyer?
tle support for the neural efficiency hypothesis underlying Hans
Eysenck's ERP research program – the majority of brain regions showed
Hans Eysenck consistently claimed he followed a rigorous scientific
a positive (rather than a negative) association between intelligence and
approach to research. Much of relevance on the issue of whether Hans
activation level.
Eysenck adopted a scientist's or a lawyer's approach was discussed by
Buchanan (2010) in his biography. His account (so accurate that no un-
7.4. Race and intelligence biased person could fail to share most of his views) strongly suggests
Hans Eysenck's approach was considerably closer to that of a lawyer.
Hans Eysenck's best-known contribution to the study of intelligence Hans Eysenck often discussed findings in a way lacking objectivity.
was his 1971 book, Race, Intelligence and Education, in which it was ar- For example, Eysenck (1952) claimed that patients with neurotic disor-
gued that some of the observed mean IQ differences between white ders were much more likely to recover if they received no systematic
and black individuals are probably due to genetic factors. The book con- treatment (spontaneous recovery) than if they received psychoanalysis.
tains even more very incorrect assumptions than generally assumed. However, the median length of treatment for the psychotherapy groups
Here are three: “The environmentalist theory is…too vague to be tested was approximately 15 weeks (unscientifically, this information is miss-
at all” (p. 119); American whites and Afro-Americans “share a common ing in the original article). If we compare improvement over a standard
cultural heritage and a common educational system…different racial length of time, psychoanalysis was moderately effective (McNeilly &
populations growing up within the same culture are being measured” Howard, 1991).
(p. 77); and “Interaction effects [between genetic and environmental Another distortion of the evidence occurred in an article (Eysenck,
factors] are rather unimportant” (p. 117). 1991b, p. 330) describing the findings of research on vitamins and IQ
216 M.W. Eysenck / Personality and Individual Differences 103 (2016) 209–219
by Schoenthaler, Amos, Eysenck, Peritz and Yudkin (1991): “A reason-
ably random selection of school children…contains a sizable number
whose IQ can be raise through [vitamin] supplementation, compared
to a control group, by some 11 points of IQ.” In fact, three IQ measures
were obtained before and after vitamin administration and there were
three different vitamin doses. Only one of the nine vitamin-control com-
parisons was statistically significant. This finding involved an IQ gain av-
eraging 3.7 points greater than the control-group gain and involved
only non-verbal IQ.
There are several distorted interpretations of the evidence on the
dangers of smoking in Eysenck (1991a). Here are three. First, he
claimed that there was very little evidence that smoking by itself is
associated with cancer or coronary heart disease. In fact, there was
pre-existing compelling evidence indicating that chronic smokers
on average lose 10 years of life through smoking (see Doll, Peto,
Boreham, & Sutherland, 2004).
Second, Eysenck (1991a) argued that it was wholly inappropriate
to use the word ‘addiction’ when describing smoking behavior. This
argument makes it hard to explain the behavior of patients with
Buerger's disease – “If a patient with the condition continues to
smoke, gangrene may eventually set in. First a few toes may have
to be amputated, then the foot at the ankle, then the leg at the
knee, and ultimately at the hip…Patients are strongly advised that
if only they will stop smoking,…the otherwise inexorable march of
gangrene up the legs will be curbed…it is not at all uncommon to
find a patient with Buerger's disease vigorously puffing away in his
hospital bed following a second or third operation” (Brecher, 1972).
Third, Eysenck (1991a) argued that exposure to environmental to-
bacco smoke was probably not a factor in cancer or coronary heart dis-
ease. Contrary evidence was available at the time (Peto & Doll, 1986).
Compelling evidence was reported by Jee, Ohrr and Kim (1999) in a
large-scale study of non-smoking Korean married women. If the
husband had smoked for more than 30 years, they had a 230% increased
probability of developing lung cancer relative to the wives of non-
smokers.
8.2. Does it matter?
Does it matter that Hans Eysenck typically adopted the lawyer's
approach to research? At least three arguments can be adduced to
suggest it does not. First, there is a continuum between scientists'
and lawyers' approaches and most scientists are not totally objective
and unbiased when interpreting their own and other researchers'
data. The location of the dividing line between the acceptable and
unacceptable is a matter of personal judgment.
