Content uploaded by Erin Baker
Author content
All content in this area was uploaded by Erin Baker on Feb 01, 2024
Content may be subject to copyright.
Content uploaded by Erin Baker
Author content
All content in this area was uploaded by Erin Baker on Jul 29, 2021
Content may be subject to copyright.
Journal of Applied Developmental Psychology 76 (2021) 101314
Available online 4 August 2021
0193-3973/Published by Elsevier Inc.
Theory of mind development in impoverished U.S. children and six
cross-cultural comparisons
Erin Ruth Baker
*
, Rong Huang , Carmela Battista , Qingyang Liu
Department of Educational and Counseling Psychology, University at Albany, State University of New York, Albany, NY 12222, United States of America
ARTICLE INFO
Keywords:
Theory of mind
Low-income
Poverty
Early childhood
Head start
ABSTRACT
Prior literature demonstrates a reliable progression of ToM development during early childhood; however, cross-
cultural studies have found a different progression of ToM development, and few studies have investigated
whether this progression is generalizable to children in poverty. The current study examined ToM development
in 106 3- to 5-year-old children enrolled in urban, low-income preschool programs in the U.S. Children
completed a 5-task ToM battery. Parents provided family demographic information. Results demonstrated that
economically disadvantaged preschool children do not develop ToM capacities in the same sequence as their
afuent peers. Comparisons were then made with six samples reported in previous studies; children here showed
a difference in age of mastery for some, but not all, aspects of ToM, compared with afuent U.S. children, and
moreover show greater similarity with low-income children worldwide. These ndings add to our understanding
of specic environmental inuences on ToM development in early childhood.
Introduction
Worldwide, approximately one in three children live in poverty
(UNICEF, 2018). In the United States, nearly 20% of all children expe-
rience poverty (U.S. Census Bureau, 2018). Poverty exposure is a sig-
nicant factor affecting children's developmental outcomes, particularly
during early childhood when children's nascent Theory of Mind (ToM)
develops considerably (Duncan & Brooks-Gunn, 2000; Petterson &
Albers, 2001). Although considerable research efforts have focused on
how poverty impacts developmental outcomes (Frankenhuis & Nettle,
2020), few studies have explored how poverty exposure impacts ToM
development. Moreover, studies that have focused on ToM development
in poverty have not examined this relationship in the U.S. As such, the
task of the current study is to consider how children's ToM development
may be inuenced by low-income conditions, specically within an
urban, Head Start sample in the U.S.
Development of theory of mind
Theory of Mind (ToM) is the understanding of others' mental states,
including desires, feelings, and thoughts, and this capacity serves in part
to predict goal-directed behaviors (Gopnik & Astington, 1988;
Shahaeian, Henry, Razmjoee, Teymoori, & Wang, 2015; Wellman & Liu,
2004). ToM has been shown to be a salient predictor of children's social
competence (Devine, White, Ensor, & Hughes, 2016; Razza & Blair,
2009) as well as academic achievement (Blair & Razza, 2007; Cavadel &
Frye, 2017). Although ToM is universal in neurotypical development,
cross-cultural evidence demonstrates that environmental features
impact both the emergence and progression of ToM development. Early
research on ToM development consistently identied that ToM un-
dergoes rapid development during early childhood, with the progression
occurring in a relatively predictable fashion. That is, although ToM
development is protracted, taking place over several years, the devel-
opment was generally thought to occur in a specic sequence. For
instance, many children can recognize that another person will have a
desire different from their own (e.g., another child preferring fruit
snacks over crackers) before they can recognize that another person has
a different belief or knowledge base (Wellman & Liu, 2004).
To this point, the precise theoretical framework for ToM is compli-
cated in the literature, but largely mirrors a more nuanced theoretical
argument of the role nature versus nurture in the presence of charac-
teristics and abilities. Arguments as to how ToM emerges consider the
intersection between natural selection favoring an ability versus the
environment encouraging it during development, and whether it results
* Corresponding author at: Division of Educational Psychology and Methodology, Department of Educational and Counseling Psychology, University at Albany,
State University of New York, Albany, NY 12222, United States of America.
E-mail address: erbaker@albany.edu (E.R. Baker).
Contents lists available at ScienceDirect
Journal of Applied Developmental Psychology
journal homepage: www.elsevier.com/locate/jappdp
https://doi.org/10.1016/j.appdev.2021.101314
Received 11 September 2020; Received in revised form 29 July 2021; Accepted 29 July 2021
Journal of Applied Developmental Psychology 76 (2021) 101314
2
from changes in how one reasons about environmental stimuli versus
differences in the amount and type of environmental information in
one's experience (Heyes, 2003). For the current purposes, ToM is a
universal phenomenon, emerging in early childhood, with differences in
sequence and timing emerging as a result of environment.
Typically developing children usually acquire foundational ToM by
4 years of age (e.g., recognition that a character may have mistaken
beliefs about reality; Cutting & Dunn, 1999). Importantly, myriad
research demonstrates that ToM involves the consideration of multiple
mental states (i.e., intentions, desires, beliefs, emotions) which may not
reect reality and which may change for a single person over time (e.g.
Kuntoro, Saraswati, Peterson, & Slaughter, 2013; Wellman, Fang, Liu,
Zhu, & Liu, 2006; Wellman & Liu, 2004). For instance, although false
belief understanding is a signicant milestone in children's ToM devel-
opment, this aspect does not extend to perceptions of differences in
emotional states. In other words, a young child may be able to recognize
that a friend will look for a toy where they placed it last, but the same
young child will have difculty understanding that a person who looks
excited may in fact be quite sad (e.g., a parent “putting on a brave face”
upon hearing distressing information).
Wellman and Liu (2004) sought to capture the nuances of ToM
development, and created a battery of ToM tasks for researchers to
examine the full ToM progression. This battery includes measures of
diverse desires (DD), diverse beliefs (DB), knowledge access (KA),
explicit false belief (FB), and hidden emotion (HE) (see Table 1 for
detailed descriptions of each conceptual task). As such, this battery, or a
facsimile, is now ubiquitous in ToM research. Each of these aspects
represents unique understandings of mental states, and together
constitute a comprehensive understanding of ToM. This battery allows
researchers to fully consider ToM development through early childhood
across different groups of children, and to make comparisons, for
instance, across culture and nationality (Peterson & Wellman, 2009;
Shahaeian, Peterson, Slaughter, & Wellman, 2011; Wellman, Fang, &
Peterson, 2011; Zhang, Shao, & Zhang, 2016). More recently, re-
searchers have agreed that although there is a large maturational
component to the development of ToM that occurs in line with brain
development (Wiesmann, Schreiber, Singer, Steinbeis, & Friederici,
2017), the environment informs the sequence and precise timing of
specic aspects of ToM. In other words, ToM in general develops during
early childhood, but the environment informs in what order and at what
time.
Cultural inuences on theory of mind development
Previously, it was believed that ToM developed in the preordained
progression as described above (i.e., diverse desires, then diverse beliefs,
and so on) for all children; that is, the capacity to understand diverse
desires would emerge rst, with continued skill mastery taking place
over some time while diverse belief understanding would emerge, and
so on. Even in children with certain disabilities from afuent Western
samples, ToM appeared to develop in the same sequence though with
delays in their emergence (Smogorzewska, Szumski, & Grygiel, 2018).
More recently, however, researchers have turned to examinations of
differences in ToM emergence and acquisition as a function of envi-
ronmental sensitivity (Shahaeian, 2015; Shahaeian et al., 2011; Well-
man et al., 2011).
Much of this research has focused on how cultural values, such as a
community's tendency to prioritize the collective group over the indi-
vidual, impacts ToM development (Shahaeian, 2015; Shahaeian, Niel-
son, Peterson, Aboutalebi, & Slaughter, 2015). Most studies that have
examined cultural differences have conceptualized culture as either
“collectivist”, characterized by prioritizing group harmony and common
in places such as China, India, and Iran, or “individualist”, characterized
by priority for individual independence and more common in places
such as the U.S. and Australia. Although the developmental sequence of
ToM has been validated in typically-developing children in the United
States (Wellman et al., 2006; Wellman & Liu, 2004), Australia (Peterson,
Wellman, & Liu, 2005; Wellman et al., 2006), and Germany (Kristen,
Thoermer, Hofer, Aschersleben, & Sodian, 2006) – that is, individual-
istic communities – children from samples in collectivist countries show
a different sequence of ToM faculties.
Consider passage rates on identical tasks for children in the U.S. and
China. Although children in Chinese and U.S. samples passed the same
overall number of tasks at the same approximate ages, the specic tasks
– and thus the conceptual mastery – was different for Chinese versus U.S.
children. (Wellman et al., 2006). Children raised in Beijing reliably
passed the Knowledge Access (KA) task before the task assessing Diverse
Beliefs (DB), which counters the sequence reliably found for children in
the United States. In other words, compared with their U.S. peers, Chi-
nese children developed the capacity to recognize that exposure to an
event yields knowledge of the event at a younger age, yet developed the
ability to recognize that two individuals can hold diverging beliefs about
an event at an older age (Wellman et al., 2006). A similar developmental
sequence has been found amongst samples of Iranian children for di-
versity of opinion and knowledge access (Shahaeian et al., 2011). Sha-
haeian et al. (2011) argued that deeply-rooted features in collectivistic
cultures, such as reverence for elders and knowledge, should ecologi-
cally prioritize knowledge access over diverse beliefs, thus leading to
differences in the sequence of ToM faculties. The impacts of poverty on
ToM development, however, remain largely underexplored.
