Trait Autism is a Better Predictor of Empathy than Alexithymia

Abstract

It has been proposed that atypical empathy in autism spectrum disorder (ASD) is due to co-occurring alexithymia. However, difficulties measuring empathy and statistical issues in previous research raise questions about the role of alexithymia in empathic processing in ASD. Addressing these issues, we compared the associations of trait alexithymia and autism with empathy in large samples from the general population. Multiple regression analyses showed that both trait autism and alexithymia were uniquely associated with atypical empathy, but dominance analysis found that trait autism, compared to alexithymia, was a more important predictor of atypical cognitive, affective, and overall empathy. Together, these findings indicate that atypical empathy in ASD is not simply due to co-occurring alexithymia.

Full text

Vol:.(1234567890)
Journal of Autism and Developmental Disorders (2019) 49:3956–3964
https://doi.org/10.1007/s10803-019-04080-3
1 3
S.I. : EMPATHY INAUTISM
Trait Autism isaBetter Predictor ofEmpathy thanAlexithymia
PunitShah1 · LucyA.Livingston2 · MitchellJ.Callan1· LoisPlayer1
Published online: 7 June 2019
© The Author(s) 2019
Abstract
It has been proposed that atypical empathy in autism spectrum disorder (ASD) is due to co-occurring alexithymia. How-
ever, difficulties measuring empathy and statistical issues in previous research raise questions about the role of alexithymia
in empathic processing in ASD. Addressing these issues, we compared the associations of trait alexithymia and autism
with empathy in large samples from the general population. Multiple regression analyses showed that both trait autism and
alexithymia were uniquely associated with atypical empathy, but dominance analysis found that trait autism, compared to
alexithymia, was a more important predictor of atypical cognitive, affective, and overall empathy. Together, these findings
indicate that atypical empathy in ASD is not simply due to co-occurring alexithymia.
Keywords Autism· Empathy· Alexithymia· Cognitive empathy· Affective empathy
Considerable research has been directed towards studying
empathy in autism spectrum disorder (ASD). Early research
indicated that empathy was impaired in ASD (e.g., Baron-
Cohen and Wheelwright 2004), but inconsistencies in con-
ceptualizing and measuring empathy led to confusion in the
literature (see Rogers etal. 2007). Addressing this issue has
involved two particularly fruitful lines of research that we
aimed to build on in the present study. First, there has been
a move towards studying different components of empathy
in ASD. Understanding or knowing what another individual
is feeling (cognitive empathy) has been dissociated from
feeling what others are feeling (affective empathy) in neuro-
science and psychological research (e.g., Yang etal. 2018).
Such research generally indicates that cognitive, not affec-
tive, empathy is lower in ASD (e.g., Rueda etal. 2015).
Second, recent work has highlighted the role of trait alex-
ithymia (difficulties in identifying and describing one’s own
emotions) in ASD. This has taken elevated rates of alexithy-
mia in ASD to argue that impaired emotional processing and
empathy, where observed in ASD, is due to co-occurring
alexithymia (e.g., Bird etal. 2010). Together, these lines of
research have challenged the view that empathy is univer-
sally impaired in ASD. There are, however, several concerns
with this research.
The first issue is that widely-used measures of empathy
in ASD research—the Interpersonal Reactivity Index (IRI;
Davis 1983) and the Empathy Quotient (Baron-Cohen and
Wheelwright 2004)—were not designed to dissociate cog-
nitive from affective empathy, and there are longstanding
concerns about the IRI’s validity (see Murphy etal. 2018).
To address this issue, Reniers etal. (2011) developed the
Questionnaire of Cognitive and Affective Empathy (QCAE)
by drawing on several empathy measures to create a more
robust measure of cognitive, affective, and overall empathy.
The QCAE has now been validated in several clinical and
non-clinical samples (e.g., Di Girolamo etal. 2017), how-
ever, apart from one recent study (see below), it has not been
used in research pertaining to ASD.
The second issue is that few studies have used appropri-
ate analyses to investigate cognitive and affective empathy
in ASD, which has contributed to the inconsistency and
mixed findings in previous research (see Yang etal. 2018).
Researchers have typically examined the link between one
Electronic supplementary material The online version of this
article (https ://doi.org/10.1007/s1080 3-019-04080 -3) contains
supplementary material, which is available to authorized users.
* Punit Shah
p.shah@bath.ac.uk
* Lucy A. Livingston
lucy.livingston@kcl.ac.uk
1 Department ofPsychology, University ofBath,
BathBA27AY, UK
2 Social, Genetic andDevelopmental Psychiatry Centre,
Institute ofPsychiatry, Psychology andNeuroscience, King’s
College London, LondonSE58AF, UK
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3957Journal of Autism and Developmental Disorders (2019) 49:3956–3964
1 3
component of empathy and ASD without accounting for the
other. This is problematic because affective and cognitive
empathy, though conceptually distinguishable constructs,
are statistically correlated and co-activated in social situa-
tions (Preckel etal. 2018). It is possible that an individual
may understand what another person is feeling (i.e., have
intact cognitive empathy), but only after controlling for their
difficulties in feeling what that person is feeling. Equally,
an individual may feel what other people are feeling (i.e.,
have intact affective empathy), yet experience difficulties
in understanding or identifying those feelings. Therefore,
investigating one component of empathy requires statistical
consideration of the other during analysis‚ for a more precise
understanding of (a)typical empathy.
