HHS Public Access

Author manuscript
Dev Psychol. Author manuscript; available in PMC 2017 November 01.
Author Manuscript

Published in final edited form as:

Dev Psychol. 2016 November ; 52(11): 1744–1762. doi:10.1037/dev0000159.

The Development of Self-Regulation across Early Childhood

Janelle J. Montroya, Ryan P. Bowlesb, Lori E. Skibbec, Megan M. McClellandd, and Frederick
J. Morrisone
Janelle J. Montroy: [email protected]; Ryan P. Bowles: [email protected]; Lori E. Skibbe:
[email protected]
aChildren’sLearning Institute, Department of Developmental Pediatrics, University of Texas
Health Science Center at Houston, 7000 Fannin Street Suite 2373H, Houston, TX 77030, USA
Author Manuscript

bDepartment of Human Development and Family Studies, Michigan State University, 552 W.
Circle Drive, 2F Human Ecology, East Lansing, MI 48823, USA
cDepartment of Human Development and Family Studies, Michigan State University, 552 W.
Circle Drive, 2E Human Ecology, East Lansing, MI 48823, USA
dDepartment of Human Development and Family Sciences, Oregon State University, 245 Hallie E.
Ford Center for Healthy Children and Families, Corvallis, OR 97331
eDepartment of Psychology, University of Michigan, 2030 East Hall, Ann Arbor, MI 48109

Abstract
The development of early childhood self-regulation is often considered an early life marker for
Author Manuscript

later life successes. Yet little longitudinal research has evaluated whether there are different
trajectories of self-regulation development across children. This study investigates the
development of behavioral self-regulation between the ages of three and seven, with a direct focus
on possible heterogeneity in the developmental trajectories, and a set of potential indicators that
distinguish unique behavioral self-regulation trajectories. Across three diverse samples, 1,386
children were assessed on behavioral self-regulation from preschool through first grade. Results
indicated that majority of children develop self-regulation rapidly during early childhood, and that
children follow three distinct developmental patterns of growth. These three trajectories were
distinguishable based on timing of rapid gains, as well as child gender, early language skills, and
maternal education levels. Findings highlight early developmental differences in how self-
regulation unfolds with implications for offering individualized support across children.
Author Manuscript

Keywords
Behavioral self-regulation; developmental trajectories; early childhood; longitudinal

The Development of Self-Regulation Across Early Childhood The development of effective

self-regulation is recognized as fundamental to an individual’s functioning, with
development during early childhood often considered an early marker for later life successes

Correspondence should be sent to Janelle J. Montroy, Children’s Learning Institute, Department of Developmental Pediatrics,
University of Texas Health Science Center at Houston, 7000 Fannin Street Suite 2373H, Houston, TX 77030, USA. Tel: +1 713 500
3831. [email protected].
 Montroy et al. Page 2

(Blair, 2002; Bronson, 2000; Calkins, 2007; Diamond, 2002; Gross & Thompson, 2007;
Author Manuscript

Kopp, 1982; McClelland & Cameron, 2012; Mischel et al, 2011; Moffitt et al., 2011; Vohs &
Baumeister, 2011; Zelazo et al., 2003). Research indicates that between ages three and seven
a qualitative shift in self-regulation may take place when children typically progress from
reactive or co-regulated behavior to more advanced, cognitive behavioral forms of self-
regulation (e.g., Diamond, 2002; Kopp 1982) that likely require the integration of many
skills such as executive functions and language skills (Calkins, 2007; Cole, Armstrong, &
Pemberton, 2010). Likewise, past research suggests wide variation in the level of self-
regulation skills children manifest during early childhood that consistently predicts a
multitude of short- and long-term outcomes such as school readiness, academic achievement
throughout primary school, adult educational attainment, feelings of higher self-worth, a
better ability to cope with stress, as well as less substance use, and less law breaking, even
among individuals at risk of maladjustment (McClelland, Acock, Piccinin, Rhea, &
Author Manuscript

Stallings, 2013; Mischel et al., 2011; Moffitt et al., 2011).

However, despite mounting evidence that early childhood is an important time period for the
development of self-regulation, little is known about how children’s trajectories of
development might vary across individuals over time (Bergman, Magnusson, & Khouri,
2002; Muthén & Muthén, 2000; Nagin, 1999). To address this gap, we examined the inter-
individual variation in children’s growth trajectories between preschool and early elementary
school based on evidence that self-regulation requires the coordination and processing of
multiple skills across several domains (Calkins, 2007; Cole et al., 2010). More specifically,
we posit that there will be differences related to when the integration of these skills begin to
manifest as well as differences in the patterns of how they are manifest as regulated behavior
(Blair, 2010; Blair & Raver, 2012; 2015; Calkins, 2007; Clark et al., 2013). In the current
study, we examined the development of behavioral self-regulation via the Head-Toes-Knees-
Author Manuscript

Shoulders task (Cameron et al., 2008) between the ages of three and seven with longitudinal
data involving up to eight measurement occasions for individual children across three
samples. We evaluated possible heterogeneity in the developmental trajectories of children’s
behavioral self-regulation using growth mixture modeling (GMM; Grimm, McArdle, &
Hamagami, 2007), and potential indicators of trajectory differences.

Defining Self-regulation
Self-regulation is a complex, multi-component construct (Blair & Raver, 2012; McClelland,
Cameron Ponitz, Messersmith, & Tominey, 2010; Schunk & Zimmerman, 1997; Vohs &
Baumeister, 2011) operating across several levels of function (e.g., motor, physiological,
social-emotional, cognitive, behavioral and motivational), that in its broadest sense
Author Manuscript

represents the ability to volitionally plan and, as necessary, modulate one’s behavior(s) to an
adaptive end (Barkley, 2011; Gross & Thompson, 2007). One approach to the complexity of
self-regulation has been to view the multiple functions of self-regulation as hierarchically
organized and, eventually, reciprocally integrated (Blair & Raver, 2012; Calkins, 2007).
Ultimately self-regulation depends on the coordination of many processes across levels of
function, with children’s ability to draw on, integrate, and manage these multiple processes
increasing across developmental time (McClelland & Cameron, 2012; McClelland et al.,
2014).

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 3

The current study focuses on self-regulation in relation to its role in successful classroom
Author Manuscript

functioning (McClelland & Cameron, 2012; McClelland, Morrison, & Holmes, 2000;
Nesbitt et al., 2015). Effective self-regulation in the classroom requires that the child
seamlessly coordinate multiple aspects of top down control (i.e., executive function) such as
attention, working memory, and inhibitory control along with motor or verbal functions to
produce overt behaviors, such as remembering multi-step directions amidst distractions
(Cameron Ponitz et al., 2008; McClelland et al., 2007). This form of self-regulation is
therefore typically termed behavioral self-regulation (c.f., emotional self-regulation; Gross
& Thompson, 2007). To evaluate individual differences in development of self-regulation
across multiple years, we used a well validated direct assessment of behavioral self-
regulation, the Head-Toes-Knees-Shoulders task (HTKS, Cameron et al., 2008) that captures
variations in behavioral self-regulation throughout the entire range of early childhood,
making it possible to accurately assess developmental change on a common scale across
Author Manuscript

time (Cameron Ponitz et al., 2008; Connor et al., 2010; McClelland et al., 2007; Skibbe et
al., 2012). The HTKS is a short, game-like task where children are asked to ‘do the opposite’
in regards to a set of paired rules. For example, if the child is asked to touch their head,
instead they must touch their toes. This task taps three executive function skills (McClelland
et al., 2014) in order to make a gross motor response: 1. attention (ability to focus on
instructions and current stimuli), 2. working memory (ability to process the current trial
while holding a rule or set of rules in mind), and 3. inhibition (ability to ignore a well
learned response in order to respond in a counter-intuitive way).