Second, Hans Eysenck came up with an ingenious defense of the
lawyer's approach: as he said, “Scientists' … finest and most original dis-
coveries are rejected by the vulgar mediocrities filling the ranks of or-
thodoxy … The figures do not quite fit, so why not fudge them a little
bit? Usually, the genius is right, of course…, and we may in retrospect
excuse the childish games, but clearly this cannot be regarded as a li-
cense for non-geniuses to foist their absurd beliefs on us. Such, then,
are the conditions in which leading scientists may be led to falsify
their data, or invent them. Convinced (rightly as it happens) that they
have made a stupendous discovery, …they see that discovery threat-
ened by enemies or hostile people…The first findings in trying to sub-
stantiate a theoretical discovery are never clear cut … But enemies
would seize upon these anomalies to destroy his theory. Obviously the
way to overcome this problem is simply to make sure the data fit the
theory!” (Eysenck, 1995a, p. 126).
Perhaps the above quotation is intended as a joke although the notion
that geniuses can (and perhaps should) suit themselves was reiterated
elsewhere (Eysenck, 1995b). However, the context is a lengthy discussion
of evidence indicating that Newton manipulated some of his data and
theoretical assumptions. The crucial point is that Hans Eysenck seems to
endorse Newton's deception. It is true that Hans Eysenck claimed that
he was not a genius on the grounds of being insufficiently high in
psychoticism. However, he repeatedly claimed to be the most cited living
psychologist (e.g., Eysenck, 1997) so perhaps his modest claim should not
be taken at face value.
Third, it is perhaps unfair to view Hans Eysenck's approach from
our contemporary perspective. As Hartley (1953, p. 1) pointed out:
“The past is a foreign country: they do things differently there.”
Perhaps Hans Eysenck's approach was more acceptable decades
ago when theory and research in psychology were at a primitive
stage.
The past also differs from the present in that various countercul-
ture movements were very influential during the 1960s. For
example, Paul Feyerabend, a philosopher of science, advocated
epistemological anarchism. He summarized his approach as follows:
“Science is an essentially anarchistic enterprise…The only principle
that does not inhibit progress is anything goes” (Feyerabend, 1975,
p. 1). According to epistemological anarchism, science advances by
a “who-shouts-the-loudest” strategy in which publicity and visibility
count for more than the intrinsic quality of the research. A central
problem with Feyerabend's views is that they imply there is no
clear dividing line between science and non-science.
9. Summary and conclusions
How can we sum up the overall impact of Hans Eysenck's enormous
research output on numerous diverse topics? Perhaps the most striking
overall characteristic is that the variance in the quality of his research
was greater than that of anyone else in the history of psychology
(with the possible intriguing exception of Sigmund Freud).
On the positive side, Hans Eysenck made key contributions including
the following:
(1) His emphasis on a few major orthogonal factors rather than nu-
merous oblique ones.
(2) The identification of the two most important personality factors
(extraversion and neuroticism) together with extensive research
to demonstrate their importance.
(3) His argument that factor analysis is intrinsically limited in that it
provides a descriptive but not an explanatory account of person-
ality.
(4) Arguing for (and helping to demonstrate) an important role for
genetic factors in accounting for individual differences in person-
ality and social attitudes.
(5) Developing theories to explain individual differences in person-
ality in terms of underlying biological/physiological mechanisms.
(6) His emphasis on the continuity hypothesis that there is no firm
dividing line between normality and psychopathology.
(7) His emphasis on the hypothesis that high levels of neuroticism
and psychoticism are risk factors for neuroses and psychoses,
respectively.
In my opinion, Hans Eysenck probably contributed more than anyone
else to the development of personality psychology. Decades ago he set out
a paradigm or agenda for a comprehensive approach to theory and re-
search in personality (represented by the major contributions identified
above). This paradigm remains current to this day. In other words, Hans
Eysenck largely asked the most important questions concerning individu-
al differences in personality. As Einstein pointed out: “The formulation of
a problem is often more essential than its solution…To raise new
questions, new problems, to regard old problems from a new angle,
requires creative imagination and makes real advances in science.”