Finally, there are methodological issues in research on
the co-occurrence of ASD and alexithymia. Such research
typically compares very small samples of people with and
without ASD, which often lacks sufficient statistical power
to test the unique associations of trait autism and alex-
ithymia (see also, Nicholson etal. 2018). Crucially, this
research also involves matching groups with and without
ASD on alexithymia, with studies reporting no associa-
tion between ASD and atypical empathy after controlling
for alexithymia (e.g., Bird etal. 2010). However, because
the prevalence of alexithymia is much lower in typically
developing compared to autistic populations (5% vs 50%,
respectively; Kinnaird etal. 2019), matching groups for
alexithymia is potentially problematic. Matching groups
in this way necessitates biased sampling, therefore neither
group is representative of autistic or typically developing
populations (see also, Lassalle etal. 2019), resulting in
inappropriate statistical group comparisons and potentially
inaccurate population-level inferences. For example, Oak-
ley etal. (2016), when investigating a small sample of 19
autistic and 23 non-autistic adults (matched for alexithy-
mia), found that affective impairments were solely associ-
ated with alexithymia, whereas theory of mind (analogous
to cognitive empathy) was only impaired in ASD. How-
ever, Oakley etal. did not statistically control for theory
of mind when examining the associations between ASD,
alexithymia and affective processing, nor control for affec-
tive processing when measuring the links between ASD,
alexithymia and theory of mind. Given the small sample
size, it is also unclear whether ASD has no association
with affective impairments over and above alexithymia, or
whether previous findings have been Type II errors due to
suboptimal statistical analysis. Furthermore, it is question-
able whether their statistical inferences—i.e., generalizing
from samples to the population—were appropriate due to
the biased sampling required for alexithymia-matched
groups. Notwithstanding these concerns, previous research
leads to testable predictions that trait autism (hereafter
‘autism’), not alexithymia, should be associated with
low cognitive empathy, whereas alexithymia, not autism,
should be related to low affective empathy. These hypoth-
eses are explored in the present research.
To our knowledge, only one study has explored the rela-
tionship between ASD and alexithymia, and their relative
associations with cognitive and affective empathy. Mul
etal. (2018), in a small study of adults with (n = 26) and
without (n = 26) ASD, found that alexithymia partially
mediated the links between ASD and both low cognitive
and affective empathy. Although this supports the idea that
alexithymia may partly contribute to atypical empathy in
ASD, it does not support claims (i.e., Bird and Cook 2013)
that empathy impairments, where observed in ASD, are
entirely due to alexithymia. Moreover, Mul etal. noted
the small sample in their study as a limitation, which also
resulted in limited variance in alexithymia in the control
compared to the ASD group. In addition, the relationships
of ASD and alexithymia, separately with cognitive and
affective empathy, were not examined whilst accounting
for the other component of empathy in their mediation
analyses. Mul etal.’s study was also not designed to com-
pare the statistical importance of ASD and alexithymia as
predictors of atypical empathy.
In view of the limitations of previous research, we suggest
that the extent to which ASD and alexithymia are related
to empathy requires further investigation. Addressing con-
cerns with previous work, we designed studies using the
QCAE and measures of trait autism and alexithymia in two
large community samples drawn from the general popula-
tion. Despite potential limitations with this approach (see
“Discussion”), this avoided inappropriate comparisons of
small and biased samples of adults with and without ASD
(matched for alexithymia), and poorly powered statistical
analyses, commonly found in previous research. Instead, we
aimed to conduct the most well-powered statistical examina-
tion of the interrelationships between autism, alexithymia,
and different components of empathy to date. Specifically,
we compared the associations of autism and alexithymia
with overall empathy, and each component of empathy
whilst controlling for the other component. Critically, we
compared the statistical importance of autism and alexithy-
mia as predictors of empathy by using dominance analysis
for the first time in this field of research.
Methods
Participants, Measures, andProcedure
Participants formed a community sample drawn from
online sources of 306 adults (45% female), aged between
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3958 Journal of Autism and Developmental Disorders (2019) 49:3956–3964
1 3
18 and 85years (M = 34.0years, SD = 11.9years). A
power analysis (Faul etal. 2007) revealed that we had 95%
power to detect “small-to-medium” unique associations
in our regression analyses (f2 = 0.03, α = 0.05, 2-tailed).1
Participants completed self-report measures of trait autism
(28-item Short Autism-Spectrum Quotient, AQS; Hoekstra
etal. 2011), alexithymia (20-item Toronto Alexithymia
Scale, TAS-20; Bagby etal. 1994), and empathy (31-
item Questionnaire of Cognitive and Affective Empathy,
QCAE; Reniers etal. 2011). The AQS, measuring (dis)
agreement with statements on autism-like symptoms on a
4-point Likert scale, is a validated and widely used quan-
titative measure of autistic traits; scores range between
28 (few autistic traits) and 112 (many autistic traits). It
has been shown to measure the same latent construct in
adults with and without a clinical diagnosis of ASD (Mur-
ray etal. 2014) and in males and females (Grove etal.
2017). The TAS-20, measuring (dis)agreement with state-
ments about difficulties identifying and describing one’s
own emotions on a 5-point Likert scale, quantified alex-
ithymia; scores range between 20 (low alexithymia) and
100 (high alexithymia). The TAS-20 has been used exten-
sively in samples with and without ASD, notably to test
the competing influences of autism and alexithymia on
psychological variables (e.g., Shah etal. 2016a, b), as in
the current study. The QCAE measured (dis)agreement
with statements about understanding others’ feelings and
feeling others’ feelings on a 4-point Likert scale; scores for
overall empathy range between 31 (low empathy) and 124
(high empathy), cognitive empathy between 19 and 76,
and affective empathy between 12 and 48. The QCAE has
also been validated and used widely to measure (a)typi-
cal levels of empathy (see Lockwood 2016). All measures
had good reliability in the current study (AQS: α = .82,
TAS-20: α = .90, QCAE: α = .91). The questionnaires were
presented in a randomized order, followed by questions
about age and sex.
Results
A wide range of autism and alexithymia scores were present
in the sample (Table1), confirming adequate variance in line
with previous research (e.g., Farmer etal. 2017; Shah etal.
2016a). All variables were moderately correlated (Table1).