Executive functions help an individual understand, monitor, and control their own reaction to
the environment, as well as problem solve regarding desired future behaviors and/or
outcomes. Put another way, the coordination of these skills often forms the basis of a child’s
ability to respond adaptively within the classroom. Notably a distinction has been made in
Author Manuscript

recent years between executive functions at the service of abstract or decontextualized

environments, and executive functions at the service of adapting to environments that require
the regulation of affect and motivation (e.g., Hongwanishkul et al., 2005). Sometimes
referred to as ‘cool’ executive functions and ‘hot’ executive functions within cognitive
traditions (e.g., Zelazo & Carlson, 2008), these skills can be considered as necessary
(although not entirely sufficient; Ursache & Blair, 2011) for behavioral and emotional
aspects of self-regulation, respectively (Zhou & Chen, 2008). Both hot and cool aspects are
important for development; hot aspects are usually more associated with socio-emotional
health and outcomes, while cool aspects are more associated with cognitive and academic
outcomes (Kim et al, 2013). The HTKS task generally draws on cool aspects of executive
function, although in reality no task is entirely free of an emotional context, with distinctions
generally being a matter of degree (Manes et al, 2002).
Author Manuscript

The Development of Behavioral Self-regulation

The development of self-regulation begins in infancy, with many of the skills that are
important for behavioral self-regulation developing first as separate domains, then becoming
organized and integrated over time (Barkley; 2011; Corrigan, 1981; Diamond et al., 1997;
Kopp, 1989; Stifter & Braungart, 1995). Previous work indicates that not only do separate
facets of self-regulation appear to develop at different times and rates (such as emotional

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 4

self-regulation generally preceding the development of behavioral self-regulation; Howse et

Author Manuscript

al., 2003) but also the underlying skills may also develop at different times. For example, the
ability to delay a response (an outcome most strongly associated with developing inhibitory
control) appears to develop earlier than other executive skills (Lengua, et al., 2015).
However, despite differences across individual facets and skills associated with self-
regulation, previous research consistently indicates that children younger than three have
difficulty simultaneously coordinating and utilizing multiple executive function skills to
create a behavioral response that also requires a motor or verbal action (Carlson, Moses, &
Breton, 2002; Diamond, 2002; Zelazo et al., 2003). However, after age three and during
early childhood, the individual skills that support behavioral self-regulation (e.g., see Cole et
al., 2010; Diamond et al., 1997 or Rothbart et al., 2006), as well as behavioral self-regulation
itself as an integration of those skills, rapidly develop(s), signifying a qualitative shift in
children’s regulatory abilities (Best & Miller, 2010; Garon et al., 2008; Kopp, 1982; Zelazo
Author Manuscript

et al., 2008).

Specifically, cross sectional work with a multitude of tasks indicates a rapid increase or
“leap” in performance on tasks that require the integration of several executive function
skills into behavior, such as the HTKS task, the Dimensional Change Card Sort, The Day/
Night, Bear/Dragon, Fish Flanker, and Luria’s tapping task (Diamond, 2002; Gerstadt et al.,
1994; Rothbart et al., 2006; Rueda et al., 2004; Zelazo et al., 2003). For example, there are
large group differences in accuracy on a fish flanker task (see Rueda et al., 2004 for a
description of the task) between four year olds and six year olds, but by about age seven,
children’s accuracy gains level off as performance becomes similar to adults (although
reaction time continues to improve; Rothbart et al., 2006; Rueda et al., 2004). In addition,
recent work explicitly evaluating behavioral self-regulation longitudinally (Cameron Ponitz
et al., 2008), as well as several studies of underlying executive function skills (Chang, Shaw,
Author Manuscript

Dishion, Gardner & Wilson, 2014; Clark et al., 2013; Diamond et al., 1997; Wiebe, Sheffield
& Espy, 2012) indicate non-linear growth with rapid gains followed by a decelerating rate of
gain in performance.

In summary, theory and research both provide evidence of rapid gains in the ability to
regulate behavior that are likely linked to the integration of multiple processes, but
particularly processes considered under the umbrella of executive function, such as attention,
working memory and inhibition (Cameron Ponitz et al., 2008; Chang et al., 2014; Diamond,
2002; Rothbart, et al., 2006). Based on these findings, we expect that the development of
behavioral self-regulation in early childhood is likely best represented by a nonlinear
function (Diamond, 2002). Specifically, we expect that between the ages of three and seven
years, gains in self-regulation will increase rapidly as multiple processes become more
Author Manuscript

coordinated, followed by later decelerated growth (e.g., Cameron Ponitz et al., 2008; Chang
et al., 2014; Wiebe et al., 2012).

Heterogeneity in behavioral self-regulation development

Several prominent theories suggest the possibility of multiple self-regulation growth
trajectories across early childhood (see Blair, 2010; Blair & Raver, 2012; 2015; Calkins,
2007; Lerner & Overton, 2008). Specifically, theories drawing on psychobiological or

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 5

dynamic systems models (Blair & Raver, 2015; Lerner & Overton, 2008) indicate a back and
Author Manuscript

forth developmental relationship between children’s biological traits and their experiences.
These theories contend that how children learn to regulate their behavior can vary widely
given that biological predispositions such as temperament and early environmental
experiences greatly vary. However, few studies have fully tested whether there are
underlying trajectory differences in self-regulation such that children develop behavioral
self-regulation in differing ways (i.e., process differences) and/or at different rates (Posner &
Rothbart, 2000; Rothbart, Ellis, Rueda & Posner, 2003). The majority of work in this area
has noted mean differences in the amount of self-regulation children are able to exert at a
given age during early childhood; only recently have studies begun to focus on growth in
self-regulation and predictors thereof. Of these studies, few have accounted for systematic
inter-individual differences across time (but see Vallotton & Ayoub, 2011; Wanless et al.,
2016; Willoughby et al., 2016).
Author Manuscript

Instead, the majority of studies evaluating self-regulation growth have focused on utilizing
child and environmental aspects to predict aggregate variation around a general slope and/or
rate mean (e.g., Blandon et al., 2008; Cameron Ponitz et al., 2008; Clark et al., 2013),
without further consideration for whether this variation may indicate qualitatively distinct
developmental change. This makes it is difficult to conclude whether there actually are
subgroups of children with systematic differences in how self-regulation processes unfold
(Rogosa, 1988). Likewise, findings at the aggregate level do not necessarily describe the
relationship among variables for a single individual or subgroup of individuals (von Eye &
Bergman, 2003). This makes it equally difficult to accurately map out predictive relations
between children’s individual traits and environments and their self-regulation development.