What are the limitations of Hans Eysenck's contribution? First, several
of his publications (and those of his collaborators) reported dramatic and
counterintuitive findings of potentially great theoretical importance that
subsequently failed to replicate. For example, Eysenck's (1952) findings
 M.W. Eysenck / Personality and Individual Differences 103 (2016) 209–219
claiming that psychotherapy reduces the probability of recovery from
mental illness are very different from those of most other researchers.
Other examples include the claim that genetic factors account for 80% of
individual differences in neuroticism (Eysenck, 1952), that longevity
and the disease from which you will die depend almost totally on person-
ality (Grossarth-Maticek & Eysenck, 1995; Grossarth-Maticek et al.,
1988), that intelligence can be assessed very accurately by ERPs
(Hendrickson, 1982), and that introverts show a very strong reminiscence
effect in memory whereas extraverts have a typical forgetting effect
(Howarth & Eysenck, 1968).
Another limitation is that Hans Eysenck's interpretations of his find-
ings cannot always be taken at face value because he omitted crucially
important information. Here are two examples (referred to earlier).
First, when Eysenck (1952) compared the effectiveness of spontaneous
recovery and psychotherapy, he failed to indicate that the time period
within which recovery could occur was 700% greater for patients in
the former group than those in the latter group.
Second, much of the research by Hans Eysenck and Grossart-Maticek
on personality and longevity (e.g., Grossarth-Maticek et al., 1988) failed
to mention that the findings depended critically on the person adminis-
tering the questionnaire generating trust and providing detailed explana-
tions of the questions. The importance of the factor was only revealed
several years later (Grossarth-Maticek et al., 1993).
Hans Eysenck sometimes provided partisan interpretations of the
findings of others. Examples discussed earlier include his rejection of
the evidence indicating clearly the dangers of smoking (Eysenck,
1991a), his endorsement of the notion that race differences in IQ owe
much to genetic factors (Eysenck, 1971), and the assertion that all cures
of anxiety disorders are based on Pavlovian conditioning.
Third, Hans Eysenck systematically de-emphasized the importance of
cognitive processes and structures. Since I am a cognitive psychologist, it
is certainly possible that my views here are biased. However, there is
compelling evidence that individual differences in neuroticism depend
importantly on various cognitive biases (M. Eysenck, 1992; M. Eysenck,
1997). There is also strong evidence that cognitive therapy is an effective
form of treatment for all the anxiety disorders (Clark & Wells, 1997)
which is inconsistent with Hans Eysenck's emphasis on changing behav-
ior through conditioning.
Finally, Hans Eysenck differed from the overwhelming majority of
researchers and theorists in lacking any ‘doubt of certitude’. One of my
main recurring thoughts while writing this article echoes a famous
quotation by William Lamb (2nd Viscount Melbourne) about the poet,
historian, and politician Thomas Macaulay:
“I wish I were as cocksure of anything as Tom Macaulay is of every-
thing.”
References
Amelang, M., & Schmidt-Rathjens, C. (1992). Psychometric properties of modified
Grossarth-Maticek and Eysenck inventories. Psychological Reports, 71, 1251–1263.
Amelang, M., & Schmidt-Rathjens, C. (2003). Personality, cancer and coronary heart dis-
ease: Fictions and facts in the etiological research. Psychologische Rundschau, 54,
12–23.
Amelang, M., Schmidt-Rathjens, C., & Matthews, G. (1996). Personality, cancer and coro-
nary heart disease: Further evidence on a controversial issue. British Journal of Health
Psychology, 1, 191–205.
Andersen, A. M., & Bienvenu, O. J. (2011). Personality and psychopathology. International
Review of Psychiatry, 23, 234–247.
Barrett, P. T., & Eysenck, H. J. (1992a). Brain evoked potentials and intelligence: The
Hendrickson paradigm. Intelligence, 16, 361–382.
Barrett, P. T., & Eysenck, H. J. (1992b). Brain evoked potentials and intelligence. In A. Gale,
& M. W. Eysenck (Eds.), The handbook of individual differences: Biological perspectives,
Vol. 3. (pp. 255–285). Chichester: Wiley.
Barrett, P. T., & Kline, P. (1982). An item and radial parcel analysis of the 16PF question-
naire. Personality and Individual Differences, 3, 259–270.
Basten, U., Hilger, K., & Fiebach, C. J. (2015). Where smart brains are different: A quantitative
meta-analysis of functional and structural brain-imaging studies on intelligence.