Notably, both autism and alexithymia were negatively cor-
related with cognitive, affective, and overall empathy. There
were positive correlations between autism and alexithymia,
and between cognitive and affective empathy. Male partici-
pants also reported lower levels of empathy than females
(Online Resource—Table1). Multiple regression analyses
measured the unique associations of autism and alexithymia
with (i) overall empathy, (ii) cognitive empathy, and (iii)
affective empathy. Sex was included in all regressions, given
the sex differences in empathy. Results showed that autism
and alexithymia were both significant predictors of low (i)
overall empathy, (ii) cognitive empathy after accounting for
affective empathy, whereas (iii) alexithymia, not autism, was
associated with higher affective empathy after accounting for
cognitive empathy (see Table2).
Multicollinearity was not a concern as the variables
were moderately correlated (Table1), in line with research
finding that autism, alexithymia, and empathy are different
constructs. Equally, however, given the finding that autism
and alexithymia were both significantly associated with
atypical empathy, it was not appropriate to determine the
relative importance of each predictor by simply compar-
ing the size of their beta coefficients (see Budescu 1993).
To overcome this problem, we employed dominance analy-
sis, which involves computing each predictor’s incremental
validity (or semi-partial correlation squared, sr2) across all
possible subset regression models involving that predictor.
Table 1 Means and correlations
Trait autism was measured using the 28-item Short Autism-Spectrum Quotient (AQS; Hoekstra et al.
2011), alexithymia using the 20-item Toronto Alexithymia Scale (TAS-20; Bagby etal. 1994), and empa-
thy using cognitive and affective subscales and overall scores of the Questionnaire of Cognitive and Affec-
tive Empathy (QCAE; Reniers etal. 2011)
**p < .01
***p < .001
Measure M (SD) 1 2 3 4
1. Trait autism (AQS) 65.08 (10.27) –
2. Trait alexithymia (TAS-20) 47.47 (14.03) .46*** –
3. Cognitive empathy (QCAE cognitive subscale) 57.94 (8.54) − .50*** − .44***
4. Affective empathy (QCAE affective subscale) 33.68 (5.70) − .26*** − .19** .51***
5. Overall empathy (QCAE overall score) 91.62 (12.45) − .46*** − .39*** .92*** .81***
1 Fourteen additional participants were recruited but excluded from
the final sample. Four participants failed to complete the study and
3 participants completed the study twice. Seven participants were
excluded as they were multivariate outliers with residuals more than
3SDs from the mean in the regression analyses.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3959Journal of Autism and Developmental Disorders (2019) 49:3956–3964
1 3
These validities are then used to establish the relative impor-
tance of each predictor to the criterion, yielding General
Dominance Weights (GDW), which represent the average sr2
across submodels for a given predictor. The GDW sum to the
overall model R2 for a given criterion and are used to rank-
order each predictor’s relative importance to the criterion.
In other words, dominance analysis permits the ranking of
statistical importance, which is not possible through con-
ventional regression (see Nimon and Oswald 2013). Includ-
ing participant sex in the models, we performed dominance
analyses using the yhat package in R (Nimon etal. 2013).
Results showed that autism dominated alexithymia as a pre-
dictor of cognitive, affective, and overall empathy (Table2).
Further, bootstrapping (1000 resamples) estimated repro-
ducibility rates (RR) of how likely the dominance relation-
ship would be observed in the population from how often
it occurred in the bootstrapped samples. This showed that
autism dominated alexithymia for overall (RR = 90.5%),
cognitive (RR = 75.1%), and affective (RR = 79.6%) empa-
thy. RRs ≥ 70% indicate high confidence that the dominance
relationship observed in the sample would exist in the popu-
lation (Azen 2013).
Exploratory Analyses andReplication Study
Exploratory analyses2 showed that males reported signifi-
cantly more autistic and alexithymic traits than female par-
ticipants (Online Resource—Table1). We also explored
whether the associations between autism and empathy, and
alexithymia and empathy, were moderated by sex by includ-
ing sex × autism and sex × alexithymia interaction terms in
the original regression analyses. These interaction terms
were not statistically significant predictors of overall and
cognitive empathy scores (Online Resource—Tables2, 3).
However, there was a statistically significant sex × alexithy-
mia interaction for affective empathy (Online Resource—
Table4). Simple slopes analysis revealed that an association
between alexithymia and higher affective empathy was sig-
nificant in male (β = 0.24, t = 3.40, p < .001) but not female
(β = − 0.02, t = − 0.27, p = .79) participants.
Following recommendations to improve the replica-
bility of clinical psychological science (Tackett etal.