Only one study to date has evaluated multiple trajectories across children in behavioral self-
Author Manuscript

regulation (Wanless et al., 2016). This studied focused specifically on a Taiwanese sample of
children and indicated two distinct behavioral self-regulation trajectories: an “increasing”
developers trajectory with children rapidly gaining in self-regulation and then leveling off
across early childhood, and a “steady-then-increasing” trajectory with children
demonstrating few regulatory gains between ages 3 – 5 years and rapid gains after 5 years of
age. However, this study only includes a relatively small sample, and focuses on a
homogenous population in Taiwan.

The current study builds upon and extends this previous work by directly examining the
possibility of qualitatively different behavioral self-regulation growth trajectories between
the ages of three and seven in a large heterogeneous population. We focus specifically on
behavioral self-regulation as theoretical considerations indicate that the regulation of
Author Manuscript

behavior is expected to include multiple trajectories during early childhood given that the
multiple executive function inputs that support it are sensitive to not only genetic inputs but
experiential inputs and that these inputs are rapidly developing and differentiating during
this time period (Blair & Raver, 2012; Lonigan & Allan, 2014). As part of investigating
potential trajectories, we also evaluate one rough environmental proxy and two child level
predictors in order to validate potential trajectory differences, and better understand patterns
of how and, possibly when, these factors matter for self-regulation development across
children.

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 6

Child factors
Author Manuscript

There are several early characteristics that previous studies have identified as having an
association with the development of behavioral self-regulation (Blair et al., 2011; Calkins,
Dedmon, Gill, Lomax, & Johnson, 2002; Cole et al., 2010; Matthews et al., 2009). The
current study focuses on children’s gender and language skills as these attributes are fairly
consistently linked to individual differences in self-regulation (Bohlmann et al., 2015;
Matthews et al., 2009; Ready et al., 2005), and potentially trajectory differences (Vallotton
& Ayoub, 2011).

Gender—Previous findings generally indicate that boys have lower levels of self-regulation
than girls (Kochanska et al., 2001; Matthews et al., 2009, 2014; McClelland et al., 2007),
with gender differences often increasing across time (Matthews et al., 2014). It is not well
understood why such gender differences occur (though see Entwistle, Alexander, & Olson,
Author Manuscript

2007), although recent work suggests gender differences may in part relate to cultural beliefs
and expectations (von Suchodoletz et al., 2013; Wanless et al., 2016). However, there is
evidence that, from an early age, gender is associated with what type of self-regulation
developmental trajectory a child is likely to follow (Vallotton & Ayoub, 2011). For example,
during toddlerhood, boys’ self-regulation generally dips around age two then rises, while
girls’ self-regulation rises steadily, resulting in gender differences at ages two and three
favoring girls. Additional research focused on kindergarteners suggests that a subset of boys
persist in demonstrating very low levels of behavioral self-regulation (Matthews et al.,
2009), potentially signifying these boys not only continue to developmentally lag behind
girls, but that they may also not be acquiring self-regulation in the same way that peers are.
Given these past findings, we expected that boys may be more likely to follow a potentially
lagged trajectory.
Author Manuscript

Language—Language is another child attribute that affects developing self-regulation, and

may be an important factor for understanding potential self-regulation trajectory differences
across children. Theoretically, language is thought to give children “mental tools” to help
them organize and modify their thoughts and behaviors (Vygotsky, 1934/1986). During early
childhood, expressive language in particular may be important as it enhances the ability of
the child to both name their own current state and manipulate that state in relation to a
specific context (Cole et al., 2010). It also seemingly enhances children’s ability to hold task
requirements in mind (Karbach, Eber, & Kray, 2008). Research evaluating how expressive
language helps toddlers to self-regulate suggests that trajectories of self-regulation vary
between children based on the child’s observed expressive vocabulary skills (Vallotton &
Ayoub, 2011). Likewise, early expressive language skills are also associated with higher
levels of early self-regulation, with greater language gains across preschool and the
Author Manuscript

transition to kindergarten associated with greater self-regulation gains (Bohlmann, Maier, &
Palacios, 2015). This suggests that children with higher levels of expressive language
develop self-regulation faster compared to children with lower levels of language. We also
expected expressive language to be related to self-regulation growth on the HTKS because
children use both expressive and receptive language when completing the task (and can
answer verbally if needed/verbalize actions). Based on these previous findings, the pattern of
associations between expressive language at the start of schooling and potential self-

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 7

regulation trajectories should follow a similar pattern such that lower levels of expressive
Author Manuscript

language are associated with a distinct, potentially lagged trajectory compared to higher
levels of expressive language.

Mother education
In addition to child attributes and competencies, past research consistently demonstrates that
children’s environments affect developing behavioral self-regulation (Blair, 2010; Grolnick
& Farkas, 2002; Landry et al., 2006). One particularly salient aspect of children’s
environments that may affect developing self-regulation is their mothers’ education levels
(e.g., see Miech, Essex, & Goldsmith, 2001). Mother education often serves as a rough yet
important proxy of family socioeconomic status and resources (Bradley & Corwyn, 2002;
Hoff, Laursen & Tardif, 2002). Low maternal education levels have been linked to lower
socioeconomic resources and higher stress levels that, over time, can affect children’s
Author Manuscript

developing neuroendocrine processes (e.g., such as cortisol levels). These processes are
theorized to directly shape developing self-regulatory response patterns (see Blair & Raver,
2015). Maternal education levels are also associated with distinct parenting profiles that
include mothers’ warmth, responsiveness, use of rich language inputs, and ability to
maintain their children’s attention (Guttentag, Pedrosa-Josic, Landry, Smith & Swank,
2006), all factors that predict individual differences in children’s self-regulation levels (see
Grolnick & Farkas, 2002). Thus, mother education levels are also expected to serve as
indicator of valid differences in children’s developing self-regulation patterns.

Current Study
Past theory and research indicate that behavioral self-regulation rapidly develops during
early childhood with possible heterogeneity of early self-regulation trajectories (e.g., Blair &
Author Manuscript

Raver, 2015; Vallotton & Ayoub, 2011; Wanless et al., 2016). To better understand
behavioral self-regulation development and heterogeneity across children, we used the
HTKS measure to assess and evaluate development via latent growth curve modeling. We
then directly focused on potential trajectory differences in early childhood utilizing growth
mixture modeling. We hypothesized that most children would demonstrate rapid gains in
their behavioral self-regulation trajectory (e.g., Cameron Ponitz et al., 2008; Matthews et al.,
2009), compared to peers, with these gains occurring early in schooling (Blair & Raver,
2015). However, we hypothesized a subset of children would demonstrate a lagged
behavioral self-regulation trajectory across early childhood as they are not ready to integrate
the multiple processes required by advanced behavioral self-regulation when they first reach
school (Wanless et al., 2016; Willoughby et al., 2016). As part of trajectory validation, we
utilized multiple diverse samples that included the same measure of behavioral self-
Author Manuscript

regulation within similar age ranges, and with similar data collection procedures in order to
evaluate whether trajectory findings replicate across a diverse population of children in
different areas of the United States. We then further validated trajectories in relation to
predicted associations between three characteristics: gender, language ability, and maternal
education levels. We expected that these factors would distinguish what trajectory a child
was likely to follow, with patterns of association matching previous findings, offering
evidence indicating that trajectories capture meaningful individual difference as well as

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 8

increasing our understanding how these characteristics relate to individual differences in

Author Manuscript

development over time.