Intelligence, 51, 10–27.
217
Blewett, D. B. (1953). An experimental study of the inheritance of neuroticism. (Ph.D. thesis)
University of London.
Brecher, E. M. (1972). The case of dr. Sigmund Freud. The consumers union report on licit
and illicit drugs (http://www.druglibrary.org/schaffer/library/studies/cu/cu24.html).
Brooks-Gunn, J., Klebanov, P. K., & Duncan, G. J. (1996). Ethnic differences in children's intel-
ligence test scores: Role of economic deprivation, home environment, and maternal
characteristics. Child Development, 67, 396–408.
Buchanan, R. D. (2010). Playing with fire: The controversial career of Hans J. Eysenck.
Oxford: Oxford University Press.
Burt, C. (1915). The general and specific factors underlying the primary emotions. Report to
the British Association for the Advancement of Science, Vol. 69. (pp. 45).
Burt, C. (1940). Factors of the mind: An introduction to factor-analysis in psychology.
London: University of London Press.
Canli, t., Zhao, Z., Desmond, J. E., Kang, E. J., Gross, J., & Gabrieli, J. D. E. (2001). An fMRI study
of personality influences on brain reactivity to emotional stimuli. Behavioral
Neuroscience, 115, 33–42.
Cattell, R. B., Eber, H. W., & Tatsouka, M. M. (1970). Handbook for the sixteen personality
factor questionnaire (16 PF). Champaign, Ill: Institute for Personality and Ability
Testing.
Chapman, J. P., Chapman, L. T., & Kwapil, T. R. (1994). Does the Eysenck psychoticism scale
predict psychosis? A ten-year longitudinal study. Personality and Individual Differences,
17, 369–375.
Claridge, G., Robinson, D. L., & Birchall, P. (1983). Characteristics of schizophrenics' and
neurotics' relatives. Personality and Individual Differences, 4, 651–664.
Clark, D. M., & Wells, A. (1997). Cognitive therapy for anxiety disorders. Review of
Psychiatry, 16, 1–9.
Clark, L. A., & Watson, D. (1999). Temperament: A new paradigm for trait anxiety. In L. A.
Pervin, & O. P. John (Eds.), Handbook of personality: Theory and research (pp. 265–286)
(3rd. ed.). New York, NY: Guilford Press.
Cloninger, C. R., Svrakic, D. M., & Przybeck, T. R. (1993). A psychobiological model of temper-
ament and character. Archives of General Psychiatry, 50, 975–990.
Corcoran, D. W. J. (1965). Personality and the inverted-U relation. British Journal of
Psychology, 56, 267–273.
Corr, P. J. (2010). The psychoticism–psychopathy continuum: A neuropsychological
model. Personality and Individual Differences, 48, 695–703.
Costa, P. T., & McCrae, R. R. (1992). Revised NEO Personality Inventory (NEO-PI-R) and NEO
Five Factor Inventory (NEO-FFI) Professional Manual. Odessa, FL: Psychological Assess-
ment Resources.
Cournand, A. (1977). The code of the scientist and its relationship to ethics. Science, 198,
699–705.
DeYoung, C. G., Hirsh, J. B., Shane, M. S., Papademetris, X., Rajeevan, N., & Gray, J. R. (2010).
Testing predictions from personality neuroscience: Brain structure and the Big Five.
Psychological Science, 21, 820–828.
Doll, R., Peto, R., Boreham, J., & Sutherland, I. (2004). Mortality in relation to
smoking: 50 years' observations on male British doctors. British Medical
Journal, 328, 1519–1528.
Eaves, L. J., & Eysenck, H. J. (1974). Genetics and the development of social attitudes.
Nature, 249, 288–289.
Eysenck, H. J. (1952). The effects of psychotherapy: An evaluation. Journal of Consulting
Psychology, 16, 319–324.
Eysenck, H. J. (1954). The psychology of politics. London: Routledge and Kegan Paul.
Eysenck, H. J. (1959). Learning theory and behavior therapy. Journal of Mental Science, 105,
61–75.
Eysenck, H. J. (1967). The biological basis of personality. Springfield, Ill: Charles C. Thomas.
Eysenck, H. J. (1971). Race, intelligence and education. London: Temple Smith.