2017), we conducted a replication study in another large
sample that completed the same procedure (see Online
Table 2 Regression and dominance analyses for overall, cognitive
and affective empathy
Examination of VIF values across the regression analyses indicated
that multicollinearity was not a concern (all < 10), and the residuals
were normally distributed. Durbin–Watson statistics were inspected
and found to be ~ 2 across the regression analyses, suggesting that
errors were uncorrelated and thus independent. Together, the data
were suitable for multiple linear regression analysis
Β standardized regression coefficient, t Student’s t-statistic, pp value,
sr2 semi-partial correlation squared, GDW General Dominance
Weight (higher GDW values indicate a more important predictor)
Predictor β t p sr2GDW
(i) Overall empathy—F(3, 302) = 39.22, R2 = 0.28, p < .001
Sex (1 = male,
0 = female)
− .18 − 3.64 < .001 0.042 0.050
Autism − .34 − 6.26 < .001 0.115 0.147
Alexithymia − .19 − 3.33 .001 0.035 0.084
(ii) Cognitive empathy—F(4, 301) = 65.99, R2 = 0.47, p < .001
Affective empathy .46 9.66 < .001 0.236 0.199
Sex .17 3.60 < .001 0.041 0.014
Autism − .28 − 5.82 < .001 0.101 0.143
Alexithymia − .26 − 5.49 < .001 0.091 0.053
(iii) Affective empathy—F(4, 301) = 50.08, R2 = 0.40, p < .001
Cognitive empathy .52 9.66 < .001 0.236 0.212
Sex − .38 − 8.34 < .001 0.187 0.148
Autism − .004 − 0.08 .94 0.00003 0.025
Alexithymia .13 2.48 .014 0.020 0.014
Table 3 Replication study—regression analyses for overall, cognitive
and affective empathy
Examination of VIF values across the regression analyses indicated
that multicollinearity was not a concern (all < 10), and the residuals
were normally distributed. Durbin–Watson statistics were inspected
and found to be ~ 2 across the regression analyses, suggesting that
errors were uncorrelated and thus independent. Together, the data
were suitable for multiple linear regression analysis
Β Standardized regression coefficient, t Student’s t-statistic, p p value
Predictor β t p
(i) Overall empathy—F(5, 348) = 30.28, R2 = 0.30, p < .001
Sex (1 = male, 0 = female) − .34 − 7.48 < .001
Sex × autism .02 0.30 .77
Sex × alexithymia .10 1.95 .052
Autism − .25 − 4.72 < .001
Alexithymia − .18 − 3.41 .001
(ii) Cognitive empathy–F(6, 347) = 40.61, R2 = 0.41, p < .001
Affective empathy .31 6.82 < .001
Sex − 1.62 − 1.62 .11
Sex × autism .03 0.56 .57
Sex × alexithymia .01 0.10 .92
Autism − .34 − 6.96 < .001
Alexithymia − .23 − 4.74 < .001
(iii) Affective empathy—F(6, 347) = 22.81, R2 = 0.28, p < .001
Cognitive empathy .38 6.82 < .001
Sex − .33 − 6.99 < .001
Sex × autism − .02 − 0.28 .078
Sex × alexithymia .12 2.29 .023
Autism .12 2.03 .044
Alexithymia .07 1.19 .24
2 We thank an anonymous reviewer for suggesting these analyses.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3960 Journal of Autism and Developmental Disorders (2019) 49:3956–3964
1 3
Resource—Replication Study). Data were submitted to
regression analyses (see Table3) to measure the unique
associations of autism and alexithymia with (i) overall, (ii)
cognitive, and (iii) affective empathy. Sex, sex × autism,
and sex × alexithymia were included in all models. Repli-
cating the original findings, autism and alexithymia were
(i) unique and significant predictors of low overall empathy,
and (ii) unique and significant predictors of low cognitive
empathy whilst accounting for affective empathy. In contrast
to the original result, (iii) autism, but not alexithymia, was
associated with higher affective empathy whilst accounting
for cognitive empathy. In line with the original study, the
sex × alexithymia interaction was statistically significant, and
simple slopes analysis revealed that the association between
alexithymia and higher affective empathy was significant for
male (β = 0.20, t = 2.50, p = .013) but not female (β = − 0.05,
t = − 0.67, p = .51) participants. Replicating the original
dominance analysis, General Dominance Weights (GDW)
and Reproducibility Rates (RR) indicated that autism
dominated alexithymia as a predictor of overall (GDW
autism = 0.10, alexithymia = 0.07; RR = 75.9%), cognitive
(GDW autism = 0.16, alexithymia = 0.11; RR = 85.3%),
and affective (GDW autism = 0.007, alexithymia = 0.005;
RR = 79.4%) empathy. Together, the original pattern of
results was replicated; that is, autism was a better predictor
of empathy than alexithymia.
Discussion
The association between autism and alexithymia, the link
between cognitive and affective empathy, and sex dif-
ferences in empathy are in line with previous research.
These results support theories that participant sex (Baron-
Cohen and Wheelwright 2004; Greenberg etal. 2018) and
alexithymia (Bird and Cook 2013) are broadly relevant
to understanding empathy in ASD. Our results also sup-
port claims that, although partly dissociable, cognitive and
affective empathy are overlapping constructs (Preckel etal.
2018). We therefore examined overall empathy (combining
cognitive and affective scores) as the starting point in our
multivariate analyses. Across both studies, we found that,
although alexithymia partly contributes to low empathy
(in line with Mul etal. 2018), autism is more predictive
of low overall empathy in the population. Critically, the
consistency of the dominance analyses across both studies
highlighted the greater statistical importance of autism
as a predictor of low overall empathy when compared
to alexithymia. Although dominance analysis only pro-
vides metrics for statistical importance, we suggest that
our results support claims that low empathy is a clinically
important feature of ASD. The present study is of course
not sufficient to substantiate this proposal, however it does
provide fresh evidence in support of longstanding (e.g.,
Baron-Cohen and Wheelwright 2004) and recent (Russ
etal. 2018) proposals that measuring overall empathy
has utility in the diagnosis and management of ASD (see
also, Robinson and Elliott 2016). Building on this previ-
ous work, we propose that measuring and managing trait
autism, compared to alexithymia, is likely to be a more
efficacious approach to investigate and ameliorate empa-
thy-related difficulties in ASD.
Given evidence for the dissociation between cognitive
and affective empathy (e.g., Reniers etal. 2011), our analy-
ses of these subcomponents provide a more precise under-
standing of empathy in relation to autism. Across both stud-
ies, autism and alexithymia were both uniquely associated
with difficulties in knowing what people are feeling (i.e.,
low cognitive empathy), whilst controlling for difficulties
in feeling what others are feeling (i.e., low affective empa-
thy). However, we note that, although alexithymia partly
contributed to atypical empathy, autism was more predictive
of lower cognitive empathy. Accordingly, autism was, across
all dominance analyses, a far more important predictor of
low cognitive empathy than alexithymia. This is consistent
with findings that ASD is characterized by poor cognitive
empathy (e.g., Rueda etal. 2015) and, importantly, the pre-
sent study is the first to demonstrate the robustness of the
association between autism and low cognitive empathy even
after accounting for alexithymia and affective empathy.
