Imaginary Companions in Children with Autism Spectrum Disorder
Journal of Autism and Developmental Disorders
Imaginary Companions in Children with Autism Spectrum Disorder
Paige E. Davis 0 1 2
Haley Simon 0 1 2
Elizabeth Meins 0 1 2
Diana L. Robins 0 1 2
0 AJ Drexel Autism Institute , 3020 Market St., Philadelphia, PA 19104 , USA
1 Psychology Department, University of York , Heslington, York YO10 5DD , UK
2 Psychology Department, Drexel University , 3141 Chestnut St., Philadelphia, PA 19104 , USA
3 Paige E. Davis
One of the deficits observed in autism spectrum disorder (ASD) is impaired imaginative play. One form of imaginative play common in many typically developing (TD) children is having an imaginary companion (IC). The occurrence of ICs has not been investigated extensively in children with ASD. We examined differences in parent report of IC between TD and ASD populations in 215 (111 with ASD) gender-matched children aged between 2 and 8 years. Findings indicate that significantly fewer children with ASD created ICs, although there were many between-group similarities in IC forms and functions. Results are discussed in terms of qualitative differences in play, social attributions, and how children with ASD conceptualize their ICs' minds.
Autism spectrum disorder; Imagination; Imaginary companions; Social attribution; Social development
Imaginary or pretend play is an activity that typically
developing (TD) children engage in frequently and spontaneously.
However, children diagnosed with an autism spectrum
disorder (ASD) show deficits in this behavior. Early in the ASD
Wing and Gould (1979)
proposed that children
with ASD had marked impairments in imagination, along
with socialization and communication deficits. The issues
with imagination that children with ASD have are so
prevalent that one of the components of the diagnostic criteria is
difficulty in sharing imaginative play
(APA 2013; Jarrold
2003; WHO 1992)
Psychology Department, University of Huddersfield,
EKG/09, Queensgate, Huddersfield HD1 3DH, UK
Past research on imaginative behavior in children
diagnosed with ASD shows that this population is profoundly
delayed in imaginative play
Rutherford et al. 2007; Wolfberg et al. 2012)
, specifically social
(Lewis and Boucher 1988; Ten Eycke and
. The generativity of pretend play is also slower
for children with ASD than TD or language matched peers
with learning disabilities (Jarrold et al. 1996). Pretend play
enables children to engage in and enact different social roles
(Howes 1988; Singer and Singer 1990)
, practice different
social interactions using an appropriate script for that
character or individual (Harris 2000), and develop an
understanding of how social rules operate by using skills such as
sharing, taking turns, or verbally interacting
. Thus, higher levels of social pretend play are
related to higher levels of peer oriented social competence
(Uren and Stagnitti 2009).
The scarcity of pretend play in children with ASD
therefore reduces opportunities for them to engage in imaginative
play as a social activity
(Wolfberg et al. 2012)
In order to explore imagination in children diagnosed
with ASD, some researchers have focused on differences
in children’s abilities to generate hypotheticals, or novel
ideas, using Karmiloff-Smith’s (1990) “draw an
impossible person” task
(Low et al. 2009; Scott and
BaronCohen 1996; Ten Eycke and Müller 2015)
. Children with
ASD showed deficits in drawing impossible (imaginary)
entities compared to both TD children and children with
(Scott and Baron-Cohen 1996)
accounting for differences in executive functioning and
generativity. In a similar paradigm, ASD and TD children
were told to imagine a person walking in a “magic” door
and coming out changed into a funny and strange looking
(Ten Eycke and Müller 2015)
. This script was used
to increase the children’s comprehension of the request.
They were then asked to draw the person and subsequently
rated on the imaginative content of their drawing by two
independent TD coders using proportional scoring criteria
Marsh et al. (1996
). Children with ASD
created significantly less imaginative drawings.
Interestingly, this imaginative deficit may be restricted to social
content; when children with ASD drew both an impossible
person and an impossible house
(Ten Eycke and Müller
, only the person drawings were significantly less
imaginative compared to TD controls. Ten Eycke and
Müller theorized impoverished social experiences might lead
to less social information processing, leading to the less
elaborate forms of mental representation much like the
early social attention negative feedback loop proposed by
Dawson et al. (2004
There are also other theories as to why children diagnosed
with ASD might have impairments in social imagination
posited that children with
ASD may be unable to use second order representations
or meta-representation, hence their difficulties in passing
false belief tasks, and this could carry over to the ability to
represent in social play situations. Deficits in play may be
restricted to play that requires meta-representation
, given that functional play may be intact in ASD.