Poverty inuences on theory of mind development
Broadly, socioeconomic status (SES) has been found to impact the
sequence and emergence of ToM during early childhood, although this
has been studied to a much lesser extent than culture. Kuntoro et al.
(2013) examined ToM development in children from two socio-
economically diverse Indonesian communities (i.e., a middle-class
community and a pemulung, that is, trash picker, community). Chil-
dren overall developed false belief understanding (FBU) at approxi-
mately the same age, however, they differed in the sequence of other
ToM aspects, namely mastery of knowledge access (KA) and hidden
emotion (HE). The pemulung children developed the understanding of
both knowledge access (KA) and hidden emotion (HE) at an older age
than their more afuent peers. Additionally, the two groups differed in
the sequence in which ToM aspects developed, with false belief under-
standing (FB) in pemulung children developing before knowledge access
(KA), and the reverse being true for the middle-class group.
Similarly, Shahaeian (2015) compared ToM development for several
Iranian communities across three different socioeconomic backgrounds:
High-SES urban, Low-SES urban, and Rural. Across each of these Iranian
samples, knowledge access (KA) developed earlier than diverse beliefs
(DB), which is counter to the progression found in collectivist samples.
However, the three different groups in the Shahaeian (2015) study
differed in emergence and age of mastery for some tasks. High-SES
urban children demonstrated earlier mastery of diverse desires (DD),
knowledge access (KA), and diverse beliefs (DB), compared to both Low-
Table 1
Descriptions of the ve-task ToM battery by Wellman and Liu (2004).
Tasks Descriptions
Diverse Desires
(DD)
Child judges that two persons (child vs. target) have different
desires about the same objects.
Diverse Beliefs (DB) Child judges that two persons (child vs. target) have different
beliefs about the same object, when the child does not know
which belief is true or false.
Knowledge Access
(KA)
Child sees the contents of a box and judges the knowledge of
an ignorant target.
Explicit False Belief
(FB)
Child judges how the target will search for an object, given
the target's mistaken belief of object's location.
Hidden Emotion
(HE)
Child judges target's capacity to display and feel disparate
emotions.
E.R. Baker et al.
Journal of Applied Developmental Psychology 76 (2021) 101314
3
SES urban and Rural children, yet children from the Low-SES urban
group performed equivalently to the High-SES urban group for false
belief understanding (FBU), while the Rural group lagged behind both.
Finally, the Rural group developed Hidden Emotions (HE) at a younger
age than the other two groups.
Within individualistic societies, differences in ToM emergence have
been found across SES. Children from low-income U.S. families tend to
reliably pass false belief understanding (FBU) tasks at an older age
compared with their middle-class peers (Curenton, 2004; Holmes, Black,
& Miller, 1996; Holmes-Lonergan, 2003; Seidenfeld, Johnson, Cavadel,
& Izard, 2014). However, prior research on poverty and ToM develop-
ment in individualistic cultures have only examined one area of ToM
(typically, FBU), and have not used a comprehensive ToM battery. As
much of the cross-cultural work has found differences for ToM tasks
other than FBU, it is possible that other differences exist but have
remained untested. The current study expands upon these prior studies
by testing the full range of ToM capacities with the 5-task battery.
In the United States, low-income families tend to experience unstable
home environments, limited resources, and greater levels of nancial
stress (Blair, Granger, & Razza, 2005; DeJoseph, Sifre, Raver, Blair, &
Berry, 2021), corresponding with different ToM performance rates for
children (Cavadel & Frye, 2017; Holmes-Lonergan, 2003). The mecha-
nisms by which poverty exposure might transmit a unique set of social
norms are numerous. For instance, infants raised in poverty are excep-
tionally skilled at detecting danger, and detecting and responding to
negative affect (Wass et al., 2019; for review, see Frankenhuis & Nettle,
2020). This extends to other types of adversity as well, as child victims of
physical or domestic abuse perform worse on cognitive ToM tasks (i.e.,
ToM tasks specically oriented around cognitive mental state under-
standing; e.g., FB), but not affective ToM tasks (i.e., oriented on
emotional mental state understanding; e.g., HE), compared with chil-
dren who experienced no abuse (Heleniak & McLaughlin, 2019). These
ndings mirror work by researchers examining adversity and emotional
knowledge in adults—women with post-traumatic stress disorder
(PTSD) from childhood abuse show faster response times in recognizing
negatively-valenced emotional mental states compared with neutral
mental states (Nazarov et al., 2014). Impoverished U.S. adults have also
shown affective ToM performance exceeding that of non-impoverished
adults (Dietze & Knowles, 2020).
Together, these ndings may suggest that individuals who experi-
ence specic types of adversity may be especially skilled at interpreting
and responding to emotionally-salient information. For instance, nega-
tive emotional experiences in childhood, such as exposure to increased
stress, may serve as a buffer for developing emotionally-salient ToM
(Anderson et al., 2015; Barbarin, 1993; Edwards & Few-Demo, 2016;
Herruzo, Raya Trenas, Pino, & Herruzo, 2020; Sexton & Pennebaker,
2004). If this is true, these children should develop hidden emotion (HE)
conceptual mastery at a rate that meets or exceeds low-risk peers.
Cognitive components of ToM may differ across income strata as
well. For example, parents that communicate more frequently and by
using more reective styles of communication tend to have children
with higher ToM mastery (Cheung, 2010; Turner, 2018). Opportunities
for such communication may be less common for families in poverty, as
caregivers may be working in several positions outside of the home. The
nancial strain placed on low-income parents may expand to less direct
parenting compared with the parenting characteristics seen in middle
class families (McLoyd, 1990; Yeung, Linver, & Brooks-Gunn, 2002).
Financial strain experienced by low-income families may restrict
parents' abilities in providing more time, energy, and resources to
facilitate children's learning and development (Dearing, 2004). More-
over, low-income families may have less access to learning materials,
and thus may inadvertently place less emphasis on children's knowledge
acquisition (KA) (DeJoseph et al., 2021; Longo, McPherran Lombardi, &
Dearing, 2017). It is therefore possible that children living in poverty in
the U.S. may experience later knowledge access (KA) development,
compared to both afuent U.S. and impoverished Iranian children, as
this aspect may not be particularly relevant for their environment.
Overall, the limited research on economically-impoverished chil-
dren's ToM in individualistic cultures suggests that their ToM develop-
ment does not mirror that of their afuent peers. These differences may
seem minor, but they hold important implications about children's
development broadly. Studies of afuent, typically white, populations
have been oversampled in developmental research (Nzinga et al., 2018;
Roberts, Bareket-Shavit, Dollins, Goldie, & Mortenson, 2020), which has
limited implications for researchers and policymakers working in
diverse populations (Frankenhuis & de Weerth, 2013; Frankenhuis &
Nettle, 2020).
In sum, the ToM development found amongst children in Iran
(Shahaeian, 2015) and China (Wellman et al., 2006), as well as children
from varying economic risk levels (Kuntoro et al., 2013; Shahaeian,
2015), suggests that – although the preschool years may be matura-
tionally ripe for ToM development – environmental and cultural dif-
ferences inform the precise emergence and sequence of specic ToM
faculties. Differences in resources and daily interactions might lead
children to attend to some aspects of social dynamics before others
(Astington, 2001; Dietze & Knowles, 2020). More simply, certain fac-
ulties may be more, or less, critical for adequate social navigation, and
these levels of necessity vary as a function of one's environment. How-
ever, this has been largely underexplored for children raised in poverty
in the U.S.
Current study
To address whether economically disadvantaged preschool children
develop ToM capacities in a way that represents their unique environ-
mental niche, and therefore differs from that reported in previous
studies of afuent children, this study assesses low-income, urban U.S.
families and their children. Based on previous studies, three hypotheses
were drawn:
(1) Reecting on the research focused on cultural differences, we
anticipate that children in this sample will develop knowledge
access (KA) sequentially after diverse desires (DD), diverse be-
liefs, (DB) and false beliefs (FB).
(2) Prior studies of impoverished U.S. communities have shown that
children in these communities often develop mastery of certain
ToM aspects at a later age than more afuent children, but this
may be especially true for diverse desires (DD), diverse beliefs,
(DB), and knowledge access (KA); as such, we expect that passage
rates for these tasks will be lower than passage rates reported for
afuent U.S. samples of the same age.
(3) Children reared in poverty may be exposed to more negatively-
valenced emotionally-salient events, which may correspond
with an ecological need for better emotional perspective taking.
We therefore hypothesize that children in this sample will
develop an understanding of hidden emotions (HE) in the same
sequence as demonstrated in both cross-cultural and low-income
studies (that is, last in comparison to other ToM aspects), but
conceptual mastery of HE will take place at the same age, if not
younger, than seen elsewhere.
Methods
Participants and procedures
Upon IRB approval, consent was obtained from school administra-
tors, and then from parents and their children. Data were analyzed for
106 preschool children (51% boys, M
age
=52.77 months, SD =6.61
[range: 37.17–63.67]) recruited from an urban Head Start program in
the northeast region of the United States. Head Start is a federally-
funded preschool program for low-income communities and families.
As a result, the current sample of children came from families living at or
E.R. Baker et al.
Journal of Applied Developmental Psychology 76 (2021) 101314
4
below the poverty threshold as dened by income-to-needs ratio (M
in-
come
=$23,030.72; Median =$18,000; Range [$1100 - $85,000]); for
comparison, the average household in this county has a median house-
hold income of roughly $43,000. Children predominantly lived in
single-parent homes (64.2%), usually with a biological mother (77%).