Our results also fit with previous evidence for a link
between alexithymia and impaired cognitive empathy (e.g.,
Di Girolamo etal. 2017; Moriguchi etal. 2006) and reports
that emotional awareness is associated with cognitive empa-
thy or theory of mind (Lane etal. 2015). However, unlike
previous research, this is the first study to detect a relation-
ship between alexithymia and low cognitive empathy even
after controlling for autism and affective empathy. These
results are therefore not consistent with recent evidence that
alexithymia is unrelated to theory of mind (synonymous
with cognitive empathy; Rueda etal. 2015) after accounting
for autism (i.e., Oakley etal. 2016). This may be due to the
fact that behavioral, instead of questionnaire, measures were
used in recent research, and/or because previous studies did
not control for affective empathy, and/or have the statisti-
cal power to detect this pattern of results. It is also debated
whether ‘cognitive empathy’ and ‘theory of mind’ are syn-
onymous because it remains unclear whether experimental
tasks of these abilities are measuring the same construct
(Happé etal. 2017; Warrier and Baron-Cohen 2018). We
therefore suggest that, following refinement of the theoreti-
cal overlap, terminology, and measures of these social cog-
nitive processes (Happé etal. 2017), it may be necessary
to conduct a follow-up of the present study that includes
refined measures of cognitive empathy and theory of mind
(see Livingston etal. 2019).
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3961Journal of Autism and Developmental Disorders (2019) 49:3956–3964
1 3
Our findings on affective empathy were novel but
inconsistent. The first regression analysis unexpectedly
showed that alexithymia, but not autism, was associated
with high affective empathy whilst controlling for cogni-
tive empathy. These results fit with research showing that
alexithymia may be associated with high levels of affec-
tive empathy (Guttman and Laporte 2002). However, the
second regression analysis failed to replicate this finding,
instead showing that autism, not alexithymia, was asso-
ciated with high affective empathy. Moreover, although
autism was not a significant predictor of affective empathy
in the first regression model including other predictor vari-
ables (sex, cognitive empathy, alexithymia), it was still a
more important predictor of affective empathy than alex-
ithymia as dominance analysis averages across all possible
subset regression models involving autism. The second
regression was, however, consistent with the dominance
analysis; autism emerged as a statistically significant and
more important predictor of affective empathy. Together,
dominance analysis, across both studies, demonstrated that
autism was a better predictor of higher affective empathy
compared to alexithymia. This result is consistent with
research that autism, but not alexithymia, may be associ-
ated with a hypersensitivity to others’ feelings (Fan etal.
2014; Smith 2009). Finally, additional exploratory analy-
ses revealed that the association between alexithymia and
higher affective empathy, evident in both our studies, was
only found in male, and not female, participants. This
is somewhat consistent with research on the association
between alexithymia and empathy (after accounting for
autism), which has typically been found in all male sam-
ples (e.g., Bird etal. 2010).
Despite some of these interesting and potentially impor-
tant findings regarding affective empathy, the inconsisten-
cies within our study and in previous research indicates
that it is most prudent to conclude that neither autism,
nor alexithymia, are robust predictors of atypical affec-
tive empathy, especially when compared to their stronger
associations with participant sex and cognitive empathy.
This is supported by emerging evidence that, in clinically
diagnosed people with ASD, neither autism, nor alexithy-
mia, are associated with affective empathy (Ziermans etal.
2018). Our data are therefore not consistent with claims
that atypical affective empathy, where observed in ASD,
is due to alexithymia (Bird etal. 2010). Rather, a growing
body of research indicates that alexithymia-based expla-
nations for the link between autism and atypical empa-
thy (e.g., Bird and Cook 2013) may have been overstated
in previous research. Overall, therefore, we suggest that
the interrelationships between sex, autism, alexithymia,
and affective empathy are more complex than previously
reported and will require further investigation in future
research (see also, Lassalle etal. 2019).
Strengths, Limitations, andFuture Directions
The current study indicates that greater consideration of
participant sex will be required in future research on the
competing influences of autism and alexithymia on atypical
empathy (see also, Greenberg etal. 2018). It seems possible
that if previous studies on alexithymia-based explanations
of empathic difficulties in ASD (e.g., Bird etal. 2010) were
conducted in samples containing females, a different pattern
of results would emerge. More broadly, there is growing
evidence for sex differences in social-emotional processing
in people with ASD (see Lai etal. 2015, for overview; Liv-
ingston and Happé 2017). Therefore, studies with large and
representative samples of males and females will, in future,
be required to investigate sex-specific empathy profiles (if
any) in ASD.
Further research will also be necessary to overcome sev-
eral other limitations of the current study. The present study
was conducted in large samples drawn from the general pop-
ulation to enable high-powered analyses that are currently
not possible to perform with equal rigor in clinical sam-
ples due to practical considerations. Well-powered statisti-
cal analyses allowed us to find new evidence in support of
existing clinical research, and detect novel associations that
will inform the design of future research in clinically diag-
nosed people with ASD. However, while the study of autistic
traits is widely used to inform understanding of autism (see
Ruzich etal. 2015), there are many limitations and ongoing
debates about the appropriateness of measuring sub-clinical
autistic traits as a proxy for understanding clinically diag-
nosed ASD. For example, it remains debated whether ASD
and population-level autism traits lie on a quantitative con-
tinuum or are qualitatively distinct (e.g., Constantino and
Charman 2016; Frazier etal. 2010; Volkmar and McPartland
2016). It will therefore be critically important to re-examine
alexithymia’s role in empathy in clinically diagnosed people
with ASD, and it is hoped that the current study provides the
impetus for such future research.