One imaginative behavior that incorporates invisible and
impossible entities and requires meta-representational ability
is the creation of imaginary companions (ICs)
Singer 1990; Taylor 1999)
. IC creation is seen in up to 50%
of TD children
(Gleason and Hohmann 2006; Motoshima
et al. 2014)
. An IC according to
invisible character named and referred to in conversation
with other persons or played with directly for a period of
time, at least several months, having an air of reality for the
child, but no apparent objective basis” (p. 988). Svendsen
did not include personified objects (POs), toys that a child
plays with for extended periods of time, but more recent
definitions consider them to be the equivalent of having an
IC as long as they have a stable personality
There are several factors that relate to the probability of
IC creation in TD children. Youngsters between the ages of
3 and 6 years are most likely to engage with an IC
and Singer 1990)
. It is also more probable for first-born
children to play with these imaginary creatures than children
(Bouldin and Pratt 1999; Manosevitz et al.
. Finally, girls are more likely than boys to develop
(Pearson et al. 2001; Carlson and Taylor 2005; Gleason
and Hohmann 2006)
, although this may relate to cultural
factors and gender normativity (Taylor 1999).
Children who create ICs have been shown to have
advantages compared to children without these entities (NIC).
Some studies report that having an IC is related to superior
ToM and emotion understanding abilities
et al. 2016; Taylor and Carlson 1997)
, although others have
reported null findings for this relation
(Davis et al. 2011,
2014; Fernyhough et al. 2007)
. Children with ICs are more
likely than their NIC counterparts (a) to know that their
minds are opaque to others (Davis et al. 2011), (b) to form
richer narratives when storytelling
(Trionfi and Reese 2009)
(c) show more sophisticated self-directed speech
et al. 2013)
, and (d) are better able to take the listener’s
perspective into account during a referential communication
(Roby and Kidd 2008)
. They also are more likely
to describe friends with reference to their mental
characteristics instead of behavioral tendencies or physical appearance
(Davis et al. 2014)
. Children with ICs also give these friends
the same status as real friends, using ICs as a way to avoid
loneliness in some instances
(Bouldin and Pratt 1999)
There is a paucity of scientific research on IC creation
by children with ASD. Given that children with ASD have
imagination deficits paired with ToM impairment, it would
seem that these children would not be strong candidates for
creating ICs. Many of the cognitive advantages explored
in TD children with ICs such as superior ToM, emotional
understanding, and social communication ability are in the
same domains as impairments seen in children with ASD.
Even the fact that girls are more likely to play with ICs
whereas children diagnosed with ASD are more typically
boys reduces the odds that a child with ASD would create
one of these friends. However, because children with ASD
were able to complete the
impossible person task” and partake in creating an
impossible being, it seems plausible that some children in the
population could have the means to create an IC spontaneously.
Furthermore, ICs have also been shown to have different
types of functions depending on their child creator
and Pratt 1999)
. For example, children with limited social
interactions will sometimes create ICs to talk with when
they have no one else to relieve loneliness. Children with
ASD often have less social interaction than their peers, so
this might be another indicator that it could be possible for
these children to create an IC on their own
(Bauminger et al.
. There have been parental accounts on internet chat
boards of children with ASD creating ICs, as well as
scholarly evidence indicating that children with ASD create ICs
The present study aimed to discover whether children
with ASD spontaneously create their own ICs, and whether
the form of IC (completely invisible or personified in an
object, the complexity of the IC’s characterization) differed
between ASD and TD children. We were also interested in
investigating whether the ASD and TD groups differed with
respect to the reported function of the IC or the age at which
ICs were created. In addition, we explored factors relating to
IC creation within the ASD group, investigating whether age
at which children were diagnosed with ASD or children’s
reported cognitive functioning related to the tendency to
create an IC. Finally, the present study asked whether
certain functions of ICs may be congruent with the needs (e.g.,
social communication) of children with ASD.
Participants were 111 parents of children with ASD (see
Table 1) between 24- and 96- months old (M = 59.10 months,
SD = 27.25); 79% (n = 88) were boys. The sample consisted
of 85.6% Caucasians, 4.5% African Americans, 2.7% Asian,
3.6% Biracial, and 3.6% preferred not to answer. Information
on economic status was not collected on this group.
Parents reported that their children were diagnosed with ASD
between 10 and 96 months of age (M = 43.05, SD = 22.52).
All diagnostic variables can be found in Table 1.
A second, previously existing data set of economically
diverse parents of typically developing children reporting on
their child’s IC status in a lab setting was used to compare
the two groups. Participants were randomly matched with
the ASD group based upon gender resulting in a TD group
of 104 parents of children (81 boys) aged between 59- and
64- months (M = 61.27, SD = 1.14). Only 81 boys
participated, so seven boys were not matched. University ethics
Parents of children with ASD were recruited through
materials stating that the researchers wanted to investigate how
children with ASD think. Materials were not directed at
parents that might be prone to have a child with an IC to provide
a better, less biased picture of how many children in this
population create ICs. Parents were recruited via (1) emails
to listservs for families of children with ASD, (2) emails
to agencies that had local, state, and national outreach, (3)
flyers posed on social media, and (4) flyers handed out at
different regional events. Parents then completed an online
questionnaire through Qualtrics. Parents were provided with
a link that they could access at their own computer or on
their phone. The questionnaire took anywhere from 3 to
20 min to fill out depending on whether the parent reported
that their child had created an IC.