Eysenck, H. J. (1979). Genetic model, theory of personality and the unification of psychology.
In J. R. Royce, & J. P. Mos (Eds.), Theoretical advances in behavior genetics (pp. 517–540).
Alphen aan den Rijn: Sitjhoff and Noordhoff.
Eysenck, H. J. (1987). Behavior therapy. In H. J. Eysenck, & I. Martin (Eds.), Theoretical
foundations of behavior therapy (pp. 3–35). New York: Plenum.
Eysenck, H. J. (1991a). Smoking, personality, and stress: Psychosocial factors in the preven-
tion of cancer and coronary heart disease. Berlin, Germany: Springer Verlag.
Eysenck, H. J. (1991b). Raising IQ through vitamin and mineral supplementation: An intro-
duction. Personality and Individual Differences, 12, 329–333.
Eysenck, H. J. (1992a). 4 ways 5 factors are not basic. Personality and Individual Differences,
13, 667–673.
Eysenck, H. J. (1992b). The definition and measurement of psychoticism. Personality and
Individual Differences, 13, 757–785.
Eysenck, H. J. (1994). The importance of theory in the taxonomy of personality. In B. De
Raad B, W. K. B. Hofstee, & G. L. Van Heck (Eds.), Personality psychology in Europe,
Vol. 5. (pp. 6–13). Tilburg: Tilburg University Press.
Eysenck, H. J. (1995a). Burt as hero and anti-hero: A Greek tragedy. In N. J. Mackintosh
(Ed.), Burt: Fraud or framed? (pp. 111–129). Oxford: Oxford University Press.
Eysenck, H. J. (1995b). Genius: The natural history of creativity. Cambridge: Cambridge Uni-
versity Press.
Eysenck, H. J. (1997). Rebel with a cause: The autobiography of Hans Eysenck (2nd ed.).
Piscataway, NJ: Transaction Publishers.
Eysenck, H. J., & Eysenck, M. W. (1985). Personality and individual differences: A natural sci-
ence approach. New York: Plenum.
Eysenck, H. J., & Eysenck, M. W. (1989). Mindwatching (2nd ed.). London: Prion/
Multimedia.
Eysenck, H. J., & Prell, D. B. (1951). The inheritance of neuroticism: An experimental study.
Journal of Mental Science, 97, 441–465.
Eysenck, M. W. (1982). Attention and arousal: Cognition and performance. Berlin: Springer
Verlag.
Eysenck, M. W. (1992). Anxiety: The cognitive perspective. Hove, UK: Psychology Press.
218 M.W. Eysenck / Personality and Individual Differences 103 (2016) 209–219
Eysenck, M. W. (1997). Anxiety and cognition: A unified theory. Hove, UK: Psychology
Press.
Eysenck, M. W. (2013). Lost in shadows? In B. D. Kirkcaldy (Ed.), Chimes of time: Wounded
health professionals. Essays in recovery. Leiden: Sidestone Press.
Eysenck, M. W., & Groome, D. (Eds.). (2015). Cognitive psychology: Revisiting the classics.
London: Sage.
Eysenck, M. W., & van Berkum, J. (1992). Trait anxiety, defensiveness, and the structure of
worry. Personality and Individual Differences, 13, 1285–1290.
Eysenck, M. W., Derakshan, N., Santos, R., & Calvo, M. G. (2007). Anxiety and cognitive
performance: Attentional control theory. Emotion, 7, 336–353.
Eysenck, M. W., Moser, J. S., Derakshan, N., & Allen, P. (2016). Trait anxiety, processing
efficiency and performance effectiveness: Developing attentional control theory
(submitted for publication) .
Feyerabend, P. (1975). Against method: Outline of an anarchist theory of knowledge.
London: New Left Books.
Feynman, R. P. (1985). Surely you're joking, Mr. Feynman: Adventures of a curious character.
New York: W. W. Norton.
Fuller, A. R. (1978). Personality and paired-associate learning. International Journal of
Psychology, 13, 123–128.
Gale, A. (1983). Electroencephalographic studies of extraversion–introversion: A case
study in the psychophysiology of individual differences. Personality and Individual
Differences, 4, 371–380.