We used a well-validated empathy questionnaire to col-
lect a large dataset to avoid problems with experimental
tasks of empathy (see Mackes etal. 2018, for recent discus-
sion), such as their lack of ecological validity and narrow
conceptualizations of empathy (e.g., empathy for pain; see
Lockwood 2016). However, it will be important to determine
whether our results can be reproduced using new video-
based experimental measures of empathy that are currently
undergoing development (e.g., Mackes etal. 2018). Mov-
ing forward, we propose that studies combining appropriate
experimental and questionnaire measures, in large samples,
will be required to further elucidate the role of alexithymia
(if any) in atypical empathy in ASD. Finally, our research,
following previous studies in this field, was cross-sectional.
Longitudinal research will of course be required to examine
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3962 Journal of Autism and Developmental Disorders (2019) 49:3956–3964
1 3
whether atypical emotional processing and empathy in ASD
is a cause or consequence of alexithymia or other co-occur-
ring traits (Livingston and Livingston 2016; Poquérusse
etal. 2018).
Conclusions
To conclude, the present study indicates that trait autism,
compared to alexithymia, is a better predictor of cognitive,
affective, and overall empathy. We suggest that the role of
alexithymia in empathic processing in ASD requires further
investigation before considering any clinical implications
arising from such research. To this end, we call for addi-
tional research using (i) appropriately large and representa-
tive samples of male and female individuals with ASD, (ii)
new and improved experimental and questionnaire measures
of empathy, and (iii) well-powered multivariate analyses as
used in the present study, for a re-examination of the role of
alexithymia in empathic processing in ASD.
Acknowledgments LAL acknowledges support from the Medical
Research Council.
Author Contributions LAL and PS contributed equally to this work.
LAL and PS conceived the study. LP, LAL, and PS participated in
data collection, PS, LP, and MC analyzed the data, PS conducted the
replication study, LAL and PS drafted the manuscript, and all authors
read and approved the final manuscript.
Funding This research was supported by the Medical Research
Council.
Compliance with Ethical Standards
Conflict of interest All authors declare no conflict of interest.
Ethical Approval All procedures were in accordance with the ethical
standards of the local ethics committee (17-214), guidelines from the
British Psychological Society, and with the 1964 Helsinki declaration
and its later amendments.
Informed Consent Informed consent was obtained from all individual
participants included in the study.
Open Access This article is distributed under the terms of the Crea-
tive Commons Attribution 4.0 International License (http://creat iveco
mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribu-
tion, and reproduction in any medium, provided you give appropriate
credit to the original author(s) and the source, provide a link to the
Creative Commons license, and indicate if changes were made.
References
Azen, R. (2013). Using dominance analysis to estimate predictor
importance in multiple regression. In Y. Petscher, C. Schatsch-
neider, & D. L. Compton (Eds.), Applied quantitative analysis
in education and social sciences (pp. 34–64). New York, NY:
Routledge.
Bagby, R. M., Parker, J. D., & Taylor, G. J. (1994). The twenty-item
Toronto Alexithymia Scale—I. Item selection and cross-validation
of the factor structure. Journal of Psychosomatic Research, 38(1),
23–32. https ://doi.org/10.1016/0022-3999(94)90005 -1.
Baron-Cohen, S., & Wheelwright, S. (2004). The empathy quotient:
An investigation of adults with Asperger syndrome or high func-
tioning autism, and normal sex differences. Journal of Autism
and Developmental Disorders, 34(2), 163–175. https ://doi.
org/10.1023/B:JADD.00000 22607 .19833 .00.
Bird, G., & Cook, R. (2013). Mixed emotions: The contribution of
alexithymia to the emotional symptoms of autism. Translational
Psychiatry, 3(7), e285. https ://doi.org/10.1038/tp.2013.61.
Bird, G., Silani, G., Brindley, R., White, S., Frith, U., & Singer, T.
(2010). Empathic brain responses in insula are modulated by lev-
els of alexithymia but not autism. Brain, 133(5), 1515–1525. https
://doi.org/10.1093/brain /awq06 0.
Budescu, D. V. (1993). Dominance analysis: A new approach to the
problem of relative importance of predictors in multiple regres-
sion. Psychological Bulletin, 114(3), 542–551. https ://doi.
org/10.1037/0033-2909.114.3.542.
Constantino, J. N., & Charman, T. (2016). Diagnosis of autism spec-
trum disorder: Reconciling the syndrome, its diverse origins, and
variation in expression. Lancet Neurology, 15(3), 279–291. https
://doi.org/10.1016/S1474 -4422(15)00151 -9.
Davis, M. H. (1983). Measuring individual differences in empathy:
Evidence for a multidimensional approach. Journal of Per-
sonality and Social Psychology, 44(1), 113–126. https ://doi.
org/10.1037/0022-3514.44.1.113.
Di Girolamo, M., Giromini, L., Winters, C. L., Serie, C. M., & de
Ruiter, C. (2017). The questionnaire of cognitive and affective
empathy: A comparison between paper-and-pencil versus online
formats in Italian samples. Journal of Personality Assessment,
101(2), 159–170. https ://doi.org/10.1080/00223 891.2017.13897
45.
Fan, Y.-T., Chen, C., Chen, S.-C., Decety, J., & Cheng, Y. (2014).
Empathic arousal and social understanding in individuals with
autism: Evidence from fMRI and ERP measurements. Social Cog-
nitive and Affective Neuroscience, 9(8), 1203–1213. https ://doi.
org/10.1093/scan/nst10 1.
Farmer, G. D., Baron-Cohen, S., & Skylark, W. J. (2017). People
with autism spectrum conditions make more consistent deci-
sions. Psychological Science, 28(8), 1067–1076. https ://doi.
org/10.1177/09567 97617 69486 7.
Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G*Power
3: A flexible statistical power analysis program for the social,
behavioral, and biomedical sciences. Behavior Research Methods,
39(2), 175–191. https ://doi.org/10.3758/BF031 93146 .
Frazier, T. W., Youngstom, E. A., Sinclair, L., Kubu, C. S., Law, P.,
Rezai, A., etal. (2010). Autism spectrum disorders as a qualita-
tively distinct category from typical behavior in a large, clini-
cally ascertained sample. Assessment, 17(3), 308–320. https ://doi.
org/10.1177/10731 91109 35653 4.