Parents of typically developing children were
participating in a larger ongoing longitudinal study of which the IC
questionnaire was a part. They completed a variation of this
questionnaire in person on paper at a lab. They did not
complete the section on demographics, diagnostics, and
education at the time they were given the questionnaire. The last
seven free-response questions about ICs were also not given
to these parents.
Parents of children with ASD first completed four
forcedchoice questions about demographics: age, sex, race, and
ethnicity. These questions were followed by three
forcedchoice questions about the child’s educational environment:
whether they were publicly or privately educated, if they had
individualized education plans, the type of school programs
they attended, and a free response item to clarify school
situation. Finally, they answered three forced-choice questions
about the child’s diagnoses, checking all labels that applied,
including a high or low functioning option, and “other”
which allowed them to type in additional diagnoses, the type
of professional who provided the diagnoses, and the age at
time of an ASD diagnosis.
Next, parents were asked if their child had ever created
an IC. This is the point where TD group began their paper
questionnaire. All parents were asked, does your child have
an imaginary friend or have they had one in the past? (This
could be a completely imaginary friend or a toy or stuffed
animal that your child has played with for over three months
and has a stable personality). If parents of children with
ASD did not report their child having an IC, they were
thanked for participating and the questionnaire closed. If
the parent responded affirmatively, they were taken to an
Taylor and Carlson’s (1997)
All parents were asked about whether the IC was a doll/
toy, or completely invisible, the IC’s name, when their child
created the IC, the IC’s gender, age, and appearance, what
the child enjoys and dislikes about the IC, and where the
IC lives and sleeps. Parents of TD children ended at these
questions, and 24 of the 47 (51%) did not choose to answer
the more elaborate IC questions in this part of the larger
longitudinal study. Only two of the 18 (11%) parents of the
children with ASD failed to report more on their child’s IC.
The online questionnaire for parents of children with ASD
also asked free response questions about activities the child
and IC participate in together, what they talk about, how
and when the child stopped playing with the IC, whether the
child uses the IC to communicate needs, if the IC has any
unique qualities. ICs were classified as completely invisible
(iIC) or personified in an object (PO); examples of iICs and
POs can be found in Table 2.
Social Attribution Coding
Parent report on the open-ended questions where parents
were asked to describe various features of their child’s IC
was coded using a variation of Klin’s (2000) animation index
for social attribution. This index was initially developed to
measure the level of social attribution children with ASD
ascribe to geometric shapes enacting a social story, and was
adapted to ascertain the level of social attribution parents
reported when describing their children’s ICs. Descriptions
were placed into 3 exclusive and exhaustive categories of
attribution: (1) behaviors (e.g., playing with the IC, or
bossing the IC around); (2) cognition, intention, and motivation
(e.g., the IC being kind, or naughty); and (3) relationships
and personality traits (e.g. ‘the pretend version of a real girl’,
‘she is his baby’). Each category contained levels of
sophistication, so, the child’s score for social attribution to their IC
would increase with greater attributive sophistication. Raw
scores for each parent response were summed to a
composite score, and divided by the total number of descriptions,
yielding a percentage score (accounting for variability in the
number of responses provided).
Open-ended IC questions were analyzed using grounded
theory, which searches for prominent themes running
throughout the interviews and naturally arising in the
parent’s answers about their child’s ICs. Both ASD and TD
populations were examined.
Descriptive Statistics and Preliminary Analysis
Based on parent report, 16.2% of children diagnosed with
ASD were reported as having created ICs. Of the children
creating ICs, 38.8% were iICs and 61.1% were POs, such
as a stuffed toy or doll. Examples of these can be seen in
Table 2. Of the TD population, 45.2% of the children were
reported to have an IC. For this group, 70.2% were iICs
and 29.8% were POs. Child gender in the ASD group was
not significantly related to IC status, χ2(1) = 1.21, p = .272,
V = .10. There was also no association between gender
and IC status in the TD population χ2(1) = 1.53, p = .216,
V = .12. Age was not related to whether or not children
created ICs in the ASD F(1,93) = 0.53, p = .585, η2 = .012, or
TD, F(1,87) = 0.06, p = .814, η2 = .001, populations.
Differences in IC Creation Between ASD and TD
Chi square analyses were used to investigate whether
children diagnosed with ASD were as likely as their TD
counterparts to create ICs. IC creation was significantly less
common in the ASD sample, χ2(1) = 21.37, p < .001, V = .32.
When only looking at the differences between iIC and PO
creation, children in the ASD sample were less likely than
the TD children to create iICs η2(1) = 5.40, p = .020, V = .29.
Group differences in the age parents reported their child
beginning to play with their IC were investigated using a one
way ANOVA. Children with ASD began playing with ICs
significantly later than TD children according to their
parents, F(1,35) = 5.09, p = .031, η2 = .130. Age of first parent
Ghosty Bubble: an invisible bubble person who was fun to talk to and slept on a bubble bed next to the child.