Families typically had more than one child living at home (Median =3;
Range: [1–7]).
The sample was predominantly children of Color: roughly 30% was
African American, Black, or African, 17% was Multi-Racial, about 15%
was White, and roughly 27% was reported as “Other” (see Table 2 for
full ethnic and racial reporting). It is also worth noting that this area of
the country boasts a high level of immigration, specically Sudanese and
Guyanese (roughly 6% of this school's population typically self-identies
as one of these two nationalities), although this data was not captured in
the current study. This sample came from the county with the highest
crime rate in the state, which is double the average U.S. violent crime
rate.
Additionally, this sample had slightly lower levels of education than
most low-income U.S. samples. As shown in Table 2, roughly 1 out of
every 5 mothers and 1 out of every 4 fathers did not complete high
school or obtain a GED, and a few did not nish primary school. Mothers
tended to be more educated than fathers, although it is important to
recall that many children did not live with a father gure (see Table 2 for
full summary of demographics).
Parents responded to an open-ended question asking them to report
their occupation (“waiter” or “homemaker”). This was then converted to
a numerical variable called Job Zone by using the Occupational Infor-
mation Network (O*NET; Peterson et al., 2001). O*NET uses a system of
categorization in which similar jobs are grouped into Job Zones based on
several criteria, including required education, relevant experience, and
necessary job training (O*NET; Peterson et al., 2001). Higher values on
the Job Zone index indicate jobs that require more education, training,
etc. The mean Job Zone was 1.90, indicating that the majority of parents
worked in occupations that required little formal preparation or
training.
In addition to demographics, parents answered questions regarding
subjective measures of nancial strain (e.g., “I can afford the things I
need to buy for my family”; Conger, Ge, Elder Jr, Lorenz, & Simons,
1994). Parents' reports of nancial strain did not overwhelmingly indi-
cate that they subjectively experienced nancial struggle. Most care-
givers indicated that they were able to afford a suitable home for their
family (67.4%), and that they could afford to meet their family's nutri-
tional (69.9%) and medical needs (58.4%). When asked explicitly about
their overall nancial difculty, most indicated they had “just enough to
get by at the end of the month”, and “some difculty” in paying bills.
Measures
Children completed the ve-task ToM battery (Wellman & Liu, 2004)
as described in Table 1. Each task was scored such that successful
demonstration of competence on that ToM concept received a score of 1
on that task, with failure to demonstrate competency earning a score of
0.
Diverse desires
Children were presented two different hypothetical snack choices (i.
e. a picture of chocolate cake and a broccoli) and asked which snack they
would like to eat (own-desire question). The children were then told,
“That's a really good choice. But, Mr. Smith doesn't like [child's choice],
he really likes [the other option]. Blech, he hates [child's choice].”
Children were then asked to select a snack for Mr. Smith (target question).
Responses were coded as 1 if the responses to the own-desire and target
questions opposed one another. This task was originally created by
Wellman and Woolley (1990).
Diverse beliefs
In the DB task (Flavell, Green, Flavell, & Watson, 1986), a sponge
painted to look like a rock was shown to the child, and the child was
asked what they believed the item to be (control question), to which all
children said “rock” or a similar item (e.g., “toy rock”). After being given
the opportunity to explore the item, the child was asked what the object
was “really and truly”, (target question a). Lastly, the child was asked
what another child would think the item was if they walked in right now
(target question b). Responses were coded as 1 if a child's response to the
target question a was different from target question b.
Knowledge access
Children were shown a black, nondescript, plastic box that – once
opened – was found to contain marbles. Prior to opening the box, chil-
dren were asked “What do you think is in this box?” The child could give
any response to indicate that he did not know the contents. Next, the
child was encouraged to open the box, and asked what the box actually
contains. Children gave a verbal indication of the contents (e.g., “It has
marbles!”). Then, the box was closed, and the child was asked “Now,
what's really in the box?”
Next, a toy gurine of a boy was brought out: “This is Davey, Davey
has never seen inside this box. Here comes Davey! Does Davey know
what's inside the box?” (target question), “Has Davey seen inside this
box?” (memory question). To earn a score of 1, the child had to respond
“no” to both the target and the memory question. This task was originally
developed by Pillow (1989), but was adapted by Wellman and Liu
(2004).
Table 2
Demographic information for children, parents, and households.
M (range) %
Child Demographics Age (months) 52.77 (37.17–63.67)
Race/Ethnicity
African/African
American/Black
29.2%
White 14.2%
Latinx 9.4%
Asian American 1.9%
Multiracial 17.0%
Other 26.8%
Parent
demographics
Job Zone 1.94 (1.00–5.00)
Mother Ed (82%
reporting)
<High School diploma 19.8%
High school or GED 22.6%
Some college 21.7%
College degree 15.1%
Graduate degree 2.8%
Father Ed (75%
reporting)
<High School diploma 23.6%
High school or GED 26.4%
Some college 10.4%
College degree 10.4%
Some graduate school 1.9%
Graduate Degree 1.9%
Household
demographics
Family Size 3.61 (1–8)
Siblings (78%
reporting)
0 13.2%
1 22.6%
2 20.8%
3 11.3%
>3 10.4%
Economic strain 2.05 (1.00–3.43)
Poverty threshold (%
below)
90.6%
Annual income $23,030 ($1100 -
$85,000)
Note. Mother Ed =mother's education. Father Ed =father's education.
E.R. Baker et al.
Journal of Applied Developmental Psychology 76 (2021) 101314
5
False belief
In the FB task, we adapted the classic Sally-Anne task (Baron-Cohen,
Leslie, & Frith, 1985). In the Sally-Anne task, children are told “Sally has
a ball, and she puts her ball in this basket. Then, Sally leaves the room.
[Sally puppet is placed out of sight.] Next, Anne comes in, and she moves
Sally's ball to be over here in this box.” Children are then asked where
they think Sally would look for her ball rst. We adapted this task
slightly by using puppets of common farm animals (a cow and a pig)
instead of puppets of children, to eliminate any possible bias about race
or gender. In scoring responses, children earned a score of 1 if they
indicated that the target character would look for her ball in the basket.
Hidden emotion
The HE task (also referred to as Real-Apparent Emotion; Wellman &
Liu, 2004) described a story about a boy named Matt. Children were rst
oriented to an image showing three faces: happy, sad, and neutral.
Children were rst asked to identify each face (e.g., “can you show me
the happy face?”).
Children were then shown an image of the back of a boy's head, and
told “Here's Matt. Matt's friends were playing together and telling jokes.
One of the older children, Rosie, told a mean joke about Matt, and
everyone laughed. Everyone thought it was very funny, but not Matt.
But, Matt didn't want the other children to see how he felt about the
joke, because he didn't want them to call him a baby. So, Matt tried to
hide how he felt on the inside.”
With the researcher pointing to the three emotion images, children
were then asked, “So, how did Matt really feel, on the inside, when
everyone laughed? Did he feel happy, sad, or okay? (target-feel question)
How did Matt try to look on his face, when everyone laughed? Did he
look happy, sad, or okay?” (target-look question). Responses were coded
such that the target-look answer had to indicate a more positive
response than the target-feel question in order to earn a score of 1 (i.e.,
Matt would look okay but feel sad, or Matt would look happy but feel
okay, or Matt would look happy and feel sad).
Procedures
Each child was interviewed for 10–15 min. Task order was ran-
domized for each child. Children were interviewed by trained under-
graduate research assistants. Interviews were conducted one-on-one at
the child's school, in a quiet area selected by the child's primary teacher
(e.g., empty classroom). All interviews were audio recorded. Trained
research assistants independently transcribed and coded 20% of the
cases (N =26), and attained perfect inter-rater reliability (100%
agreement).
Results
Preliminary analyses
Correlation analysis was used to evaluate the relationships between
the demographic factors and ToM tasks (see Table 3). Point-biserial
correlation was carried out to examine the bivariate relations between
continuous variables (age, family size, family income, nancial strain)
and the ve ToM variables. As Maternal and Paternal education level
and Job Zone were all ordinal variables, Kendall's tau-b correlation co-
efcients were obtained between these three ordinal variables and the
ve ToM variables. Four children were excluded from the analysis due to
outliers in family income. Of the family variables included, only Job
Zone was signicantly positively related to knowledge access (
τ
b
=
0.254, p =.028), and negatively related to hidden emotion (
τ
b
=
−0.311, p =.010).
Developmental sequence of ToM tasks in low-income sample
To test the rst hypothesis that the developmental sequence of ToM
tasks will be unique, given the environmental necessity of specic
perspective taking aspects, we conducted a Rasch Model (1PL), which is
an item response theory model (Wellman & Liu, 2004). As a one-
parameter logistic model for dichotomous items, the Rasch model esti-
mates respondent's ability and task difculty levels based on their re-
sponses to test items (Rasch, 1960; Wright & Masters, 1982). Thus, when
the task's difculty level is lower than the person's ability level, the
estimated probability of the person passing the task is greater than 0.5,
whereas if the task's difculty level is higher than the person's ability
level, the estimated probability of the person passing the task is less than
0.5, and when the task's difculty level equals the person's ability, the
estimated probability of the person passing the task is equal to 0.5.