Greenberg, D. M., Warrier, V., Allison, C., & Baron-Cohen, S. (2018).
Testing the empathizing-systemizing theory of sex differences
and the extreme male brain theory of autism in half a million
people. Proceedings of the National Academy of Sciences of the
United States of America, 115(48), 12152–12157. https ://doi.
org/10.1073/pnas.18110 32115 .
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3963Journal of Autism and Developmental Disorders (2019) 49:3956–3964
1 3
Grove, R., Hoekstra, R. A., Wierda, M., & Begeer, S. (2017). Explor-
ing sex differences in autistic traits: A factor analytic study
of adults with autism. Autism, 21(6), 760–768. https ://doi.
org/10.1177/13623 61316 66728 3.
Guttman, H., & Laporte, L. (2002). Alexithymia, empathy, and psycho-
logical symptoms in a family context. Comprehensive Psychiatry,
43(6), 448–455. https ://doi.org/10.1053/comp.2002.35905 .
Happé, F., Bird, G., & Cook, J. (2017). The structure of social cog-
nition: In(ter)dependence of sociocognitive processes. Annual
Review of Psychology, 68, 243–267. https ://doi.org/10.1146/annur
ev-psych -01041 6-04404 6.
Hoekstra, R. A., Vinkhuyzen, A. A., Wheelwright, S., Bartels, M.,
Boomsma, D. I., Baron-Cohen, S., etal. (2011). The construc-
tion and validation of an abridged version of the autism-spectrum
quotient (AQ-Short). Journal of Autism and Developmental Disor-
ders, 41(5), 589–596. https ://doi.org/10.1007/s1080 3-010-1073-0.
Nimon, K. Oswald, F., & Roberts, J. K. (2013). yhat: Interpreting
Regression Effects. R package version 2.0-0. https ://CRAN.R-
proje ct.org/packa ge=yhat.
Kinnaird, E., Stewart, C., & Tchanturia, K. (2019). Investigating
alexithymia in autism: A systematic review and meta-analysis.
European Psychiatry, 55, 80–89. https ://doi.org/10.1016/j.eurps
y.2018.09.004.
Lai, M.-C., Lombardo, M. V., Auyeung, B., Chakrabarti, B., & Baron-
Cohen, S. (2015). Sex/gender differences and autism: Setting
the scene for future research. Journal of the American Acad-
emy of Child and Adolescent Psychiatry, 54, 11–24. https ://doi.
org/10.1016/j.jaac.2014.10.003.
Lane, R. D., Hsu, C. H., Locke, D. E., Ritenbaugh, C., & Stonnington,
C. M. (2015). Role of theory of mind in emotional awareness
and alexithymia: Implications for conceptualization and meas-
urement. Consciousness and Cognition, 33, 398–405. https ://doi.
org/10.1016/j.conco g.2015.02.004.
Lassalle, A., Zürcher, N. R., Porro, C. A., Benuzzi, F., Hippolyte, L.,
Lemonnier, E., etal. (2019). Influence of anxiety and alexithy-
mia on brain activations associated with the perception of others’
pain in autism. Social Neuroscience, 14(3), 359–377.https ://doi.
org/10.1080/17470 919.2018.14683 58.
Livingston, L. A., Carr, B., & Shah, P. (2019). Recent advances and
new directions in measuring theory of mind in autistic adults.
Journal of Autism and Developmental Disorders, 49(4), 1738–
1744.https ://doi.org/10.1007/s1080 3-018-3823-3.
Livingston, L. A., & Happé, F. (2017). Conceptualising compensa-
tion in neurodevelopmental disorders: Reflections from autism
spectrum disorder. Neuroscience and Biobehavioral Reviews, 80,
729–742. https ://doi.org/10.1016/j.neubi orev.2017.06.005.
Livingston, L. A., & Livingston, L. M. (2016). Commentary: Alex-
ithymia, not autism, is associated with impaired interoception.
Frontiers in Psychology, 7, 1103. https ://doi.org/10.3389/fpsyg
.2016.01103 .
Lockwood, P. (2016). The anatomy of empathy: Vicarious experience
and disorders of social cognition. Behavioural Brain Research,
311, 255–266. https ://doi.org/10.1016/j.bbr.2016.05.048.
Mackes, N. K., Golm, D., O’Daly, O. G., Sarkar, S., Sonuga-Barke, E.
J., Fairchild, G., etal. (2018). Tracking emotions in the brain—
Revisiting the empathic accuracy task. NeuroImage, 178, 677–
686. https ://doi.org/10.1016/j.neuro image .2018.05.080.
Moriguchi, Y., Ohnishi, T., Lane, R. D., Maeda, M., Mori, T., Nemoto,
K., etal. (2006). Impaired self-awareness and theory of mind: An
fMRI study of mentalizing in alexithymia. NeuroImage, 32(3),
1472–1482. https ://doi.org/10.1016/j.neuro image .2006.04.186.
Mul, C. L., Stagg, S. D., Herbelin, B., & Aspell, J. E. (2018). The feel-
ing of me feeling for you: Interoception, alexithymia and empathy
in autism. Journal of Autism and Developmental Disorders, 48(9),
2953–2967. https ://doi.org/10.1007/s1080 3-018-3564-3.
Murphy, B. A., Costello, T. H., Watts, A. L., Cheong, Y. F., Berg, J.
M., & Lilienfeld, S. O. (2018). Strengths and weaknesses of two
empathy measures: A comparison of the measurement precision,
construct validity, and incremental validity of two multidimen-
sional indices. Assessment. https ://doi.org/10.1177/10731 91118
77763 6.
Murray, A. L., Booth, T., McKenzie, K., Kuenssberg, R., & O’Donnell,
M. (2014). Are autistic traits measured equivalently in individu-
als with and without an autism spectrum disorder? An invariance
analysis of the Autism Spectrum Quotient Short Form. Journal of
Autism and Developmental Disorders, 44(1), 55–64. https ://doi.
org/10.1007/s1080 3-013-1851-6.