When the child wanted to be alone he could be popped
Pretend Ada: An invisible version of a child’s school friend who plays with the child when she needs a friend
Mikey: An invisible ninja who lives in the sewer and is played with daily and read to by the child
Andrew: An invisible boy who drives a rainbow colored Lincoln and sleeps on a bunk bed
Hatch: A stuffed chicken that the child carries around and uses in stop motion films
Teddy: A stuffed bear used to help the child sleep and makes sure he has good dreams
Batman: A toy that the child plays with and likes because Batman helps people that are in trouble
reported IC interaction examining group differences between
iIC and PO creation was investigated in children with ASD
and then in TD children in an ANOVA. No effect was found
for the type of IC in ASD, F(1,16) = 0.19, p = .672, η2 = .012;
or TD populations, F(1,18) = 0.07, p = .800, η2 = .004.
Means and standard deviations for age related variables can
be found in Table 3.
Social Attribution Analysis
Parental reports on IC descriptions and interactions were
coded to reflect social attributions ascribed to the IC by
their children. The data for the ASD and TD groups are
presented in Table 3. There was no difference in social
attributes between the TD and ASD groups, t(37) = 1.36,
p = .251, d = .38. There were also no differences in social
attributions attributed in children with iICs versus POs in
either the ASD, t(16) = 1.72, p = .105, d = .84 or TD groups,
t(19) = − 0.55, p = .586, d = .24. Means and standard
deviations for social attribution variables can be found in Table 3.
IC Form and Function Analysis
The IC’s gender was then analyzed to determine if there were
group differences in children interacting with a male, female,
or non-gendered IC. No difference was found between the
TD and ASD groups, χ2(2) = 1.34, p = .512, V = .18. Parents
were also asked if their child dislikes anything about their
IC. Significantly more parents in the TD group reported their
children disliking things about their IC than in the ASD
population t(34) = 3.69, p = .002 (see Table 4).
Table 3 Mean (standard
deviation) scores as a function
of group and IC status
IC Form and Function Thematic Analyses
Parent description of their child’s ICs were then analyzed to
look further into the forms and functions that these entities
represented in the population of children with ASD. The
two main themes of social support or comfort were found
running through the questionnaires related to the function of
the child’s IC. Table 5 shows examples of parent responses
to the functionality their child’s IC. Parent reports were then
grouped into three categories of function (Social, Comfort,
and neither social nor comfort). Figure 1 shows the parent
report of functions of iICs and POs in children in the TD and
ASD populations. Fisher’s exact test indicated no significant
group difference in children’s functional use of their IC for
social or comfort purposes (p = .432).
Variables Related to ICs in Children Diagnosed with ASD
A binomial logistic regression was used to investigate
whether the age of first diagnosis related to the
likelihood of a child creating an IC. The regression model was
statistically significant χ2(1) = 6.74, p = .009. The model
explained 11.2% (Nagelkerke R2) of the variance.
Children’s age of diagnosis was further investigated by
looking at differences in age of diagnosis between iIC, NIC,
and PO children in an ANOVA finding a main effect of
group F(1,98) = 3.47, p = .035, η2 = .067. Post-hoc
analyses indicated that children creating POs were diagnosed at
a significantly younger age than NIC children (p= .037),
however there were no differences in age of diagnosis
Current age (months) IC
Age of IC creation (months) IC
Age of ASD diagnosis (months) IC
Raw attribution scores (all descrip- IC
tions including social attributions) PO
Social attribution scores (raw scores IC
divided by total descriptions) PO
*Note all ages are reported in months. Social attribution scores are raw scores divided by total number of
parent descriptions. Social attribution total scores exceeded the number of parent descriptions in four cases,
so percentages were over 100. Standard deviations are given in parentheses
aThere are missing data for some columns. The N reflects the number of parents that reported for specific
*Note this table consists solely of frequencies followed by within-group percentages in parentheses.
Percentages reflect total responses given not including missing data
aThe non-gendered variable refers to parents reporting that the IC had no gender or that a child had more
than one IC with more than one gender
TD children N (%)
between iIC and PO children (p = .822), or iIC and NIC
children (p = 1.000). All age-related variables can be found
in Table 3. Using Fisher’s Exact test, there were no
differences in IC creation between the high and low functioning
children (p = .603).
Parents of children diagnosed with ASD were also asked
whether their child uses the IC to communicate needs. Of the
17 parents reporting a child with an IC, three (17.6%, 1iIC
and two POs) of the parents responded that their child did
speak through the IC to communicate their needs, whereas
13 (76.5%, five iICs and eight POs) responded that their
child did not use the IC in this way, and one (5.9%, IC) was
This exploratory study’s main aim was to discover what
proportion of children diagnosed with ASD created ICs,
investigating potential differences in ICs both within the ASD
group and in comparison with TD children. Our findings
indicate that a substantial minority of children diagnosed
with ASD (16.2%) do spontaneously create ICs, with
similar features to TD children’s ICs in some respects (e.g., iIC
social and PO comfort functions). Furthermore, ICs’ gender
and their reported social attributions did not differ between
groups. However, parents of children with ASD were less
likely than those of TD children to indicate that their
children had created an IC to their knowledge, and when parents
did notice IC play, it was at a later age than TD parents.