The Rasch model analysis for the ve ToM tasks were analyzed using
the mirt package (Chalmers, 2012) through R (R Core Team, 2017). In
order to check that the ve tasks measured one latent trait (i.e., ToM) in
the Rasch model, parallel analysis (Lance, Butts, & Michels, 2006) and
factor analysis were conducted. Both analyses identied only one factor
(conrmed by scree plot; Cattell, 1966), indicating that the data is
unidimensional. Then, we assessed the model t overall; the value of
model t statistic M2 was not signicant (p =.161) and the root mean
square error of approximation (RMSEA =0.048) was smaller than 0.06
(Hu & Bentler, 1999; Jackson, Gillaspy, & Purc-Stephenson, 2009). Both
of these indicators suggest the Rasch model is a good t to the data.
Table 4 presents the item (task) difculty levels and item t statistics
for the ve ToM tasks from least difcult to most difcult. Following
Duh et al. (2016), we chose two types of item t statistics for each item:
int and outt. Both of these indicate how accurately an item ts the
model. The int is more sensitive to the patterns of responses to items,
while the outt is more sensitive to responses to items with difculty far
from a person's ability (Linacre, 2003; Wellman & Liu, 2004). Values
between 0.75 and 1.33 are considered reasonable in both int and outt
statistics, and therefore indicate acceptable t (Adams & Khoo, 1996).
Table 3
Correlations between demographic variables and ToM tasks.
Age (months) Family size Family income Financial Strain Mat. Ed. Pat. Ed Job Zone
DD 0.232* 0.134 0.032 0.095 0.134 0.007 −0.091
DB 0.259* −0.113 0.160 0.015 0.194 0.028 −0.060
KA 0.339* 0.120 −0.017 −0.127 0.064 0.013 0.254*
FB 0.193 0.075 −0.046 0.023 0.019 0.005 0.010
HE −0.105 0.005 −0.103 0.020 0.041 0.078 −0.311*
Note. DD =Diverse Desires. DB =Diverse Beliefs. FB =False Belief. KA =Knowledge Access. HE =Hidden Emotion. Mat. Ed. =Maternal Education. Pat. Ed. =Paternal
Education.
*
<0.05.
Table 4
Item measure summary and t statistics for ve-item Rasch model.
ToM tasks Difculty SE Int MS Outt MS
Diverse Desires (DD) −1.12 0.32 0.95 0.99
Diverse Beliefs (DB) −0.04 0.28 0.86 0.80
Explicit False Belief (FB) 0.49 0.28 0.95 0.94
Knowledge Access (KA) 0.67 0.28 0.83 0.76
Hidden Emotion (HE) 0.78 0.29 0.95 0.88
E.R. Baker et al.
Journal of Applied Developmental Psychology 76 (2021) 101314
6
Both int and outt conrmed that the ve items t the Rasch model
well.
These values revealed that the order of item difculty in the current
sample did not mirror the ToM sequence from the typically developing
children found in Wellman and Liu (2004), or those reported by others
(Kuntoro et al., 2013; Shahaeian, 2015). In the current sample, the least
difcult task was the diverse desires (DD) task, followed by the diverse
beliefs (DB), which is consistent with Wellman and Liu (2004). How-
ever, in the current sample, children then found explicit false belief (FB)
to be the next easiest, followed by knowledge access (KA). The hidden
emotion (HE) task was the most difcult task. In other words, based on
these ndings, children who are living in low-income families develop
false belief understanding (FB) before knowledge access (KA). In addi-
tion, by examining the differences of difculty levels between the
adjacent tasks, diverse desires (DD) and diverse beliefs (DB) is much
easier than the other three tasks (e.g., the difference between diverse
beliefs (DB) and false belief understanding (FB) is 0.71), while the dif-
culty of knowledge access (KA) task approximates the difculty of the
hidden emotion (HE) task, with a difference of 0.11.
We then constructed the Item Characteristics Curves (ICC) for the
ve tasks (see Fig. 1). The ICC for each task is a S-shaped curve which
shows the probability of passing a task based on the respondent's ability
(Henard, 2000). For interpretability, item difculty is dened as the
point at which a person has a 50% probability of passing the task suc-
cessfully (McCamey, 2014). Therefore, along the X axis, the items are
presented from least difcult to most difcult. The ICC conrmed that
children in the current sample passed the ToM tasks in the following
order: diverse desires (DD), diverse beliefs (DB), false belief under-
standing (FB), knowledge access (KA), and hidden emotion (HE), which
differs from the developmental sequence for typical middle-income
preschoolers in the U.S. which found in Wellman and Liu (2004), as
well as the developmental sequence reported for economically-
impoverished non-U.S. samples (e.g., Shahaeian, 2015).
Task performance across samples
Next, we compared passage rates of ToM tasks in the current sample
with those of six other samples from published studies, all of which used
the ve-task battery from Wellman and Liu (2004). These comparison
samples were categorized into three groups, all of which included
typically developing children. First, children from middle-or-upper class
families in individualistic cultures including the U.S. children from the
Wellman and Liu (2004) study, and the Australian children from Sha-
haeian et al. (2011). The second group was afuent children from
collectivist cultures: Iranian children from urban, high SES families in
the Shahaeian (2015) study, and Chinese children from urban middle-
or-upper income families in Wellman et al. (2006). The last group
included two low-income samples from collectivistic cultures: one was
Indonesian children from pemulung families in the Kuntoro et al. (2013)
study, and the other was an Iranian sample from urban, low-income
families from the Shahaeian (2015) study.
The goal here was to test the second hypothesis, which focused on
the developmental timing of conceptual mastery of diverse desires (DD),
diverse beliefs (DB), and knowledge access (KA) as compared to children
from other environments. Additionally, this tested our third hypothesis,
which specied that children from the current sample would show equal
or advanced hidden emotion (HE) conceptual understanding, as
compared to others.
The comparisons of passing rates of ToM tasks were done using The
Signicance of the Difference Between Two Independent Proportions, z-
ratio test application (Lowry, 2016). Table 5 provides the basic de-
mographic information of each sample and the difference in percentage
of children who passed each task across all samples. Comparisons were
made between the current sample and samples from each of the previous
studies for each ToM task.
Findings revealed a number of signicant differences, as indicated in
Table 5. Before considering direct comparisons of passage rates, it is
important to consider that – as ToM changes rapidly in the preschool
years – signicant age differences between samples would likely have
meaningful impacts on ToM differences. For this reason, it is important
to note that children's ages across the six samples were not equivalent;
specically, children in the current sample were signicantly younger
than children in the Kuntoro et al. (2013) pemulung group, but did not
differ from any other sample. This age difference should be considered
during interpretations.
Fig. 1. The item characteristic curves for the ve ToM tasks.
Note. The x axis (“Ability”) has been standardized and “0” indicated the average performance in the current sample.
E.R. Baker et al.
Journal of Applied Developmental Psychology 76 (2021) 101314
7
Diverse desires
Children in the current sample passed the diverse desires (DD) task at
a lower rate compared with all other samples, indicating that they may
develop this conceptual understanding at an older age compared to
more afuent children from individualistic communities (i.e., U.S. and
Australia) and collectivistic communities (i.e., Iran and China), and
compared to low-income children reared in Indonesia and Iran (see
Table 5).
Diverse beliefs
Findings reveal that children in our sample (low-income U.S.) passed
the diverse beliefs (DB) task at a rate that was signicantly lower than all
other relatively-afuent individualistic and collectivistic samples (Sha-
haeian et al., 2011; Wellman et al., 2006; Wellman & Liu, 2004), and the
low-income Indonesian sample (Kuntoro et al., 2013). However, they
performed equivalently to the low-income Iranian sample (Shahaeian,
2015), as shown in Table 5. Given that the low-income Indonesian
sample was considerably older, this indicates that children in the current
sample may master this ToM aspect at a later age, as environmentally
appropriate, as compared with more afuent, same-age peers in indi-
vidualistic communities, but that this (DB) may develop at an earlier age
for low-income U.S. children as compared to children in low-income
collectivistic children (i.e., Shahaeian, 2015).
False belief
We next turn to discuss ndings related to false belief understanding
(FB), which – as discussed previously – developed in a different sequence
in the current sample (i.e., sequentially earlier) as compared to more
afuent children in both individualistic and collectivistic communities
and low-income Iranian children, but not low-income Indonesian chil-
dren. With regard to passage rates, children in the current sample did
not differ from either of the other low-income samples and more afuent
samples from collectivistic culture, but did pass this task at signicantly
lower rates as compared to more afuent U.S. children, yet at a signif-
icantly higher rate than afuent Australian children (see Table 5).
Knowledge access
Similar to the sequential differences for false belief understanding
(FB), it is important to recall that conceptual mastery of knowledge
access (KA) occurred in a different sequence (i.e., after false belief) as
compared to afuent samples of individualistic countries (Shahaeian
et al., 2011) and collectivistic countries (Shahaeian, 2015; Wellman
et al., 2006) and low-income Iranian children (Shahaeian, 2015), but not
low-income Indonesian children (Kuntoro et al., 2013). Speaking to
passage rates across the samples, children in our sample did not differ
from the low-income Indonesian children (who, again, were consider-
ably older), but were quite different from all the other samples (see
Table 5). Indeed, children from afuent U.S. and Australian families,
afuent Chinese and Iranian families, and low-income Iranian samples –
who did not differ from the current sample in age – passed the knowl-
edge access (KA) task at more than double the rate of the low-income U.
S. children in the current study. This suggests that children from more
afuent families and low-income Iranian communities may have greater
environmental and social interactions that facilitate or require recog-
nizing that beliefs come from having direct exposure to information.