Nicholson, T., Williams, D. M., Grainger, C., Christensen, J. F., Calvo-
Merino, B., & Gaigg, S. B. (2018). Interoceptive impairments do
not lie at the heart of autism or alexithymia. Journal of Abnor-
mal Psychology, 127(6), 612–622. https ://doi.org/10.1037/abn00
00370 .
Nimon, K. F., & Oswald, F. L. (2013). Understanding the results of
multiple linear regression: Beyond standardized regression coef-
ficients. Organizational Research Methods, 16(4), 650–674. https
://doi.org/10.1177/10944 28113 49392 9.
Oakley, B. F., Brewer, R., Bird, G., & Catmur, C. (2016). Theory of
mind is not theory of emotion: A cautionary note on the Read-
ing the Mind in the Eyes Test. Journal of Abnormal Psychology,
125(6), 818–823. https ://doi.org/10.1037/abn00 00182 .
Poquérusse, J., Pastore, L., Dellantonio, S., & Esposito, G. (2018).
Alexithymia and ASD: A complex relationship. Frontiers in Psy-
chology, 9, 1196. https ://doi.org/10.3389/fpsyg .2018.01196 .
Preckel, K., Kanske, P., & Singer, T. (2018). On the interaction of
social affect and cognition: Empathy, compassion and theory of
mind. Current Opinion in Behavioral Sciences, 19, 1–6. https ://
doi.org/10.1016/j.cobeh a.2017.07.010.
Reniers, R. L., Corcoran, R., Drake, R., Shryane, N. M., & Völlm, B.
A. (2011). The QCAE: A questionnaire of cognitive and affective
empathy. Journal of Personality Assessment, 93(1), 84–95. https
://doi.org/10.1080/00223 891.2010.52848 4.
Robinson, A., & Elliott, R. (2016). Brief report: An observational
measure of empathy for autism spectrum: A preliminary study of
the development and reliability of the client emotional processing
scale. Journal of Autism and Developmental Disorders, 46(6),
2240–2250. https ://doi.org/10.1007/s1080 3-016-2727-3.
Rogers, K., Dziobek, I., Hassenstab, J., Wolf, O. T., & Convit, A.
(2007). Who cares? Revisiting empathy in Asperger syndrome.
Journal of Autism and Developmental Disorders, 37(4), 709–715.
https ://doi.org/10.1007/s1080 3-006-0197-8.
Rueda, P., Fernández-Berrocal, P., & Baron-Cohen, S. (2015). Dis-
sociation between cognitive and affective empathy in youth
with Asperger Syndrome. European Journal of Developmen-
tal Psychology, 12(1), 85–98. https ://doi.org/10.1080/17405
629.2014.95022 1.
Russ, V., Kovshoff, H., Brown, T., Abbott, P., & Hadwin, J. A. (2018).
Exploring the role of empathy in understanding the social-cogni-
tive profile for individuals referred for Autism Spectrum Disorders
assessment in adulthood. Journal of Autism and Developmental
Disorders. https ://doi.org/10.1007/s1080 3-018-3693-8.
Ruzich, E., Allison, C., Smith, P., Watson, P., Auyeung, B., Ring,
H., etal. (2015). Measuring autistic traits in the general popu-
lation: A systematic review of the Autism-Spectrum Quo-
tient (AQ) in a nonclinical population sample of 6,900 typical
adult males and females. Molecular Autism, 6, 2. https ://doi.
org/10.1186/2040-2392-6-2.
Shah, P., Catmur, C., & Bird, G. (2016a). Emotional decision-making
in autism spectrum disorder: The roles of interoception and alex-
ithymia. Molecular Autism, 7, 43. https ://doi.org/10.1186/s1322
9-016-0104-x.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

3964 Journal of Autism and Developmental Disorders (2019) 49:3956–3964
1 3
Shah, P., Hall, R., Catmur, C., & Bird, G. (2016b). Alexithymia, not
autism, is associated with impaired interoception. Cortex, 81,
215–220. https ://doi.org/10.1016/j.corte x.2016.03.021.
Smith, A. (2009). The empathy imbalance hypothesis of autism: A
theoretical approach to cognitive and emotional empathy in autis-
tic development. The Psychological Record, 59(2), 273–294. https
://doi.org/10.1007/BF033 95663 .
Tackett, J. L., Lilienfeld, S. O., Patrick, C. J., Johnson, S. L., Krueger,
R. F., Miller, J. D., etal. (2017). It’s time to broaden the replica-
bility conversation: Thoughts for and from clinical psychological
science. Perspectives on Psychological Science, 12(5), 742–756.
https ://doi.org/10.1177/17456 91617 69004 2.
Volkmar, F. R., & McPartland, J. C. (2016). Moving beyond a categori-
cal diagnosis of autism. Lancet Neurology, 15(3), 237–238. https
://doi.org/10.1016/S1474 -4422(15)00299 -9.
Warrier, V., & Baron-Cohen, S. (2018). Genetic contribution to ‘theory
of mind’ in adolescence. Scientific Reports, 8(1), 3465. https ://doi.
org/10.1038/s4159 8-018-21737 -8.
Yang, C.-C., Khalifa, N., & Völlm, B. (2018). The effects of repeti-
tive transcranial magnetic stimulation on empathy: A systematic
review and meta-analysis. Psychological Medicine, 48(5), 737–
750. https ://doi.org/10.1017/S0033 29171 70023 2X.
Ziermans, T., de Bruijn, Y., Dijkhuis, R., Staal, W., & Swaab, H.
(2018). Impairments in cognitive empathy and alexithymia occur
independently of executive functioning in college students with
autism. Autism. https ://doi.org/10.1177/13623 61318 81771 6.
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com