Furthermore, the type of IC (iIC or PO) differed in that,
compared with their TD peers, the ASD sample created
fewer iICs, and were less likely to talk to their parent about
disliking parts of their IC’s personalities. Within group
differences in children diagnosed with ASD revealed earlier
diagnoses for children creating ICs and IC type.
Finding a subset of children with ASD who
spontaneously create ICs is contrary to much of the research on
imagination in children with ASD, and in some instances,
even the diagnosis itself
(APA 2013; Baron-Cohen 1987;
Jarrold et al. 1996)
. However, the argument has been made
that certain individuals presenting with ASD are very
creative and imaginative (Roth 2007), based on examining art
and literature created by ASD and TD individuals. Children
who did create ICs created them at later age than their TD
peers which is congruent with previous research on
imagination in children with ASD
(Rutherford et al. 2007; Wolfberg
et al. 2012)
, however in line with Roth (2007), our
findings showed that there were more similarities than
differences in ICs between the ASD and TD groups, suggesting
that this imaginative behavior is not necessarily affected in
children diagnosed with ASD. Children in both groups
created ICs and there were no gender differences in creation.
Furthermore, many of the features of the ICs created by TD
and ASD children were also similar. A similar degree of
social attributions was reported to be given to the ICs of both
groups, even when taking into account how many
opportunities parents had to describe their child’s IC. These results are
surprising, as past research on social attribution has found
that even children with ASD who are higher functioning
and can pass ToM tests have deficits in social cognition
. The similarities between parent descriptions
of their children’s social attributions across the ASD and TD
groups suggest that perhaps ICs among children with ASD
serve socially comparable purposes to those of TD children.
Although, since a higher proportion of parents in the ASD
group completed the more in-depth IC questions, it also may
be the case that the parents of children with ASD might be
more responsive to surveys about their child’s inner lives.
The finding that children diagnosed with ASD were not
reported to dislike things about their ICs as frequently as
TD children suggests that the ICs that are created by
children with ASD might have less well developed personalities
and minds, due to documented problems in ASD children
in conceptualizing other minds
(Baron-Cohen et al. 1985)
Alternatively, children diagnosed with ASD may be
spontaneously creating ICs that are qualitatively different from
those of TD children and may thus engage with them in
qualitatively different ways, for example using their IC for
comfort, social, or other purposes which have not been
analyzed statistically due to the small sample size of the IC
children. It has been shown that play often takes on
different forms in ASD
(Honey et al. 2007)
; perhaps IC play is
another example of this phenomenon. A second possibility
is that the parents of children with ASD and ICs may be
more focused on the mental lives as well as behaviors of
their children as a result of their diagnosis and emphasis put
on monitoring their child’s symptoms, thus making them
more likely to attend to their children’s interactions with
their IC, as the lack of parent report in the TD population
This parental attentiveness could also be a reason for
the finding within the ASD group that age of diagnosis was
related to IC status. Children who created POs were
diagnosed significantly younger than children who did not have
ICs; children who created iICs were not significantly
different in age from the other two groups. There are several
different reasons for children diagnosed earlier to create POs.
It may be the case that children diagnosed younger had more
access to early intervention, which promoted development
of social engagement. In addition, many factors, including
symptom presentation, early detection strategies used by the
child’s healthcare provider, urban versus rural locale, and
other sociodemographic variables relate to age of
(e.g., see Daniels and Mandell 2014 for review)
research could investigate whether these sociodemographic
variables relate to the type of IC created by children with
Other potential differences in ICs within the ASD
group were investigated in relation to parent reports of
whether their child was high or low functioning. Children
reported to be high versus low functioning did not differ
with respect to their tendency to create an IC. However
only four parents reported their child as low functioning,
despite the fact that 33 were reported to have an additional
developmental disorder. It is therefore difficult to draw
strong conclusions on the relation between functioning
level and tendency to create an IC in children with ASD.
To look at the functions played by ICs in both TD and
ASD populations, a thematic analysis was carried out on
both groups of parent descriptions of their child’s IC. In
terms of using an IC to communicate needs to a parent,
only three (17.6%) of the children with ASD with ICs
reportedly used their IC for communicative purposes. ICs
have been shown to have many communicative functions
in TD children
, but communication was not
probed in the TD questionnaire, so direct comparison is
not possible. Thematic analysis identified two main
functions of ICs: social and comfort purposes in both samples.
The hypothesis made about the functions of ICs entailed
the possible relief of loneliness, as this is reported as a
typical function of the IC, and research highlights higher
reported levels of loneliness in children with ASD
(Bauminger et al. 2003; Hoff 2005)
. Alleviating loneliness was
not, however, reported as a function of ICs in parents of
children with ASD.