Hidden emotion
Children in our sample (low-income U.S.) passed the hidden emo-
tions (HE) task at a rate that was signicantly higher than the low-
income Indonesian sample (Kuntoro et al., 2013) and the afuent
Australian sample (Shahaeian et al., 2011), as indicated in Table 5.
Given that the low-income Indonesian sample was older, children in the
current sample may master this ToM aspect earlier, as their social
ecology might require, in comparison with the older low-income chil-
dren in Indonesia. Importantly, the low-income U.S. children did not
differ from the more afuent U.S., Chinese, and Iranian samples, or the
low-income Iranian sample in terms of hidden emotion (HE) passage
rates, indicating that perceptions of delayed ToM development thought
to occur due to poverty exposure may not hold true for all ToM aspects.
Discussion
Conceptually, ToM focuses on our ability to engage in complex social
interactions and so it follows that this conceptual mastery would be
informed by culturally- and socially-salient values and mores trans-
mitted through dynamic social interactions. As children reared in
poverty are exposed to unique environmental features which may not
occur for children at all income levels, our goal was to determine if
economically-impoverished U.S. children develop ToM in a way that
differed from more afuent U.S. samples, given that cross-cultural work
has found similar differences. This study is critical as it follows a call for
more studies examining the variability in ToM development as related to
differences in culture and social backgrounds (Shahaeian, 2015), and
demonstrates that even within individualistic cultures, the development
Table 5
Basic sample information and the percentages of children that passed the ve ToM tasks amongst the current low-income sample, as compared with two middle-class
individualistic samples, two middle-class collectivistic samples, and two low-income collectivistic samples.
Current Sample Wellman and Liu
(2004)
Shahaeian et al.
(2011)
Kuntoro et al.
(2013)
Shahaeian
(2015)
Shahaeian
(2015)
Wellman et al.
(2006)
Sample
Characteristics
Source U.S. U.S Australia Indonesia Iran Iran China
Culture individualist individualist individualist collectivist collectivist collectivist collectivist
SES low income – middle class low income low income high SES middle class
Mean Age (SD) 52.77 (6.61) 54.33 (−) 54.6 (−) 66.5
a
(16.89) 59.1 (5.83) 56.9(5.53) 54.82(−)
Age Range [37–64] [35–78] [36–77] [37–94] – [36–78] [35–73]
Gender (%
male)
51% 56% 47% 52% 49% 52% 53%
Avg. Parent
Education
Attended high
school
– Completed
university
Completed
elementary school
Completed high
school
Completed
university
Completed
university
Task Performance DD 77.3
a,c
95
a
95
a
85
c
89
c
94
a
89
c
DB 58.7
a
84
a
77
a
85
a
51 70 71
FB 42.7
b
57
b
36 40 49 46 49
KA 37.3
a
73
a
68
a
33 86
a
91
a
79
a
HE 34.7
a
32 16
a
12
a
32 27 37
Note. Dashes indicate that the information was not reported in the previous study. Superscripts refer to comparison with the current sample, as indicated by the test of
signicance of the difference between two independent proportions. Age is reported in months. DD =Diverse Desires. DB =Diverse Beliefs. FB =False Belief. KA =
Knowledge Access. HE =Hidden Emotion.
a
<0.001.
b
<0.01.
c
<0.05.
E.R. Baker et al.
Journal of Applied Developmental Psychology 76 (2021) 101314
8
of the order of ToM is not preordained.
First, children in the current study showed some similarities with
studies of more afuent children across cultures, and children from low-
income collectivistic families. Children here performed several tasks
(DD, DB) in the same way found in both individualistic and collectivistic
cultures across income levels (Kuntoro et al., 2013; Shahaeian, 2015;
Wellman et al., 2006; Wellman & Liu, 2004). This pattern (DD >DB)
was found for all samples, including ours. Moreover, FB also developed
sequentially after DB and DD for all samples. It seems that one universal
ToM aspect may be that children develop the understanding that
different people can hold different wants earlier than understanding that
different people can have different beliefs, and then that beliefs may be
incongruent with reality.
However, it is the differences revealed here that expand our
knowledge about ToM. Beyond the tasks that followed the same pro-
gression (i.e., DD >DB >FB), children in the current sample showed a
different progression from more afuent samples in individualistic cul-
tures, but tended to follow the progression found for impoverished
children from collectivistic cultures (i.e., impoverished children from
Indonesia; Kuntoro et al., 2013). Children in the current study developed
mastery of the recognition that knowing comes from exposure after
mastery of the recognition that beliefs are distinct from reality and that a
change in reality may not correspond with a change in belief (FB >KA).
This could indicate that children in this sample may have less of an
environmental emphasis on recognizing whether others have seen the
same things or been exposed to the same information they have, or that
perception leads to knowing. For example, when playing a game, they
may not realize that another child did not know the rules because that
child was not present when the rules were made or stated.
Alternatively, Kuntoro et al. (2013) proposed that knowledge access
mastery differed across cultures and groups due to differences in values.
For instance, collectivistic cultures place importance on knowledge
acquisition and education (i.e., China; Wellman et al., 2006) while
individualistic cultures tend to place greater value on independence and
individual beliefs and opinions (i.e., U.S.). Children in collectivistic
cultures are reared to consider and plan for advanced education quite
early in life. The rst step towards that goal is to acquire knowledge,
which can only be done if one recognizes the importance of access to
knowledge, and should then correspond with children developing the
knowledge access component earlier (Yang & Shin, 2008). That is,
children reared in poverty may have less awareness of how people come
to know things, perhaps due to less formal education opportunities
(Kuntoro et al., 2013).
In contrast, children from individualistic cultures tend to develop the
conceptual understanding of diversity of opinions and beliefs rather
early during the childhood years, as compared with children from
collectivistic cultures that value group harmony (Shahaeian et al., 2011;
Wellman et al., 2006). Our ndings, however, showed that KA devel-
oped later than FB, which may indicate that low-income families in the
U.S. communicate less directly about information and knowledge.
Children's development of knowledge access occurring later may be an
artifact of parental employment reported in the current sample. Parents
with occupations that do not require any training, for instance, may not
communicate with their children about how knowledge and skill
development might serve them in obtaining their desired occupation (e.
g., “if you want to be a teacher, you need to....”). This could correspond
with a broader pattern indicating that low-income parents may not
understand the precise behaviors or mechanisms that lead to high
educational attainment unless they themselves have attained a similar
or higher level (Robinson, 2017).
A different possible explanation for the pattern of knowledge access
mastery may be that low-income parents feel uninformed about how to
impact their child's occupational accomplishments. Parents' attitudes
about knowledge and wisdom have been found to directly and positively
impact parental involvement with their child's educational outcomes,
and perhaps occupational outcomes, both at school and at home
(Puccioni, Froiland, & Moeyaert, 2020). Parents who prioritize and
value knowledge and information, even at the preschool level, are more
likely to work to prepare their child for academic settings and skills, and
– as related to the current ndings – these practices also correspond with
children's socio-emotional capacities (Puccioni, Baker, & Froiland,
2020). For example, these parents may spend time reading with their
children or sing songs about numbers and colors. Behaviorally, when
these parents are engaging in these practices with their children, they
are also transmitting and modeling important social and cognitive ca-
pacities (Gutman & McLoyd, 2000; Kohm, Holmes, Romeo, & Koolidge,
2016), such as ToM. Importantly, low-income parents may be less likely
to engage in numeracy- and literacy-related activities (Phillips, 2011;
Vandermaas-Peeler, Nelson, Bumpass, & Sassine, 2009), not for lack of
interest but for lack of access, due to multiple competing demands for
time. Low-income parents are often forced to work multiple jobs, work
evenings and weekends, have jobs with limited paid-leave, etc., in order
to meet the nancial needs of their family. Therefore, government pol-
icies that serve to support low-income parents in having greater access
to their children's academic preparedness (i.e., mandated paid leave, to
allow parents to attend to their children's needs with greater availabil-
ity), may be one application of this work.
Overall, previous research using samples from individualistic cul-
tures have demonstrated that children from low-income families pass
false belief tasks later than their more afuent peers (Holmes et al.,
1996; Holmes-Lonergan, 2003; Seidenfeld et al., 2014), which was also
found here. Moreover, although the order of diverse desires and diverse
beliefs seems universal across environments, the rates at which children
in this study passed these tasks (DD, DB) indicates that low-income
children may arrive at a developmental niche for recognizing diversity
of beliefs later than their afuent counterparts. This has been explained
previously by culturally salient norms of parenting (Shahaeian, 2015;
Shahaeian et al., 2011). For instance, parents who are more likely to
intervene in or prevent children's disagreements may have children who
arrive later to conceptually understanding of belief diversity, as these
children are thought to have fewer natural opportunities to consider
diverging beliefs. Shahaeian, Nielson, et al. (2015) noted that Iranian
parents who communicated less frequently and less directly with their
children tended to have children who performed more poorly on the
ToM battery. Additionally, these parents indicated that they would
avoid correcting their child's misbehavior, and would avoid talking to
the child about the consequences of such behavior, which Shahaeian,
Nielson, et al. (2015) suggested may have given children fewer oppor-
tunities to recognize that beliefs may differ amongst individuals during
times of disagreement. That is, children were not guided in identifying
disparate beliefs and opinions. In contrast, parents who correct their
children's behavior by reasoning with the child and explaining why a
particular action is harmful to others has been found to positively pre-
dict children's understanding of divergent beliefs (Shahaeian, Nielson,
et al., 2015), as the parents would oftentimes make direct reference to
the victims' belief state (e.g., “I told him that he shouldn't touch others'
things without them knowing”).