Methods
Participants
Participants consisted of 1,386 children across three samples that had at least two
assessments of self-regulation between the ages of three and seven. Children were
administered the same direct assessment of behavioral self-regulation in all three studies (the
Head-Toes-Knees-Shoulders Task; Cameron Ponitz et al., 2008). The samples are described
below.

Michigan longitudinal sample—The first sample was collected in predominantly

middle- to upper-SES suburban area with a range of ethnic diversity in southeast Michigan.
Author Manuscript

Participants included 351 (51% female) children followed from preschool through second
grade as part of the “Pathways to Literacy” longitudinal study evaluating children’s socio-
emotional and cognitive development (e.g., ***; blinded for review; 32 of the full sample of
383 were not included due to having fewer than 2 assessments of self-regulation). Students
attended 314 classrooms located within 16 schools in a single suburban school district. All
schools within this district that included at least one preschool classroom were represented
and preschool classrooms included Head Start classrooms (n = 49) as well as those that
charged tuition. On average, children were 48.16 months (SD = 7.35) old at the start of the
study: just over four years of age. The bulk of parents who provided information about their
child’s ethnicity (n =257) reported that their child was White/Caucasian (80%). The
remainder of children were described as African-American (4%), Asian/Indian (5%),
Hispanic (1%), and Multi-racial (3%). Several parents (8%) noted that another ethnicity
Author Manuscript

would describe their child best. Most (n = 278) families noted that their child’s native
language was English, although some families (n = 73) did not respond to this question1.
Median household income was high (i.e., $115,000; Range = $11,000 to $650,000) as were
parent education levels, with over 75% of mothers (n = 233) reporting that they had earned
at least a bachelor’s degree.

Families were recruited via flyers sent home in children’s backpacks at the beginning of the
school year(s). Children’s self-regulation was evaluated in the fall and spring of each year
that the child was in the study until they finished first grade (up to 8 times) as part of a
battery of measures administered by trained research assistants. Language assessments were
administered during the fall of children’s first preschool year. Parents also filled out
demographic information including child gender and information related to education level
Author Manuscript

in the fall of their child’s first preschool year.

MLSELD preschool sample—The second sample consisted of 642 (51% female)

preschool aged children from middle-SES communities with data waves collected over four

1In the Michigan longitudinal sample, > 7 languages other than English were reported as children’s primary language, the largest
language sub-group consisted of Arabic (n=4). For the MLSELD preschool sample, 18 languages other than English were reported,
the largest language sub-groups consisted of Spanish, and Korean speakers (n = 8).

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 9

years in Michigan as part of the Michigan Longitudinal Study of Early Literacy

Author Manuscript

Development (MLSELD preschool sample; ***; blinded for review). Children were drawn
from 78 classrooms across six schools: two in central Michigan and four in western
Michigan. Schools in central Michigan were accredited by the National Association for the
Education of Young Children. One was associated with a university and the other was a joint
public/university preschool that also had a population of Head Start eligible children (less
than 5% of the current sample). The four schools in western Michigan were part of the
area’s public schools. In western Michigan, families were recruited for participation at a
parent information night, while at the central Michigan schools families were recruited via
flyers sent home in children’s backpacks. Children were on average approximately four
years of age at the start of the study (M = 47.74, SD = 7.02). Most parents who provided
information about their child’s ethnicity (n = 479) reported that their child was White/
Caucasian (81%). Children who were African American (2%), Hispanic (3%), Asian (7%),
Author Manuscript

multi-racial (4%), and those from ‘other’ (3%) ethnicities also participated in the present
work. Among families reporting primary language spoken at home, almost all reported
English (n = 443) although some families did not respond to this question (n = 158). Over
half of mothers (n = 374) reported that they had earned at least a bachelor’s degree.
Household income levels were not collected as part of this study.

Children’s self-regulation was collected in the fall and spring of each year by a trained
research assistant in a quiet setting; self-regulation was also collected two additional times in
winter (about a month and a half apart) in two years of the study, and one additional time in
the winter in one year of study (i.e., in the first two years of study self-regulation was
assessed four times, year three it was assessed three times, and it was assessed twice in year
four). Children’s language skills were tested in the fall of their first year of preschool and
parents filled out demographic information including child gender and information related to
Author Manuscript

education level at this time as well. Some children (n = 160) participated in the study over
the course of two years and a small subset of children (n = 13) were included in the study for
three years. Thus these children had their self-regulation evaluated more frequently. Across
the larger MLSELD study (n = 888), 246 children either had only one self-regulation
assessment (n = 133) or no assessments (n = 113; by design, only half of the sample had
self-regulation assessed in year 4).

Oregon sample—The third sample was recruited from a mixed-SES rural site in Oregon
and consisted of 393 (50% female) children followed from preschool through kindergarten
as part of a measurement study focused on improving measures of school readiness and self-
regulation (***; blinded for review; 38 of the full sample of 431 were not included due to
having fewer than 2 assessments of self-regulation related to study attrition). Children were
Author Manuscript

drawn from 37 classrooms in 17 schools, with 54% (n = 209) of children in Head Start
programs. Children were on average over four and a half years old at the start of the study
(M = 56.14, SD = 3.65). Most parents who provided information about their child’s ethnicity
(n = 354) reported that their child was White/Caucasian (63%), or Hispanic (19%). Children
who were African American (1%), Asian/Pacific Islander (4%), multi-racial (13%), and
those from ‘other’(1%) ethnicities also participated in the present work. Families reported
that English was the primary language spoken at home for most children (n = 297); however

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 10

this sample also included a subsample of children whose primary language was Spanish (n =
Author Manuscript

60) who were tested in Spanish (all Spanish speakers were enrolled in Head Start). On
average, mothers reported having attended some college, but only 43% reported that they
had earned a bachelor’s degree or higher. Of respondents, 58% indicated that their families
qualified for public assistance such as WIC or food stamps in the past four years.

Families were recruited through letters sent home with an enrollment packet sent during the
summer before the beginning of the preschool year. Self-regulation was assessed each year
in the fall and spring by trained research assistants (up to four time points). Language skills
were assessed in the fall of children’s preschool year, and parents filled out demographic
surveys at this time.