The results reported here should be interpreted in light
of some limitations of the present study. All measures were
assessed by parental report. Although parents of TD children
have been shown to be the best at reporting the
imaginative lives of their children
, future research
should include child report, as suggested by
Parents may not be aware of the existence of an IC, and
this may be particularly the case in parents of children with
ASD, who may not be able to talk to or play with ICs as
obviously as TD children. The finding that ICs were
predominantly reported to be POs by parents in the ASD group
is in line with this suggestion. Future research should
therefore attempt to gather information on the existence of ICs in
children with ASD from the children themselves, and should
also assess autistic traits in the parents to explore whether
these may help explain the lower reported frequency of ICs
in children with ASD. Our findings open up some interesting
avenues for future research into the relation between ICs and
autism symptomology. For example, face-to-face research
would enable children with ASD’s cognitive abilities to be
assessed directly in order to investigate in greater depth how
individual differences in functioning (e.g. social vs. comfort)
and symptom severity in children with ASD relate to their
creation of ICs.
A second limitation is that the group of ASD children
reported to have ICs was small, meaning that the
withingroup analyses were lacking in power. The third issue is one
of generalizability. The parents of children with ASD were
recruited mainly from Facebook support groups and local
ASD events, and is therefore not a representative sample.
Finally, matching participants on age, and including a
comparison group of children with developmental delay to allow
for matching ASD children for cognitive functioning, would
provide more information on which aspects of IC creation
and play may be the unique to children diagnosed with ASD.
Broadly, our results confirm much of the research on
deficits in ASD children’s spontaneous imagination behavior.
However, the similarities identified between the ICs of TD
children and those with ASD who had created ICs suggest
that differences between TD and ASD children on this
specific high-level imaginative behavior may be quantitative
rather than qualitative.
Acknowledgments Research with the TD parents was made possible
by Grant RES-000-23-1073 from the Economic and Social Research
Council. Lorna Elliot, Alexandra Hearn, Beth Liddle, and Athena
Fernyhough also should be thanked for their work in data collection
Author Contributions PED and DLR conceived of the study, crafted
the experimental design, recruited participants, drafted portions of the
manuscript, and edited revisions. HS wrote up and submitted the ethics
proposal to gain ethical approval, implemented software for the data to
be collected, contributed to methods section of manuscript dealing with
the questionnaire. EM drafted portions of the manuscript and revised
the manuscript, contributed a previously existing data set, and edited
revisions. All authors read and approved the final manuscript.
Compliance with Ethical Standards
Conflict of interest Paige E. Davis declares that she has no conflict of
interest; Haley Simon declares that she has no conflict of interest;
Elizabeth Meins has been awarded a grant from the Economic and Social
Research Council RES-000-23-1073. Diana L. Robins declares that
she has no conflict of interest.
Ethical Approval All procedures performed in studies involving human
participants were in accordance with the ethical standards of the
institutional and/or national research committee and with the 1964 Helsinki
declaration and its later amendments or comparable ethical standards.
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
Creative Commons Attribution 4.0 International License (http://creativeco
mmons.org/licenses/by/4.0/), which permits unrestricted use,
distribution, 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.
American Psychiatric Association ( 2013 ). Diagnostic and statistical manual of mental disorders (5th edn .). Washington, D.C: American Psychiatric Association.
Baron-Cohen , S. ( 1987 ). Autism and symbolic play . British Journal of Developmental Psychology , 5 , 139 - 148 . https://doi.org/10.1111/ j.2044- 835X . 1987 .tb01049.x.
Baron-Cohen , S. , Leslie , A. M. , & Frith , U. ( 1985 ). Does the autistic child have a “theory of mind”? Cognition , 21 , 37 - 46 . https://doi. org/10.1016/ 0010 - 0277 ( 85 ) 90022 - 8 .
Bauminger , N. , Shulman , C. , & Agam , G. ( 2003 ). Peer interaction and lonliness in high-functioing children with autism . Journal of Autism and Developmental Disorders , 33 ( 5 ), 489 - 507 . https://doi. org/10.1023/A: 1025827427901 .
Bouldin , P. , & Pratt , C. ( 1999 ). Characteristics of preschool and schoolage children with imaginary companions . Journal of Genetic Psychology , 160 ( 4 ), 397 - 410 . https://doi.org/10.1080/0022132990 9595553.
Bruder , M. B. , & Chen , L. H. ( 2007 ). Measuring social competence in toddlers: Play tools for learning . Early Childhood Services , 1 , 49 - 70 .
Calver , A. ( 2009 ). The unicorn that came to stay and other stories… Mothers' experiences of having a child with autism with an imaginary friend (Unpublished doctoral dissertation ). University of East London, London.
Carlson , S. M. , & Taylor , M. ( 2005 ). Imaginary companions and impersonated characters: Sex differences in children's fantasy play . Merrill-Palmer Quarterly Journal of Developmental Psychology , 51 ( 1 ), 93 - 118 . https://doi.org/10.1353/mpq. 2005 . 0003 .