Differences in parent intervention and correction misbehavior also
aligns with research describing how U.S. parents who assert their power
as an authority gure (described as behaviors characterized by “physi-
cally and verbally negative responses such as spanking and yelling”, akin
to the authoritarian parenting style) tend to have children who per-
formed more poorly on belief tasks (Pears & Moses, 2003). Interestingly,
these children also performed better on tasks of divergent emotions – in
other words, parenting behaviors characterized by harsh responses to
children's misbehavior may impede children's understanding of di-
versity of beliefs, but may promote an understanding of the complexity
of human emotion. Parenting approaches focusing on power assertion
also demonstrate high levels of negative affect as a part of this approach.
For example, yelling and spanking is often accompanied by anger or
intense frustration (Pears & Moses, 2003), and so events of this nature
likely allow children to develop a keen sense of emotional awareness
E.R. Baker et al.
Journal of Applied Developmental Psychology 76 (2021) 101314
9
(Wass et al., 2019).
The children in the current sample also showed somewhat preco-
cious emotional perspective taking capacities, especially when consid-
ered in conjunction with the performance found on other tasks
compared with other samples. It may be that nancial hardship and
adversity creates a parenting environment focused more on day-to-day
maintenance demands, such as working several jobs in order to nan-
cially provide necessities (as evidenced here by parents' reports of sub-
jective nancial well-being), yet allows little time for lengthy inductive
discussions related to other's beliefs. It is possible that inductive
parenting strategies, which require more one-on-one interaction time
with the parent, help to develop children's appreciation for diversity of
belief yet do not contribute to children's appreciation for emotional
expression.
Overall, children in the current sample most resembled children from
the low-income Indonesian sample (i.e., Kuntoro et al., 2013) in terms of
ToM progression and emergence. Importantly, children from these two
studies may share some environmental similarities not considered in
either study, which could have been impactful for their ToM above and
beyond poverty. For instance, both groups are from urban areas and
their parents have only basic literacy prociency, and so it is possible
that these two samples (ours, and Kuntoro et al., 2013) performed
similarly because of some unique factor of urban poverty, rather than
rural poverty, or because of the parents' literacy rates. However, it is
important to remember that children within our sample were younger,
receiving more early prevention/intervention through Head Start, and
may have been experiencing more parental involvement than children
from low income families generally. Therefore, although striking par-
allels were found, it is possible that these parallels do not rest upon
socio-economic similarities.
Implications
The ndings of the current study offer a number of important im-
plications and applications for educators and researchers alike. First, it is
important to recognize the impacts of the ecological niche on develop-
ment (Frankenhuis & Nettle, 2020). As has been demonstrated across
several domains of development, certain faculties seem to emerge earlier
in some populations compared to others, which is thought to occur as a
way of development meeting the demands of the environment (Frank-
enhuis & de Weerth, 2013; Wass et al., 2019). That is, the differences we
see across populations in the emergence of certain capacities should be
viewed as an adaptive environmental sensitivity; this is a rather recent
perspective, as decades of previous research on impoverished and
otherwise marginalized communities interpreted differences as decits
(see Frankenhuis & de Weerth, 2013 for review). Therefore, the impli-
cations and interpretations of ndings here should be viewed as evi-
dence that children's capacities are adaptive, indeed responsive, to their
environment. More precisely, we propose that children in the current
sample develop certain ToM capacities at an older age because such
capacities are less important for their daily lives until later in
development.
From this vantage point, the current ndings suggest that children in
impoverished communities arrive at three general ToM competencies at
about the same time (KA, FBU, and HE) – a stark difference from afuent
children. This could indicate that – for children in this sample – a sudden
change in environment necessitates a relatively synchronous develop-
ment of these ToM capacities, rather than sequential development. For
instance, it could be that the start of formal education (i.e., preschool)
offers several new challenges that were not present prior to exposure to
education. Additionally, as these were the most difcult tasks, it is
possible that these capacities do develop sequentially, but at a later age
than studied here, which would imply that they do perhaps do not un-
dergo a robust change in a relatively short timeframe, but that this
timeframe is protracted beyond the preschool years.
Here, it is important to recall that the Head Start preschool education
program in the U.S. is an intervention designed and implemented by the
government to serve impoverished communities. Moreover, preschool
education in the U.S. is not mandatory, and so parents who utilize Head
Start are opting in and indeed must apply and can be put on a waitlist or
denied. This is all to say that parents of children who attend Head Start
programs are motivated to seek out and acquire this intervention for
their children. Put simply, these are dedicated and committed parents
who lack the economic means to educationally support their children,
but nevertheless recognize the importance of it.
The question remains, however, as to whether ToM differences
across populations offer meaningful differences for children's develop-
ment. Given the strengths-based approach utilized here, we offer that
these ToM differences are likely indicative of many other differences in
socio-cognitive and emotional competencies. Although all neurotypical
children will arrive at ToM competence, the pattern by which this occurs
indicates a sensitivity to the environment, which in turn reects
appropriate and healthy development. With this in mind, educators and
researchers would do well to consider what these patterns signal about
the environment and capitalize upon that strength.
For instance, as children here were more advanced in their emotional
perspective taking compared to some samples, it is possible that the
children here hold a precocious and robust emotional awareness. This
could be leveraged by educators in a number of ways, such as using
emotionally-salient narratives during numeracy- and literacy-related
activities (e.g., learning to read by identifying labels for emotional
terms like “sad” and “happy”). As schools increasingly recognize the
importance of social-emotional learning for academic achievement and
overall development, they are beginning to incorporate social-emotional
learning into their grade level expectations for school aged students
(Eklund, Kilpatrick, Haider, & Eckert, 2018). In order to fairly assess all
children in this area, we must have a foundation for what is typical
development for children from different backgrounds with research that
considers cultural and socioeconomic differences in addition to age.
Research identifying how ToM may develop differently for children
from a variety of backgrounds can inform why some young children may
not have certain ToM components in place yet. For example, a young
child from a low socioeconomic background may not demonstrate the
diverse beliefs component of ToM compared to their afuent peers.
Instead of this being considered as a lag or decit within the child, this
research suggests it may be more typical given the child's ecological
niche, and educators can work with the student to develop this skill
rather than penalizing the child or the parent.
Limitations and future directions
The current study conrmed ndings from previous studies showing
a different pattern of ToM development for low-income children, and
expanded on this by (1) considering the full range of ToM capacities, (2)
demonstrating that ToM progression occurs in a different order for
children from impoverished communities, and (3) that the low-income
difference in ToM is not true for all aspects of ToM. Although these
are novel and worthwhile contributions, the current study has limita-
tions that future studies should consider.
First, the adverse impact of abuse and neglect exposure on children's
ToM development was not examined in this study, nor were parenting
styles. This was a decision made by our community partners (i.e., the
administrators of the Head Start program). They reported being aware of
the high rates of abuse in the area from which our sample was drawn and
within the families of their center specically, which is consistent with
the extremely high violent crime rate in the area. They also noted that
contacting the ofce for Child Protective Services was not uncommon.
For this reason, they believed that attempting to assess abuse or neglect
by questioning parents or children could cause greater harm to the
children, and we agreed that including this assessment would be
disadvantageous to the families. Nevertheless, it is probable that at least
some of the sample was exposed in some way to abuse or neglect.
E.R. Baker et al.
Journal of Applied Developmental Psychology 76 (2021) 101314
10
Another interpretation of our ndings, with this in mind, would be to
disentangle how abuse and poverty might impact ToM.
Secondly, it is important to consider that slight variations in tasks
between studies might hold important implications for children's per-
formance. For instance, in the current study we used an adapted
appearance-reality task as a task for Diverse Belief; some studies
consider a version of this assessment a task of FBU, although not all do
(e.g., Warnell & Redcay, 2019). Although the task used here does
developmentally map onto the timing for the Diverse Belief task used by
Wellman and Liu (2004), it is possible that task differences between that
used here and used by Wellman and Liu (2004) were more meaningful
than we originally believed. Moreover, this also speaks to a possible area
of confusion within the ToM literature—the number of ToM tasks that
are used in current research is quite staggering, and may make it difcult
to draw parallels between studies. Therefore, one possible area of
importance for future studies would be to create a ToM task library, with
noted equivalencies and distinctions, in order to aid researchers in
developing greater uniformity across studies.
Conclusions
Low-income families experience stressors related to having limited
or insufcient resources and time (Blair et al., 2005), and children from
low-income families likely develop in ways that are sensitive and
responsive to such stressors. Overall, the ndings of this study lend
support for differential development of ToM in low-income families.
This study supports and expands on previous work that has reliably
shown that although ToM is a universal developmental phenomenon the
timing and development of ToM is a consequence of numerous envi-
ronmental factors. The environmental impact of poverty exposure seems
partially sufcient to alter the well-established sequence of ToM. As a
result, researchers and educators should consider how children's back-
grounds may play a role in their development.
Acknowledgements
The authors would like to acknowledge the Head Start programs of
the Schenectady Community Action Partnerships for their collaboration
on this study, as well as the children who participated, their caregivers,
and their teachers. We thank the following students for their help in data
collection: Michelle Meyer, Matilda Armstrong, Kristen Cocca, Kiana
Roundtree, Brandy Severino, Rose Alicia Oliveras, Stephanie Silva,
Aaron Striano. We also extend our thanks to the editor and the two re-
viewers for their time and contributions in strengthening this
manuscript.