Measures
Self-regulation—Children’s self-regulation was measured directly using the Head-Toes-
Author Manuscript

Knees-Shoulders task (Cameron Ponitz et al., 2008; Connor et al., 2010; Matthews, et al.,
2009). During the task, children are provided with paired behavioral rules (e.g., touch your
head/touch your toes) and asked to do the opposite of what they were instructed to do. For
example, when a child is asked to touch her toes, she should complete the opposite action
(touch her head). The first ten items include one paired rule (e.g., head/toe). If children
respond correctly to four or more items, they are given ten additional items with two paired
rules (e.g., head/toes, knees/shoulders). Children earned two points for each correct
response, one point for each self-correction (i.e., an initial movement to the incorrect
response, but ultimately ending with the correct response), and zero points for each incorrect
response. Scores ranged from 0–40, with higher scores indicating higher self-regulation. In
the first year of the Michigan longitudinal sample data collection, when all children were in
preschool, only the first half of the HTKS was administered, as the second half had not yet
Author Manuscript

been developed. We therefore used a Rasch measurement approach to extrapolate an

expected score on the entire 40 item task (details are provided in Bindman, Hindman,
Bowles, & Morrison, 2013).

The HTKS has good construct and predictive validity within many culturally diverse
samples, and across languages (Cameron Ponitz, McClelland, Matthews, & Morrison, 2009;
McClelland et al., 2007; von Suchodoletz et al., 2013; Wanless, et al., 2011). Scores on this
measure are significantly correlated with reported self-regulation in the classroom, parental
reports of attention (Cameron Ponitz et al., 2009; McClelland et al., 2007) and other
measures of self-regulation and executive function tasks. The HTKS also loads well onto a
self-regulation factor with other similar measures (Allan & Lonigan, 2014). In addition, past
evidence indicates that growth in HTKS performance does not appear to be a function of
Author Manuscript

practice effects (Cameron Ponitz et al., 2008).

In terms of predictive validity, the HTKS consistently predicts academic achievement across
diverse sample populations (McClelland et al., 2007; Montroy et al., 2014; von Suchodoletz
et al., 2013; Wanless et al., 2011). Notably, evidence suggests HTKS scores and growth are
generally stronger predictors of growth in academic achievement than other self-regulation
and executive function measures, particularly measures that mostly capture one skill versus
an integration of skills (Lipsey et al., 2014; McClelland et al., 2014).

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 11

The HTKS has strong reliability (Cameron Ponitz et al., 2008; Matthews et al., 2009;
Author Manuscript

Montroy, et al., 2014; Wanless, et al., 2011). Past studies consistently report high levels of
inter-rater reliability (kappa > .90; Cameron Ponitz et al., 2008), and internal consistency
estimates above .80 (Montroy et al., 2014; Wanless et al., 2011). Within the current study
internal consistency was also good, with Cronbach’s alpha values ranging from .85–.94
across samples.

Language—Language skills were assessed across all three samples. In the Michigan
longitudinal sample and the Oregon sample, the Picture Vocabulary subtest of the Woodcock
Johnson III was used as an indicator of language (Woodcock & Mather, 2001), while the
Test of Preschool Early Literacy picture vocabulary subtest (TOPEL; Lonigan, Wagner,
Torgeson, & Rashotte, 2007) was used in the MLSELD preschool sample. Both the TOPEL
and WJ vocabulary tests have been well validated and extensively used in the literature as
indicators of expressive vocabulary (Bohlmann et al., 2015; Pence, Bojczyk & Williams,
Author Manuscript

2007; Wilson & Lonigan, 2009; Vallotton & Ayoub, 2011).

The Woodcock Johnson picture vocabulary subtest is an untimed picture naming task where
children are shown a series of pictures and are asked to verbally identify the image. Children
speaking Spanish in the Oregon sample were administered the Picture Vocabulary subtest of
the Spanish version of the Woodcock-Johnson, the Bateria III Woodcock-Munoz. Picture
Vocabulary has strong evidence of reliability (e.g., split half reliability between 0.76–0.81
for English speaking children and 0.88–0.89 for Spanish speakers) and validity. We used W-
scores, a Rasch-type measure of ability, for all analyses. This type of score ensures
measurement on an equal-interval scale and takes into account the level of item difficulty in
relation to a children’s age.
Author Manuscript

The TOPEL picture vocabulary subtest consists of 35 items including various untimed
picture naming tasks where children name pictures (1 point) and then describe aspects or
functions associated with the picture presented to them (e.g., What are they for? 1 point)2.
Thus raw scores range from 0–70. Test-retest reliability for this subtest is .81 and test
developers indicated that scores were strongly related to the Early One-Word Picture
Vocabulary Test (r = .71, Brownell, 2000). The TOPEL was administered only in years 2 and
3 of the study. The remaining n =77 in year 1 and n = 113 in year 4 were not administered
by design.

Mother education and gender—Across samples, demographics questionnaires were

provided to parents including information regarding the child’s gender and parent education
levels. For the Michigan longitudinal and the Oregon sample, mothers were asked to report
Author Manuscript

education in terms of the number of years of schooling they had completed, while the
MLSELD preschool sample was asked to report education levels via an 11-point survey
question where education level categorically increased with 1 = less than a high school level
education, 7 = a bachelor’s degree, and 11 = an advanced graduate degree (e.g., Ph.D or

2We also ran analyses with only points for naming an object correctly included (i.e., excluding the points received for describing
functions). The results were similar with both scoring methods, thus we present findings using the total (naming and functions) score
as this is commonly how the subtest is scored and used in previous studies.

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 12

M.D). For comparability across samples, data from the Michigan longitudinal and Oregon
Author Manuscript

sample were converted to the 11-point scale used by the MLSELD sample.

Analytic Approach
Analyses were done in two parts to (1) describe the general growth trajectory of self-
regulation and (2) evaluate heterogeneity in self-regulation trajectories across children. First,
we used latent growth curve models (Bowles & Montroy, 2013; McArdle, 1986; Meredith &
Tisak, 1990; Singer & Willett, 2003) to examine the general trajectory of development of
self-regulation. These models provide information about the average values of children’s
self-regulation (level of self-regulation) at a specified time, how rapidly their skills increase
or decrease (i.e., slope), and whether this change is constant or might accelerate or
decelerate (i.e., linear versus nonlinear growth). The general equation for the latent growth
curve models we used was:
Author Manuscript

where Self-reg[t]n is the HTKS score for child n at age t; A[t] or the basis coefficient(s), are
a function defining the shape of the growth trajectory, determining both the precise
interpretation of the Level and the Slope, and the nature of change; Leveln represents child
n’s predicted level of self-regulation at the point where A[t] is 0; and Slopen generally
reflects child n’s predicted rate of growth on the HTKS per unit of the basis coefficients. We
considered five models for the trajectory: linear, quadratic, exponential, logistic, and the
latent basis model. Due to variation in what age children received assessments and the time
between assessments, scores were grouped by child age into three month windows in each
dataset3. To evaluate what model optimally described the general growth trajectory of
Author Manuscript

behavioral self-regulation, we utilized the Akaike Information Criterion (AIC) and the
Adjusted Bayesian Information Criterion (aBIC) fit indices.