Goldberg, L. R. (1990). An alternative “description of personality”: The Big-Five
factor structure. Journal of Personality and Social Psychology, 59, 1216–1229.
Gray, J. A. (1964). Strength of the nervous system and levels of arousal: A reinterpretation.
In J. A Gray (Ed.), Pavlov's typology. Oxford: Pergamon Press.
Gray, J. A. (1981). A critique of Eysenck's theory of personality. In H. J. Eysenck (Ed.), A
model for personality (pp. 246–279). Berlin: Springer Verlag.
Grossarth-Maticek, R., & Eysenck, H. J. (1995). Self-regulation and mortality from cancer,
coronary heart disease and other causes: A prospective study. Personality and
Individual Differences, 19, 781–795.
Grossarth-Maticek, R., Eysenck, H. J., & Barrett, P. (1993). Prediction of cancer and coro-
nary heart disease as a function of method of questionnaire administration.
Psychological Reports, 73, 943–959.
Grossarth-Maticek, R., Eysenck, H. J., & Vetter, H. (1988). Personality type, smoking habit
and their interaction as predictors of cancer and coronary heart disease. Personality
and Individual Differences, 9, 479–495.
Guilford, J. P. (1934). Introversion–extraversion. Psychological Bulletin, 31, 331–354.
Guilford, J. P. (1939). General psychology. New York, NY: Van Nostrand.
Hartley, L. P. (1953). The go-between. London: Hamish Hamilton.
Hatemi, P. K., & Verhulst, B. (2015). Political attitudes develop independently of personal-
ity traits. PloS One, 10(3).
Hendrickson, D. E. (1982). The biological basis of intelligence: 2. Measurement. In H. J.
Eysenck (Ed.), A model for intelligence (pp. 197–228). New York: Springer Verlag.
Hendrickson, D. E., & Hendrickson, A. E. (1980). The biological basis of individual differ-
ences in intelligence. Personality and Individual Differences, 1, 3–33.
Hepburn, L., & Eysenck, M. W. (1989). Personality, average mood and mood variability.
Personality and Individual Differences, 10, 975–983.
Hettema, J. M., Neale, M. C., Myers, J. M., Prescott, C. A., & Kendler, K. S. (2006). A
population-based study of the relationship between neuroticism and internalizing
disorders. American Journal of Psychiatry, 163, 857–864.
Hettema, J. M., Prescott, C. A., Myers, J. M., Neale, M. C., & Kendler, K. S. (2005). The struc-
ture of genetic and environmental risk factors for anxiety disorders in men and
women. Archives of General Psychiatry, 41, 1987–1995.
Hogg, M. A., & Vaughan, G. M. (2005). Social psychology (4th ed.). Harlow, UK: Pearson
Education Limited.
Howarth, E., & Eysenck, H. J. (1968). Extraversion, arousal, and paired-associate learning.
Journal of Experimental Research in Personality, 3, 114–116.
Jang, K., Wolf, H., & Larstone, R. (2006). What is the role of personality in psychopatholo-
gy? A view from behavior genetics. In R. F. Krueger, & J. L. Tackett (Eds.), Personality
and psychopathology (pp. 166–186). New York: Guilford Press.
Jee, S. H., Ohrr, H., & Kim, I. S. (1999). Effects of husbands' smoking on the incidence
of lung cancer in Korean women. International Journal of Epidemiology, 28,
824–828.
Jung, C. G. (1923). Psychological types. London: Routledge & Kegan Paul.
van Kampen, D. (2009). Personality and psychopathology: A theory-based revision
of Eysenck's PEN model. Clinical Practice & Epidemiology in Mental Health, 5,
9–21.
Keppel, G. (2014). Consolidation and forgetting. In H. Weingartner, & E. S. Parker (Eds.),
Memory consolidation: Psychobiology of cognition (pp. 149–162). New York: Psychology
Press.
Kleinsmith, L. J., & Kaplan, S. (1963). Paired-associate learning as a function of arousal and
interpolated interval. Journal of Experimental Psychology, 65, 190–193.
Knezović, Z. (1985). Multivariate and classical-experimental test of unitarity of Eysenck's
dimensions of introversion–extraversion. Unpublished doctoral dissertation,
University of Zagreb, Croatia.
Knezović, Z., & Bauer, K. (2003). Extraversion and paired-associate recall. Review of
Psychology, 10, 125–130.