References
Adams, R. J., & Khoo, S. T. (1996). Quest: The interactive test analysis system [computer
software]. Camberwell, Australia: Australian Council for Educational Research.
Anderson, R. E., Hussain, S. B., Wilson, M. N., Shaw, D. S., Dishion, T. J., & Williams, J. L.
(2015). Pathways to pain: Racial discrimination and relations between parental
functioning and child psychosocial well-being. Journal of Black Psychology, 41,
491–512.
Astington, J. W. (2001). The future of theory-of-mind research: Understanding
motivational states, the role of language, and real-world consequences. Child
Development, 72(3), 685–687. https://doi.org/10.1111/1467-8624.00305
Barbarin, O. A. (1993). Coping and resilience: Exploring the inner lives of African
American children. Journal of Black Psychology, 19(4), 478–492.
Baron-Cohen, S., Leslie, A. M., & Frith, U. (1985). Does the autistic child have a “theory
of mind”? Cognition, 21, 37–46.
Blair, C., Granger, D., & Razza, R. P. (2005). Cortisol reactivity is positively related to
executive function in preschool children attending head start. Child Development, 76
(3), 554–567. https://doi.org/10.1111/j.1467-8624.2005.00863.x
Blair, C., & Razza, R. P. (2007). Relating effortful control, executive function, and false
belief understanding to emerging math and literacy ability in kindergarten. Child
Development, 78(2), 647–663. https://doi.org/10.1111/j.1467-8624.2007.01019.x
Cattell, R. B. (1966). The scree test for the number of factors. Multivariate Behavioral
Research, 1, 245–276.
Cavadel, E. W., & Frye, D. A. (2017). Not just numeracy and literacy: Theory of mind
development and school readiness among low-income children. Developmental
Psychology, 53(12), 2290–2303. https://doi.org/10.1037/dev0000409
Chalmers, R. P. (2012). Mirt: A multidimensional item response theory package for the R
environment. Journal of Statistical Software, 48(6), 1–29. https://doi.org/10.18637/
jss.v048.i06
Cheung, C. (2010). Environmental and cognitive factors inuencing children’s theory-of-mind
development. Doctoral dissertation. University of Toronto. Retrieved from https://pd
fs.semanticscholar.org/a2a3/d96ceb11b62887ccabfd11835c4b1269ca51.pdf.
Conger, R. D., Ge, X., Elder, G. H., Jr., Lorenz, F. O., & Simons, R. L. (1994). Economic
stress, coercive family process, and developmental problems of adolescents. Child
Development, 65(2), 541–561. https://doi.org/10.1111/j.1467-8624.1994.tb00768.x
Core Team, R. (2017). R: A language and environment for statistical computing. Vienna,
Austria: R Foundation for Statistical Computing. Retrieved from https://www.R-pro
ject.org/.
Curenton, S. M. (2004). The association between narratives and theory of mind for low-
income preschoolers. Early Education and Development, 15(2), 124–146. https://doi.
org/10.1207/s15566935eed1502_1
Cutting, A. L., & Dunn, J. (1999). Theory of mind, emotion understanding, language, and
family background: Individual differences and interrelations. Child Development, 70
(4), 853–865. https://doi.org/10.1111/1467-8624.00061
Dearing, E. (2004). The developmental implications of restrictive and supportive
parenting across neighborhoods and ethnicities: Exceptions are the rule. Journal of
Applied Developmental Psychology, 25, 555–575. https://doi.org/10.1016/j.
appdev.2004.08.007
DeJoseph, M. L., Sifre, R. D., Raver, C. C., Blair, C. B., & Berry, D. (2021). Capturing
environmental dimensions of adversity and resources in the context of poverty across
infancy through early adolescence: A moderated nonlinear factor model. Child
Development.. https://doi.org/10.1111/cdev.13504
Devine, R. T., White, N., Ensor, R., & Hughes, C. (2016). Theory of mind in middle
childhood: Longitudinal associations with executive function and social competence.
Developmental Psychology, 52(5), 758–771. https://doi.org/10.1037/dev0000105
Dietze, P., & Knowles, E. D. (2020). Social class predicts emotion perception and
perspective-taking performance in adults. Personality and Social Psychology Bulletin..
https://doi.org/10.1177/0146167220914116
Duh, S., Paik, J. H., Miller, P. H., Gluck, S. C., Li, H., & Himelfarb, I. (2016). Theory of
mind and executive function in Chinese preschool children. Developmental
Psychology, 52(4), 582–591. https://doi.org/10.1037/a0040068
Duncan, G. J., & Brooks-Gunn, J. (2000). Family poverty, welfare reform, and child
development. Child Development, 71(1), 188–196. https://doi.org/10.1111/1467-
8624.00133
Edwards, A. L., & Few-Demo, A. L. (2016). African American maternal power and the
racial socialization of preschool children. Sex Roles, 75(1–2), 56–70.
Eklund, K., Kilpatrick, K. D., Kilgus, S. P., Haider, A., & Eckert, T. (2018). A systematic
review of state-level social-emotional learning standards: Implications for practice
and research. School Psychology Review, 47(3), 316–326. https://doi.org/10.17105/
SPR-2017.0116.V47-3
Flavell, J. H., Green, F. L., Flavell, E. R., & Watson, M. W. (1986). Development of
knowledge about the appearance-reality distinction. Monographs of the Society for
Research in Child Development, 51(1).
Frankenhuis, W. E., & de Weerth, C. (2013). Does early-life exposure to stress shape or
impair cognition? Current Directions in Psychological Science, 22(5), 407–412.
Frankenhuis, W. E., & Nettle, D. (2020). The strengths of people in poverty. Current
Directions in Psychological Science, 29(1), 16–21.
Gopnik, A., & Astington, J. W. (1988). Children’s understanding of representational
change and its relation to the understanding of false belief and the appearance-
reality distinction. Child Development, 59(1), 26–37. https://doi.org/10.2307/
1130386
Gutman, L. M., & McLoyd, V. C. (2000). Parents’ management of their children’s
education within the home, at school, and in the community: An examination of
African-American families living in poverty. The Urban Review, 32(1), 1–24.
Heleniak, C., & McLaughlin, K. A. (2019). Social-cognitive mechanisms in the cycle of
violence: Cognitive and affective theory of mind, and externalizing psychopathology
in children and adolescents. Development and Psychopathology, 1–16. https://doi.org/
10.1017/S0954579419000725
Henard, D. H. (2000). Item response theory. In L. G. Grimm, & P. R. Yarnold (Eds.),
Reading and understanding MORE multivariate statistics (pp. 67–97). American
Psychological Association.
Herruzo, C., Raya Trenas, A., Pino, M. J., & Herruzo, J. (2020). Study of the differential
consequences of neglect and poverty on adaptive and maladaptive behavior in
children. International Journal of Environmental Research and Public Health, 17(3),
739.
Heyes, C. (2003). Four routes of cognitive evolution. Psychological Review, 110(4),
713–727. https://doi.org/10.1037/0033-295X.110.4.713
Holmes, H. A., Black, C., & Miller, S. A. (1996). A cross-task comparison of false belief
understanding in a head start population. Journal of Experimental Child Psychology,
63(2), 263–285. https://doi.org/10.1006/jecp.1996.0050
Holmes-Lonergan, H. A. (2003). Understanding of affective false beliefs, perceptions of
parental discipline, and classroom behavior in children from head start. Early
Education and Development, 14(1), 29–46. https://doi.org/10.1207/
s15566935eed1401_3
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for t indexes in covariance structure
analysis: Conventional criteria versus new alternatives. Structural Equation Modeling:
A Multidisciplinary Journal, 6(1), 1–55. https://doi.org/10.1080/
10705519909540118
E.R. Baker et al.
Journal of Applied Developmental Psychology 76 (2021) 101314
11
Jackson, D. L., Gillaspy, J. A., & Purc-Stephenson, R. (2009). Reporting practices in
conrmatory factor analysis: An overview and some recommendations. Psychological
Methods, 14(1), 6–23. https://doi.org/10.1037/a0014694
Kohm, K. E., Holmes, R. M., Romeo, L., & Koolidge, L. (2016). The connection between
shared storybook readings, children’s imagination, social interactions, affect,
prosocial behavior, and social play. International Journal of Play, 5, 128–140. https://
doi.org/10.1080/21594937.2016.1203895
Kristen, S., Thoermer, C., Hofer, T., Aschersleben, G., & Sodian, B. (2006). Validation of
the theory of mind scale. Zeitschrift für Entwicklungspsychologie und P¨
adagogische
Psychologie, 38, 186–195.
Kuntoro, I. A., Saraswati, L., Peterson, C., & Slaughter, V. (2013). Micro-cultural
inuences on theory of mind development: A comparative study of middle-class and
pemulung children in Jakarta, Indonesia. International Journal of Behavioral
Development, 37(3), 266–273. https://doi.org/10.1177/0165025413478258
Lance, C. E., Butts, M. M., & Michels, L. C. (2006). The sources of four commonly
reported cutoff criteria what did they really say? Organizational Research Methods, 9,
202–220.
Linacre, J. M. (2003). Size vs. signicance: Standardized chi-square t statistics. Rasch
Measurement Transactions, 17(1), 918.
Longo, F., McPherran Lombardi, C., & Dearing, E. (2017). Family investments in low-
income children’s achievement and socioemotional functioning. Developmental
Psychology, 53(12), 2273–2289. https://doi.org/10.1037/dev0000366
Lowry, R. (2016). The signicance of the difference between two independent proportions.