Daniels , A. M. , & Mandell , D. S. ( 2014 ). Explaining differences in age at autism spectrum disorder diagnosis: A critical review . Autism: The International Journal of Research and Practice , 18 ( 5 ), 583 - 597 . https://doi.org/10.1177/1362361313480277.
Davis , P. E. , Meins , E. , & Fernyhough , C. ( 2011 ). Self-knowledge in childhood: Relations with children's imaginary companions and understanding of mind . British Journal of Developmental Psychology , 29 ( 3 ), 680 - 686 . https://doi.org/10.1111/j.2044- 835X . 2011 . 02038 .x.
Davis , P. E. , Meins , E. , & Fernyhough , C. ( 2013 ). Individual differences in children's private speech: The role of imaginary companions . Journal of Experimental Child Psychology , 116 ( 3 ), 561 - 571 . https://doi.org/10.1016/j.jecp. 2013 . 06 .010.
Davis , P. E. , Meins , E. , & Fernyhough , C. ( 2014 ). Children with imaginary companions focus on mental characteristics when describing their real-life friends . Infant and Child Development , 23 ( 6 ), 622 - 633 . https://doi.org/10.1002/icd. 1869 .
Dawson , G. , Toth , K. , Abbott , R. , Osterling , J. , Munson , J. , Estes , A. , & Liaw , J. ( 2004 ). Early social attention impairments in autism: Social orienting, joint attention, and attention to distress . Developmental Psychology , 40 ( 2 ), 271 - 283 . https://doi. org/10.1037/ 0012 - 1649 . 40 .2.271.
Fernyhough , C. , Bland , K. , Meins , E. , & Coltheart , M. ( 2007 ). Imaginary companions and young children's responses to ambiguous auditory stimuli: Implications for typical and atypical development . Journal of Child Psychology and Psychiatry , 48 , 1094 - 1101 . https://doi.org/10.1111/j.1469- 7610 . 2007 . 01789 .x.
Giménez-Dasi , M. , Pons , F. , & Bender , P. ( 2016 ). Imaginary companions, theory of mind and emotion understanding in young children . European Early Childhood Education Research Journal , 24 ( 2 ), 186 - 197 . https://doi.org/10.1080/1350293X. 2014 . 919778 .
Gleason , T. R. ( 2004 ). Imaginary companions: An evaluation of parents as reporters . Journal of Infant and Child Development , 13 ( 3 ), 199 - 215 . https://doi.org/10.1002/icd.349.
Gleason , T. R. , & Hohmann , L. M. ( 2006 ). Concepts of real and imaginary friendships in early childhood . Social Development , 15 ( 1 ), 128 - 144 . https://doi.org/10.1111/j.1467- 9507 . 2006 . 00333 .x.
Harris , P. L. ( 2000 ). The work of the imagination . Massachusetts: Blackwell.
Hoff , E. V. ( 2005 ). A friend living inside me-The forms and functions of imaginary companions . Imagination, Cognition and Personality , 24 ( 2 ), 151 - 189 . https ://doi. org/10.2190/4M9J-76M2 - 4Q4Q-8KYT.
Honey , E. , Leekam , S. , Turner , M. , & McConachie , H. ( 2007 ). Repetitive behaviour and play in typically developing children and children with autism spectrum disorders . Journal of Autism and Developmental Disorders , 37 ( 6 ), 1107 - 1115 . https://doi. org/10.1007/s10803-006-0253-4.
Howes , C. ( 1988 ). Peer interaction of young children . Monographs of the Society for Research in Child Development , 53 , 1 - 92 .
Jarrold , C. ( 2003 ). A review of research into pretend play in autism . Autism , 7 ( 4 ), 379 - 390 . https://doi.org/10.1177/136236130300700 4004.
Jarrold , C. , Boucher , J. , & Smith , P. K. ( 1996 ). Generativity deficits in pretend play in autism . British Journal of Developmental Psychology , 14 ( 3 ), 275 - 300 . https://doi.org/10.1111/j.2044- 835X . 1996 . tb00706.x.
Karmiloff-Smith , A. ( 1990 ). Constraints on representational change: Evidence from children's drawing . Cognition , 34 ( 1 ), 57 - 83 . https ://doi.org/10.1016/ 0010 - 0277 ( 90 ) 90031 -E.
Klin , A. ( 2000 ). Attributing social meaning to ambiguous visual stimuli in higher-functioning autism and Asperger syndrome: The social attribution task . Journal of Child Psychology and Psychiatry , 41 ( 7 ), 831 - 846 . https://doi.org/10.1111/ 1469 - 7610 . 00671 .
Leslie , A. M. ( 1994 ). Pretending and believing: Issues in the theory of ToMM . Cognition, 50 , 211 - 238 .