Kohn, N., Eickhoff, S. B., Scheller, M., Laird, A. R., Fox, P. T., & Habel, U. (2014). Neural net-
work of cognitive emotion regulation: An ALE meta-analysis and MACM analysis.
NeuroImage, 87, 345–355.
Kotov, R., Gamez, W., Schmidt, F., & Watson, D. (2010). Linking “big” personality traits to
anxiety, depression, and substance abuse disorders: A meta-analysis. Psychological
Bulletin, 136, 768–821.
Lee, P. N. (1991). Personality and disease: A call for replication. Psychological Inquiry, 2,
251–253.
Lee, Y. S. C., Suchday, S., & Wylie-Rosett, J. (2012). Perceived social support, coping styles,
and Chinese immigrants' cardiovascular responses to stress. International Journal of
Behavioral Medicine, 19, 174–185.
Lonnqvist, J., Verkasalo, M., Haukka, J., Nyman, K., Tiihonen, J., Laaksonen, I., et al.
(2009). Premorbid personality factors in schizophrenia and bipolar disorder:
Results from a large cohort study of male conscripts. Journal of Abnormal
Psychology, 118, 418–423.
MacLeod, C., & Clarke, P. J. F. (2013). Cognitive bias modification: A new frontier in
cognition and emotion research. In M. D. Robinson, E. Watkins, & E. Harmon-
Jones (Eds.), Handbook of cognition and emotion (pp. 540–562). New York:
Guilford Publications.
McCrae, R. R., & Costa, P. T. (1985). Updating Norman's “adequate taxonomy”: Intelligence
and personality dimensions in natural language and in questionnaires. Journal of
Personality and Social Psychology, 49, 710–721.
McCrae, R. R., & Costa, P. T. (1997). Personality trait structure as a human universal.
American Psychologist, 52, 509–516.
McLaughlin, R. J. (1968). Retention in paired-associate learning related to extroversion
and neuroticism. Psychonomic Science, 13, 333–334.
McNeilly, C. L., & Howard, K. I. (1991). The effects of psychotherapy: A re-evaluation
based on dosage. Psychotherapy Research, 1, 74–78.
Meyer, G. J., & Shack, J. R. (1989). Structural convergence of mood and personality: Evi-
dence for old and new directions. Journal of Personality and Social Psychology, 57,
691–706.
Olson, J. M., Vernon, P. A., Harris, J. A., & Jang, K. L. (2001). The heritability of
attitudes: A study of twins. Journal of Personality and Social Psychology, 80,
845–860.
Peto, J., & Doll, R. (1986). Passive smoking. British Journal of Cancer, 54, 381–383.
Pickering, A. D., & Gray, J. A. A. (1999). The neuroscience of personality. In L. Pervin, & O.
John (Eds.), Handbook of personality (pp. 277–299) (2nd ed.). New York: Guilford
Press.
Pickering, A. D., Cooper, A. J., Smillie, L. D., & Corr, P. J. (2013). On the shoulders of giants.
The Psychologist, 26, 22–25.
van der Ploeg, H. (1991). What a wonderful life it would be: A re-analysis of some of the
work of Grossarth-Maticek. Psychological Inquiry, 2, 280–285.
Rachman, S. (2015). The evolution of behavior therapy and cognitive behavior therapy.
Behaviour Research and Therapy, 64, 1–8.
Rust, J. (1975). Cortical evoked potential, personality, and intelligence. Journal of
Comparative and Physiological Psychology, 89, 1220–1226.
Saville, P., & Blinkhorn, S. (1976). Undergraduate personality by factored scores. London:
NFER Publishing Company.
Saville, P., & Blinkhorn, S. (1981). Reliability, homogeneity and the construct validity of
Cattell's 16PF. Personality and Individual Differences, 2, 325–333.
Schneller, W. F., & Garske, J. P. (1976). Effects of extroversion and neuroticism on learning
and memory: A test of Eysenck's theory of individual differences in arousal. Journal of
Psychology, 94, 27–32.
Schoenthaler, S. J., Amos, S. P., Eysenck, H. J., Peritz, E., & Yudkin, J. (1991). Controlled trial
of vitamin-mineral supplementation: Effects on intelligence and performance.