Next, we utilized growth mixture modeling (GMM; Muthén, 2001) to evaluate if there were
multiple growth trajectories of early childhood behavioral self-regulation. In GMM, the
trajectory classes are formed based on the growth factor means and variances (e.g., Level
and Slope means and variances) with each class defining a different growth trajectory
(Muthén, 2001). GMM also captures individual variation around these growth curves by
estimating the growth factor variances within each class (Muthén & Muthén, 2000). Within
the GMM models, we chose to restrict trajectory shape to the shape indicated by the latent
growth curve models. This is common practice in the GMM literature when there is not a
strong theory regarding shape of trajectory differences across the population. However, slope
Author Manuscript

and rate parameters (but not functional form) were ultimately allowed to vary across
trajectories, thus providing information regarding different developmental progressions and
patterns. In all models, errors were specified to be uncorrelated. We determined best model
fit based on AIC and aBIC indices (Tofighi & Enders, 2007), entropy, and bootstrapped
likelihood ratio tests (BLRT) which compares the fit of the estimated model with k classes to

3In years one and two of the MLSELD preschool study, most children had two assessments within three months. To ensure local
independence within the three month windows, only one assessment was included, chosen at random.

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 13

the same model with one less class (k-1), with p-values less than .05 indicating that the
Author Manuscript

estimated k class model fits better than the k-1 model (Grimm, Ram & Estabrook, 2010).
Note, BLRTs can only test differences in relation to what number of classes fits best, they
provides little information when comparing within class solutions with differing parameters
(e.g., whether a solution with constrained random effects versus variable random effects fit
best). In addition we also considered whether results were interpretable and meaningful, and
we took into account estimation parameters as well as estimation history as these are all
relevant indicators of model comparison and selection (Grimm et al., 2010). To evaluate the
predictors of trajectory classes, we assigned each child to the class with the highest
probability, and used logistic regression based analyses to predict class membership4. All
analyses were completed with Mplus version 7.2 (Muthén & Muthén, 1998–2010), utilizing
full information maximum likelihood to account for missing data. In all analyses, year of
study was included as a saturated covariate given its relationship with missing data in all
Author Manuscript

samples, and the MLR estimator was used as this estimator provides the most accurate
parameter estimates when missing data are present (Enders, 2010; Graham, 2003).

Results
Descriptive Statistics of Behavioral Self-regulation
On average, children demonstrated gains in behavioral self-regulation as measured by the
HTKS between the ages of three and seven; see Table 1 for a comparison of average gains
across samples. Individual observed trajectories for a random subset of 25 children’s scores
per sample are presented in Figure 1. In all samples, there were substantial individual
differences, and periods of acceleration and deceleration in growth both within and across
children. Correlations are provided in the supplementary materials.
Author Manuscript

The Development of Behavioral Self-regulation

Fit statistics for the five latent growth curve models are reported in Table 2 by sample. In all
samples, both AIC and aBIC suggested that the changes and between person differences in
early childhood behavioral self-regulation development were best described by an
exponential curve; see Figure 2. Across samples patterns varied such that: in the MLSELD
preschool sample, children’s growth accelerated across preschool. However, in the Oregon
and Michigan longitudinal samples that followed children across early elementary grades,
children demonstrated faster gains early in preschool with gains slowing in early elementary
school5.

Heterogeneity in Behavioral Self-regulation Development

Author Manuscript

Growth mixture modeling allows for the estimation of different trajectories with the
possibility of every estimated parameter differing across groups (e.g., means, variances,

4We also ran the 3-step approach discussed in Asparouhov & Muthen (2014). However, this approach results in listwise deletion of
missing data at the predictor level. Thus, in addition to the initial 3-step approach models we also used multiple imputations of missing
predictor level data and re-ran all 3-step approach models. The pattern of results was the same in all cases.
5In order to rule out the possibility that the exponential shape was a product of floor and ceiling effects across time points, all analyses
were re-ran utilizing Tobit growth curve models (Wang, Zhang, McArdle & Salthouse, 2009), which can account for floor or ceiling
effects. In all samples, an exponential model still fit best compared to other models, even when accounting for floor or ceilings.

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 14

covariances, and basis coefficients; Grimm et al., 2007; McArdle & Bell, 2000; McArdle &
Author Manuscript

Nesselroade, 2003). Currently, there is no generally accepted strategy for how GMMs should
be evaluated in terms of which constraints to relax first (Grimm et al., 2007). However,
similar to past studies, we evaluated models based on the principles of factorial invariance
studies (e.g., Grimm et al., 2007), followed by an examination of models with different
constraints related to within trajectory variation patterns (e.g., Kreuter & Muthén, 2008).
Specifically, we evaluated fit across separate datasets starting with a two growth trajectory
model where only level and slope means were allowed to vary across trajectory groups. In
all models, the exponential shape indicated by the latent growth curve analyses was specified
for all classes. Rate of acceleration/deceleration within the exponential trajectory(ies) was
initially constrained to be the equal across curves (i.e., only timing differences were allowed
with identical developmental form and rate of change). This would offer strong evidence that
variation in behavioral self-regulation trajectory growth across children is similar, but with
Author Manuscript

differences in developmental timing. All within curve variations (random effects) were also
initially constrained (Kreuter & Muthén, 2008). We then allowed rates of acceleration/
deceleration to vary in order to evaluate possible differences in how children develop self-
regulation across early childhood. Specifically across different class/trajectory solutions, one
could potentially see changes in the sign for rate of change indicating whether rate was
accelerating or decelerating across the specific study time period, as well as differences in
rate parameter magnitude (i.e., it was possible for a non-significant rate parameter to be
found, which would be similar to if a linear trajectory was specified). Additional growth
trajectories were then added to determine what number of trajectories best fit the data. Once
number and trajectory rate of change differences were determined, we progressively relaxed
within trajectory level and growth variances and covariances to investigate how closely
individual children followed group trajectories.
Author Manuscript

Results indicated that the three trajectory solution fit best in all samples based on the
evaluation of global fit statistics in association with bootstrapped LRTs, iteration history,
convergences, estimated parameters, and entropy values; see Table 3 for a summary of fit
statistics across the different models and samples. In all samples, the three trajectory
solution was also interpretable such that children generally demonstrated timing differences
in early childhood self-regulatory gains but with some variation in rate across trajectories.
As seen in Figure 3, children demonstrated either early gains, intermediate gains, or later
gains relative to sample peers. In general, children’s individual self-regulation trajectories
also conformed closely to the three trajectories. Specifically, the models where within
trajectory level, slope or rate variations were constrained to zero fit best for all three
samples.
Author Manuscript

As expected, the percent of children predicted to follow a given trajectory varied across
samples, yet results still demonstrated clear consistency. Descriptively, in the Michigan
longitudinal sample, 20% of children were classified as early developers who demonstrated
higher initial levels of self-regulation and earlier gains, 45% as intermediate developers who
had low initial self-regulation, followed by rapid gains, and 35% as later developers, who
started with lower levels and gained more slowly compared to other groups. A similar
pattern occurred within the MLSELD preschool sample with 29% of children classified as
early developers, 45% of children as intermediate developers, and 26% as later developers.

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 15

Likewise in the Oregon sample: 50% of children were early developers, 32% were
Author Manuscript

intermediate developers, and 18% were later developers.