Retrieved from http://vassarstats.net/propdiff_ind.html.
McCamey, R. (2014). A primer on the one-parameter Rasch model. American Journal of
Economics and Business Administration, 6(4), 159–163. https://doi.org/10.3844/
ajebasp.2014.159.163
McLoyd, V. C. (1990). The impact of economic hardship on black families and children:
Psychological distress, parenting, and socioemotional development. Child
Development, 61(2), 311–346. https://doi.org/10.1111/j.1467-8624.1990.tb02781.x
Nazarov, A., Frewen, P., Parlar, M., Oremus, C., MacQueen, G., McKinnon, M., &
Lanius, R. (2014). Theory of mind performance in women with posttraumatic stress
disorder related to childhood abuse. Acta Psychiatrica Scandinavica, 129(3), 193–201.
https://doi.org/10.1111/acps.12142
Nzinga, K., Rapp, D. N., Leatherwood, C., Easterday, M., Rogers, L. O., Gallagher, N., &
Medin, D. L. (2018). Should social scientists be distanced from or engaged with the
people they study? Proceedings of the National Academy of Sciences, USA, 45,
11435–11441. https://doi.org/10.1073/pnas.1721167115
Pears, K. C., & Moses, L. J. (2003). Demographics, parenting, and theory of mind in
preschool children. Social Development, 12(1). https://doi.org/10.1111/1467-
9507.00219
Peterson, C., Wellman, H., & Liu, D. (2005). Steps in theory-of-mind development for
children with deafness or autism. Child Development, 76(2), 502–517. https://doi.
org/10.1111/j.1467-8624.2005.00859.x
Peterson, C., & Wellman, H. M. (2009). From fancy to reason: Scaling deaf and hearing
children’s understanding of theory of mind and pretence. British Journal of
Developmental Psychology, 27(2), 297–310. https://doi.org/10.1348/
026151008X299728
Peterson, N. G., Mumford, M. D., Borman, W. C., Jeanneret, P. R., Fleishman, E. A.,
Levin, K. Y., … Gowing, M. K. (2001). Understanding work using the occupational
information network (O* NET): Implications for practice and research. Personnel
Psychology, 54(2), 451–492. https://doi.org/10.1111/j.1744-6570.2001.tb00100.x
Petterson, S. M., & Albers, A. B. (2001). Effects of poverty and maternal depression on
early child development. Child Development, 72(6), 1794–1813. https://doi.org/
10.1111/1467-8624.00379
Phillips, M. (2011). Parenting, time use, and disparities in academic outcomes. In
G. Duncan, & R. Murnane (Eds.), Whither opportunity? (pp. 207–228). New York:
Russell Sage Foundation.
Pillow, B. H. (1989). Early understanding of perception as a source of knowledge. Journal
of Experimental Child Psychology, 47, 116–129.
Puccioni, J., Baker, E. R., & Froiland, J. M. (2020). Academic socialization and the
transition to kindergarten: Parental beliefs about school readiness and involvement.
Infant and Child Development, 28(6).
Puccioni, J., Froiland, J. M., & Moeyaert, M. (2020). Preschool Teachers’ transition
practices and Parents’ perceptions as predictors of involvement and Children’s
school readiness. Children and Youth Services Review, 104742.
Rasch, G. (1960). Probabilistic models for some intelligence and attainment tests. Chicago:
University of Chicago Press.
Razza, R. A., & Blair, C. (2009). Associations among false-belief understanding, executive
function, and social competence: A longitudinal analysis. Journal of Applied
Developmental Psychology, 30(3), 332–343. https://doi.org/10.1016/j.
appdev.2008.12.020
Roberts, S. O., Bareket-Shavit, C., Dollins, F. A., Goldie, P. D., & Mortenson, E. (2020).
Racial inequality in psychological research: Trends of the past and recommendations
for the future. Perspectives on Psychological Science, 1–15. https://doi.org/10.1177/
1745691620927709
Robinson, D. W. (2017). Collaborative partnerships between high poverty and minority
parents and educational leaders: Reversing the school and home divide. Journal for
Multicultural Education, 11(1), 2–18. https://doi.org/10.1108/JME-11-2015-0035
Seidenfeld, A. M., Johnson, S. R., Cavadel, E. M., & Izard, C. E. (2014). Theory of mind
predicts emotion knowledge development in head start children. Early Education and
Development, 25(7), 933–948. https://doi.org/10.1080/10409289.2014.883587
Sexton, J. D., & Pennebaker, J. W. (2004). Stigma and nonexpression of emotion and self.
In I. Nyklicek, & A. Vingerhoets (Eds.), Emotional expression and health (pp.
321–333). New York: Brunner-Routledge.
Shahaeian, A. (2015). Sibling, family, and social inuences on children’s theory of mind
understanding: New evidence from diverse intra-cultural samples. Journal of Cross-
Cultural Psychology, 46(6), 805–820. https://doi.org/10.1177/0022022115583897
Shahaeian, A., Henry, J. D., Razmjoee, M., Teymoori, A., & Wang, C. (2015). Towards a
better understanding of the relationship between executive control and theory of
mind: An intra-cultural comparison of three diverse samples. Developmental Science,
18(5), 671–685. https://doi.org/10.1111/desc.12243
Shahaeian, A., Nielson, M., Peterson, C. C., Aboutalebi, M., & Slaughter, V. (2015).
Knowledge and belief understanding among Iranian and Australian preschool
children. Journal of Cross-Cultural Psychology, 45(10), 1643–1654. https://doi.org/
10.1177/0022022114548484
Shahaeian, A., Peterson, C., Slaughter, V., & Wellman, H. (2011). Culture and the
sequence of steps in theory of mind development. Developmental Psychology, 47(5),
1239–1247. https://doi.org/10.1037/a0023899
Smogorzewska, J., Szumski, G., & Grygiel, P. (2018). Same or different? Theory of mind
among children with and without disabilities. PLoS One, 13(10), 1–20. https://doi.
org/10.1371/journal.pone.0202553
Turner, B. (2018). Raising awareness of reective functioning among African American single
mothers. Doctoral dissertation. Fielding Graduate University. Retrieved from https
://search.proquest.com/docview/2022436679?pq-origsite=gscholar.
UNICEF. (). Child poverty. Retrieved from https://www.unicef.org/social-policy/childpo
verty. In Income and poverty in the United States: Current population reports. U.S.
Census Bureau. Retrieved from https://www.census.gov/content/dam/Census/lib
rary/publications/2019/demo/p60-266.pdf.
U.S. Census Bureau. (2018). Income and poverty in the United States: Current population
reports. Retrieved from https://www.census.gov/content/dam/Census/library/publi
cations/2019/demo/p60-266.pdf.
Vandermaas-Peeler, M., Nelson, J., Bumpass, C., & Sassine, B. (2009). Social contexts of
development: Parent-child interactions during reading and play. Journal of Early
Childhood Research Literacy, 9(3), 295–317.
Warnell, K. R., & Redcay, E. (2019). Minimal coherence among varied theory of mind
measures in childhood and adolescence. Cognition, 191. https://doi.org/10.1016/j.
cognition.2019.06.009
Wass, S. V., Smith, C. G., Daubney, K. R., Suata, Z. M., Clackson, K., Begum, A., &
Mirza, F. U. (2019). Inuences of environmental stressors on autonomic function in
12-month-old infants: Understanding early common pathways to atypical emotion
regulation and cognitive performance. Journal of Child Psychology and Psychiatry, 60,
1323–1333. https://doi.org/10.1111/jcpp.13084
Wellman, H. M., Fang, F., Liu, D., Zhu, L., & Liu, G. (2006). Scaling of theory of mind
understanding in Chinese children. Psychological Science, 17(12), 1075–1081.
https://doi.org/10.1111/j.1467-9280.2006.01830.x
Wellman, H. M., Fang, F., & Peterson, C. C. (2011). Sequential progressions in a theory-
of-mind scale: Longitudinal perspectives. Child Development, 82(3), 780–792.
https://doi.org/10.1111/j.1467-8624.2011.01583.x
Wellman, H. M., & Liu, D. (2004). Scaling of theory of mind tasks. Child Development, 75
(2), 523–541. https://doi.org/10.1111/j.1467-8624.2004.00691.x
Wellman, H. M., & Woolley, J. D. (1990). From simple desires to ordinary beliefs: The
early development of everyday psychology. Cognition, 35, 245–275.
Wiesmann, C. G., Schreiber, J., Singer, T., Steinbeis, N., & Friederici, A. D. (2017). White
matter maturation is associated with the emergence of theory of mind in early
childhood. Nature Communications, 8(1), 1–10.
Wright, B. D., & Masters, G. N. (1982). Rating scale analysis: Rasch measurement. Chicago,
IL: MESA Press.
Yang, S., & Shin, C. S. (2008). Parental attitudes towards education: What matters for
children’s well-being? Children and Youth Services Review, 30(11), 1328–1335.
Yeung, W. J., Linver, M. R., & Brooks-Gunn, J. (2002). How money matters for young
children’s development: Parental investment and family processes. Child
Development, 73(6), 1861–1879. https://doi.org/10.1111/1467-8624.t01-1-00511
Zhang, T., Shao, Z., & Zhang, Y. (2016). Developmental steps in theory of mind of typical
Chinese children and Chinese children with autism spectrum disorder. Research in
Autism Spectrum Disorders, 23, 210–220. https://doi.org/10.1016/j.
rasd.2015.10.005
E.R. Baker et al.