Personality and Individual Differences, 12, 351–362.
Schürer-Necker, E. (1990). Arousal and paired-associate learning: Evidence refuting the
action decrements theory of Walker and Tarte (1963). Pavlovian Journal of
Biological Science, 25, 195–200.
Servaas, M. N., Geerligs, L., Renken, R. J., Marsman, J. -B. C., Ormel, J., Riese, H., & Aleman, A.
(2015). Connectomics and neuroticism: An altered functional network organization.
Neuropsychopharmacology, 40, 296–304.
Shagass, C., Roemer, R. A., Straumanis, J. J., & Josiassen, R. C. (1981). Intelligence as a factor
in evoked-potential studies in psychopathology. 1. Comparison of low and high IQ
subjects. Biological Psychiatry, 11, 1007–1029.
Shields, J. (1954). Personality differences and neurotic traits in normal twin
schoolchildren: A study in psychiatric genetics. The Eugenics Review, 45,
213–246.
Siebel, W. A., Winkler, T., Seitz-Bernhard, B., & Noosomatik, I. (1990). Theoretische
Grundlegung [theoretical foundations]. Langwedel: Glaser & Wohlschlegel.
Soto, C. J. (2015). Is happiness good for your personality? Concurrent and
prospective relations of the Big Five with subjective well-being. Journal of
Personality, 83, 45–55.
Spinhoven, P., Penelo, E., de Rooij, M., Penninx, B. W., & Ormel, J. (2014). Reciprocal
effects of stable and temporary components of neuroticism and affective disor-
ders: Results of a longitudinal cohort study. Psychological Medicine, 44,
337–348.
Steel, P., Schmidt, J., & Shultz, J. (2008). Refining the relationship between personality and
subjective well-being. Psychological Bulletin, 134, 138–161.
Sylvester, C. M., Corbetta, M., Raichle, M. E., Rodebaugh, T. L., Schlaggar, B. L., Sheline, Y. I.,
et al. (2012). Functional network dysfunction in anxiety and anxiety disorders. Trends
in Neurosciences, 35, 527–535.
Tambs, K., Czajkowsky, N., Roysamb, E., Neale, M. C., Reichborn-Kjennerud, T.,
Aggen, S. H., et al. (2009). Structure of genetic and environmental risk factors
for dimensional representations of DSM-IV anxiety disorders. British Journal
of Psychiatry, 195, 301–307.
Tesser, A., & Shaffer, D. (1990). Attitudes and attitude change. Annual Review of
Psychology, 41, 479–523 (Palo Alto, CA: Annual Reviews).
Turkheimer, E., Haley, A., Waldron, M., D'Onofrio, B., & Gottesman, I. I. (2003). Socioeco-
nomic status modifies heritability of Q in young children. Psychological Science, 14,
623–628.
Uher, J. (2013). Personality psychology: Lexical approaches, assessment methods, and
trait concepts reveal only half the story — Why it is time for a paradigm shift?
Integrative Psychological and Behavioral Science, 47, 1–55.
 M.W. Eysenck / Personality and Individual Differences 103 (2016) 209–219 219

Vetterli, C. F., & Furedy, J. J. (1985). Evoked potential correlates of intelligence: Some prob-
lems with Hendrickson's string measure of evoked potential complexity and error
theory of intelligence. International Journal of Psychophysiology, 3, 1–3.
Vogel, F., Kruger, J., Schalt, E., Schnobel, R., & Hassling, L. (1987). No consistent relation-
ships between oscillations and latencies or visually and auditorily evoked EEG poten-
tials and measures of mental performance. Human Neurobiology, 6, 173–182.
Vukasović, T., & Bratko, D. (2015). Heritability of personality: A meta-analysis of behavior
genetic studies. Psychological Bulletin, 141, 769–785.
Watson, D., & Tellegen, A. (1985). Toward a consensual structure of mood. PsychoIogical
Bulletin, 98, 219–235.
Wood, D. (2015). Testing the lexical hypothesis: Are socially important traits more dense-
ly reflected in the English lexicon? Journal of Personality and Social Psychology, 108,
317–335.
Young, G. B., & Pigott, S. E. (1999). Neurobiological basis of consciousness. Archives of
Neurology, 56, 153–157.