Broad patterns that were replicated across samples indicated that more than half of the
children assessed within a given sample (including both early and intermediate developers)
demonstrated rapid growth across preschool. On average, these children accurately
responded (correct or self-correct) 75% of the time or more by age five (58–60 months). Yet,
about 20% of children (later developers) consistently demonstrated relatively few early
gains, responding at less than 35% accuracy on the behavioral self-regulation task
throughout preschool. This pattern persisted into kindergarten in both samples that spanned
into early elementary school with gains for later developers lagging same sample peers; in
the Oregon sample gains do not pick up until 70–72 months of age (i.e., nearly 6 years of
age). In short, children classified as later developers were, on average, six months to a year
behind their intermediate developing peers and at least a year and a half behind early
Author Manuscript

developers.

Descriptively there were also several other similarities across samples, see Table 4 for means
across samples. In all three samples boys made up the majority of the later developers group
whereas the majority of early developers were girls. Likewise, across all three samples, early
developers had the highest mean levels of language and, on average, their mother’s obtained
the highest education levels. Children in the later developers group demonstrated the lowest
levels of language and had mothers with comparatively lower levels of education compared
to the other two groups.

Predictors of self-regulation growth trajectories—In this section we investigated

whether child attributes and mother education levels were indicative of which children are
Author Manuscript

more likely to follow which trajectory. Given that a subset of children in each sample
demonstrated a ‘later developer’ behavioral self-regulation trajectory, we first tested whether
these children differed from children who globally demonstrated rapid gains in preschool,
classified as ‘preschool developers’ in behavioral self-regulation (i.e., ‘intermediate’ and
‘early developers’ considered together). We followed this analysis up with a second analysis
comparing intermediate to early developers. In all analyses, gender, language, and maternal
education levels were included as predictors, controlling for children’s age at the first
measurement time point, race/ethnicity, and whether English was a indicated as a child’s
primary language.

We used the Benjamini-Hochberg procedure to account for multiple predictors (Benjamini

& Hochberg, 1995). This procedure controls for the false discovery rate (i.e., Type I errors)
associated with conducting multiple comparisons by ranking post-hoc the individual p-value
Author Manuscript

associated with each predictor from smallest p-value to largest. Each individual p-value is
then compared to its Benjamini-Hochberg critical value. The predictor with the largest p-
value that is less than the Benjamini-Hochberg critical value, and all predictors ranked
before it are considered significant. In the current study, we used a false discovery rate of .
05, thus all predictors with a Benjamini-Hochberg critical value below .05, regardless of raw
p-values, were considered significant. See Table 5 for all results including raw p-values and
calculated Benjamini-Hochberg critical values.

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 16

Gender: Child gender was linked to what trajectory a child was likely to follow in both the
Author Manuscript

Michigan longitudinal sample and the Oregon sample. Girls included in the Michigan
longitudinal sample were 1.79 times more likely to be classified as preschool developers
(following either an intermediate or early developers’ trajectory) versus later developers.
Further, in the Oregon sample, girls were 2.34 times more likely to be classified as early
developers versus intermediate developers. No other comparisons were significant.

Language: Children’s language skills also predicted within sample trajectory differences.
For a 1 SD increase in expressive vocabulary, children in the MLSELD preschool sample
were 1.40 times more likely to be classified as a preschool developer versus a later
developer. Furthermore, for a 1 SD increase in vocabulary children were 1.68 times more
likely to be identified as early developers versus intermediate developers. For the Oregon
sample, children were 1.23 times more likely to be classified as demonstrating a preschool
developer trajectory compared to later developer trajectory, and 1.63 times as likely to be
Author Manuscript

identified as early developers versus intermediate developers per a 1 SD increase in

expressive vocabulary. No other comparisons were significant.

Mother education: Mother education predicted trajectory differences in the MLSELD

preschool sample and the Oregon sample. In the MLSELD preschool sample, children
whose mothers completed one additional level of education (e.g., moving from associate to
bachelors) were 1.30 times more likely to be classified as a preschool developer than a later
developer. Likewise, of the children who demonstrated growth across preschool, there were
significant differences between intermediate developers and early developers such that for a
1 unit difference in mother education, children were 1.26 times more likely to be classified
as early developers. Within the Oregon sample, for every 1 unit increase in reported maternal
education, children were 1.23 times more likely to be classified as demonstrating a
Author Manuscript

preschool developer trajectory versus a later developer trajectory. No other comparisons

were significant.

Discussion
The development of effective self-regulation is widely recognized as an early marker for
later life successes (Blair & Raver, 2015; Calkins, 2007; Diamond, 2002; Gross &
Thompson, 2007; Moffitt et al., 2011). The primary goal of the current paper was to evaluate
the trajectory of behavioral self-regulation across early childhood in three distinct samples
across two states in order to determine if there was consistent heterogeneity in this
trajectory/ies across children. In addition, we evaluated whether several predictors that are
often associated with self-regulation could be used to indicate the type of trajectory a child
Author Manuscript

was likely to follow, with an ultimate goal of validating trajectory differences as meaningful
and consistent with previous findings, while further increasing our understanding of
developmental differences in early childhood behavioral self-regulation across children.
Overall, our findings indicate that self-regulation rapidly increases across early childhood,
with children in all samples following three distinct trajectories that were distinguishable
based on several important factors.

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 17

The Development of Behavioral Self-regulation

Author Manuscript

The general growth trajectory of behavioral self-regulation across early childhood was best
represented by an exponential function. This was the case across all three samples, although
there were differences across samples in whether growth was accelerating or decelerating,
likely related to sample specific differences in relation to what ages were captured. This
result is consistent with previous findings suggesting that behavioral self-regulation (and the
executive function skills that support behavioral self-regulation) develop(s) in a nonlinear
fashion with early, rapid gains during the preschool (e.g., Cameron-Ponitz et al., 2008;
Diamond, 2002; Wiebe et al., 2012).

Heterogeneity in Behavioral Self-regulation Development

The growth mixture modeling results indicated heterogeneity across children in the
developmental trajectories. Despite differences in sample locations, background
Author Manuscript

characteristics, and study windows, in all three samples children could be characterized as
either early developers, intermediate developers or later developers relative to their within
sample peers. Likewise, an identifiable overall pattern emerged such that for the majority of
(but not all) children in all samples there was a period of rapid development of behavioral
self-regulation in preschool with individual differences in when rapid growth occurred, how
rapidly it occurred, and what level of behavioral self-regulation children demonstrated at the
beginning of preschool.

Yet, overall, approximately 20% of children appear to make few gains in preschool. This
was especially the case in the MLSELD preschool and Oregon sample, with a subset of
children in both samples showing little to no growth until 60 months (i.e., 5 years) of age.
Children classified as later developers in the Michigan longitudinal sample did show gains
Author Manuscript

before age five, but these gains still lagged behind their sample peers (e.g., it takes children
classified as later developers an additional nearly three years, with children ~80 months old
or age 6.75, to obtain relatively identical mean levels of self-regulation their early
developing peers demonstrated at four years of age).