Lewis , V. , & Boucher , J. ( 1988 ). Spontaneous, instructed and elicited play in relatively able autistic children . British Journal of Developmental Psychology , 6 ( 4 ), 325 - 339 . https://doi.org/10.1111/ j.2044- 835X . 1988 .tb01105.x.
Low , J. , Goddard , E. , & Melser , J. ( 2009 ). Generativity and imagination in autism spectrum disorder: Evidence from individual differences in children's impossible entity drawings . British Journal of Developmental Psychology , 27 , 425 - 444 . https://doi. org/10.1348/026151008X334728.
Manosevitz , M. , Prentice , N. M. , & Wilson, F. ( 1973 ). Individual and family correlates of imaginary companions in preschool children . Developmental Psychology , 8 ( 1 ), 72 - 79 . https://doi.org/10.1037/ h0033834.
Marsh , R. L. , Landau , J. D. , & Hicks , J. L. ( 1996 ). How examples may (and may not) constrain creativity . Memory and Cognition , 24 ( 5 ), 669 - 680 . https://doi.org/10.3758/BF03201091.
Motoshima , Y. , Shinohara , I. , Todo , N. , & Moriguchi , Y. ( 2014 ). Parental behavior and children's creation of imaginary companions: A longitudinal study . European Journal of Developmental Psychology , 11 ( 6 ), 716 - 727 . https://doi.org/10.1080/17405 629. 2014 . 932278 .
Pearson , D. , Burrow , A. , Fitzgerald , C. , Green , K. , Lee , G. , & Wise , N. ( 2001 ). Auditory hallucinations in normal child populations . Personality and Individual Differences , 31 , 401 - 407 . https://doi. org/10.1016/S0191- 8869 ( 00 ) 00145 - 8 .
Roby , A. C. , & Kidd , E. ( 2008 ). The referential communication skills of children with imaginary companions . Developmental Science , 11 ( 4 ), 531 - 540 . https://doi.org/10.1111/j.1467- 7687 . 2008 . 00699 .x.
Roth , I. ( 2007 ). Autism and the imaginative mind . In I. Roth (Ed.), Imaginative minds. Oxford: British Academy and Oxford University Press.
Rutherford , M. D. , Young , G. S. , Hepburn , S. , & Rogers , S. J. ( 2007 ). A longitudinal study of pretend play in autism . Journal of Autism and Developmental Disorders , 37 ( 6 ), 1024 - 1039 . https://doi. org/10.1007/s10803-006-0240-9.
Scott , F. J. , & Baron-Cohen , S. ( 1996 ). Imagining real and unreal things: Evidence of a dissociation in autism . Journal of Cognitive Neuroscience , 8 ( 4 ), 371 - 382 . https://doi.org/10.1162/ jocn. 1996 . 8 .4.371.
Singer , D. G. , & Singer , J. L. ( 1990 ). The house of make-believe: Children's play and the developing imagination . Cambridge: Harvard University Press.
Svendsen , M. ( 1934 ). Children's imaginary companions . Archives of Neurology and Psychiatry , 32 ( 5 ), 985 - 999 . https://doi. org/10.1001/archneurpsyc. 1934 . 02250110073006 .
Taylor , M. ( 1999 ). Imaginary companions and the children who create them . Oxford: Oxford University Press.
Taylor , M. , & Carlson , S. ( 1997 ). The relation between individual differences in fantasy and theory of mind . Child Development , 68 ( 3 ), 436 - 455 . https://doi.org/10.2307/1131670.
Ten Eycke , K. D. , & Müller , U. ( 2015 ). Brief report: New evidence for a social-specific imagination deficit in children with autism spectrum disorder . Journal of Autism and Developmental Disorders , 45 , 213 - 220 . https://doi.org/10.1007/s10803-014-2206-7.
Trionfi , G. , & Reese , E. ( 2009 ). A good story: Children with imaginary companions create richer narratives . Child Development , 80 ( 4 ), 1301 - 1313 . https://doi.org/10.1111/j.1467- 8624 . 2009 . 01333 .x.
Uren , N. , & Stagnitti , L. ( 2009 ). Pretend play, social competence and involvement in children aged 5-7 years: The concurrent validity of the child-initiated pretend play assessment . Australian Occupational Therapy , 56 ( 1 ), 33 - 40 . https://doi.org/10.111 1/j.1440- 1630 . 2008 . 00761 .x.
Wing , L. , & Gould , J. ( 1979 ). Severe impairments of social interaction and associated abnormalities in children: Epidemiology and classification . Journal of Autism and Developmental Disorders , 9 ( 1 ), 11 - 29 .
Wolfberg , P. J. , Bottema-Beutel , K. , & DeWitt , M. ( 2012 ). Including children with autism in social and imaginary play with typical peers: Integrated play groups model . American Journal of Play . https://doi.org/10.1007/s10803-014-2245-0.
World Health Organization. ( 1992 ). The ICD-10 Classification of mental and behavioural disorders: Clinical descriptions and diagnostic guidelines . Geneva: World Health Organization.