In general, we conclude that many children are still developing behavioral self-regulation
skills as they leave preschool and enter kindergarten, and that they may need behavioral
supports in kindergarten with a subset of children just beginning to develop advanced
behavioral self-regulation. The current study’s findings mirror teacher observations, with
nearly half of teachers indicating that they feel many children enter kindergarten without the
self-regulation skills necessary to be ready to learn in formal education setting (Rimm-
Kaufman et al., 2000). These finding are also fairly consistent with previous growth
trajectory findings within a smaller, Taiwanese sample (Wanless et al., 2016). Interestingly,
Author Manuscript

this previous study reports that both a two and three trajectory solution fit the data, but the
three class solution produced an extremely small class (p. 109; Wanless et al., 2016). It is
plausible that three classes were consistently found in the current study due to inclusion of
larger, more heterogeneous samples.

The overarching similar patterns observable across samples and in the literature at large,
indicate that once children start to utilize executive function skills to carry out complex

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 18

directions, it is a relatively short developmental time before they are able to accurately and
Author Manuscript

consistently follow multiple abstract rules such as those utilized in the Head-Toes-Knees-
Shoulders self-regulation task. Importantly, what is clear from the current study that was less
clear in previous work is that: 1) it generally takes just about 2–3 years to go from little
ability to self-regulate in the face of complex instructions to task mastery in this context and
2) All children will master the basic skills needed to participate in self-regulatory tasks,
however the age at which they do so varies across children and is related to child
characteristics and contextual factors Thus, the important relations between children’s levels
of behavioral self-regulation and later development outcomes documented in previous work
(e.g., McClelland et al., 2006) may be better described as reflecting differences in the timing
of the development of behavioral self-regulation.

Indicators associated with self-regulation growth trajectories—Our findings

indicate that early childhood behavioral self-regulation trajectories are distinguishable based
Author Manuscript

on associations with other variables. Consistent with past findings (e.g., Matthews et al.,
2009) the identification of a child as a girl was associated with earlier development
trajectories. In all samples, there were more boys in the later developers group, with results
reaching statistical significance in the Michigan longitudinal sample. Interestingly in the
Oregon sample there was also a statistically significant difference in the number of boys in
the intermediate trajectory compared to the early developer’s trajectory. These findings are
generally consistent with previous findings (e.g., Matthews et al., 2009) and suggest that
boys may need additional supports at the beginning or during preschool to ensure they
develop the self-regulation skills necessary for entry into kindergarten. Notably, gender
differences were observed in the two samples that spanned into the formal years of
education (kindergarten and beyond), possibly providing support to previous evidence
suggesting that gender differences appear more substantially over time (Entwistle et al.,
Author Manuscript

2007; Matthews et al., 2009; 2014). More work is necessary to determine why gender
differences are linked to different pattern of self-regulation development. Likewise
systematic investigations of how these associations change across time are also warranted.
One possibility beyond the scope of the current study is that boys can be more sensitive to
environmental experiences including chaos (Cameron Ponitz, Rimm-Kaufman, Brock, &
Nathanson, 2009; Wachs, 1992; Wachs et al., 2004), as well as parent and teacher
expectations of school success (Entwistle, et al., 2007; Wanless et al., 2011). It is possible
that overtime, continued chaos or continued high or low expectations from parents and
teachers may have cumulative effects on skill development.

Likewise, language was also predictive of what behavioral self-regulation trajectory a child
was likely to follow. Within the MSELD preschool and Oregon sample, higher levels of
Author Manuscript

expressive language at the start of preschool was associated with earlier development,
similar to other field findings (Bohlmann et al., 2015; Vallotton & Ayoub; 2011). There is
long standing theoretical support for the role of language as an organizational tool used to
aid self-regulation development (Cole et al., 2010; Vygotsky 1934/1986), and the current
study provides some evidence that early language skills may affect the timing and rate of
development of early self-regulation growth across early childhood. Higher levels of
language may give children the ability to organize and better understand incoming

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 19

information such as complex behavioral rules, contributing to the use of more complex self-
Author Manuscript

regulation that relies on attending to and keeping track of information (i.e., working
memory), while inhibiting a well learned dominant response pattern. Future studies are
necessary to better understand the ongoing relations between language and self-regulation
development (Bohlman et al., 2015). Specifically, more studies evaluating the potential
mechanisms regarding language’s role (e.g., see Winsler et al., 2009) and the specifics of
what other aspects of language lead to the rapid develop of self-regulation.

Mothers’ education levels also affected what trajectory a child was likely to follow. Early
developers generally lived in homes where mothers reported the highest education levels,
with children in Oregon and the MLSELD significantly more likely to be classified as an
intermediate or early developer based on maternal education. Mother education serves as a
proxy for aspects of the environment, including available household resources. These
resources include physical resources such as toys, games, learning materials, and books that
Author Manuscript

can support self-regulation development (Brooks-Gunn & Duncan, 1997), but also abstract
resources such as a less stressful home environment (see Blair & Raver, 2012). Additionally,
more highly educated mothers often hold different beliefs compared to their lower educated
counterparts that affect their parenting behavior towards their children (see Bradley &
Corwyn, 2002 for review), that could in turn affect developing self-regulation. Maternal
education may also help to explain why children within the Oregon sample developed self-
regulation slightly later relative to both Michigan samples. There was a higher
preponderance of mothers with lower education levels and families were more likely to be
living at or below the poverty line, perhaps indicating that children had fewer familial
supports to help them to develop their self-regulation. Furthermore, although beyond the
purview of the current study, it is also important to note that poverty consistently predicts the
complexity of children’s developing language (Fernald, Marchman, & Weisleder, 2013; Hart
Author Manuscript

& Risley, 1995), and is often related to gender differences in children’s performance
(Entwistle et al., 2007). A greater investigation across multiple indicators of a child’s
environment is a necessary next step in fully understanding what attributes affect children’s
self-regulation development in context.

Practical Implications
These findings have several implications. First, given that majority of children appear to
demonstrate rapid gains in behavioral self-regulation between the ages of three and seven
(but particularly during the preschool years), this research supports previous work
emphasizing this time period as a potential critical period (Blair, 2002; 2010; Diamond, et
al., 2007; McClelland & Cameron, 2012). Utilizing preschool curricula such as Tools of the
Mind (Bodrova & Leong, 2007) that center on scaffolding children’s early self-regulation
Author Manuscript

skills or targeted games and activities focusing on promoting self-regulation (e.g., Schmitt,
McClelland, Tominey & Acock, 2015) may provide children the support they need to
develop behavioral self-regulation skills early. On the other hand, programs and curricula
that focus on self-regulation and consider/target multiple aspects of the developing child
may ultimately prove even more impactful for preparing all children for formal education
(see Dickinson, McCabe & Essex, 2006; Lonigan et al., 2015). Future investigations

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 20

evaluating school contexts and their associations with different self-regulation

Author Manuscript

developmental trajectories are necessary to provide further support to these assertions.

In addition to offering insights into the developmental trajectories of behavioral self-

regulation, this work also offers tentative evidence indicating which trajectory a child is
likely to follow may be predictable based on background characteristics. Specifically, it may
be possible to screen for children who are at risk of entering kindergarten without the self-
regulation skills teachers feel are necessary to succeed (Rimm-Kaufman et al., 2000). For
example, our study indicates that boys from families with lower education levels who score
lower on early vocabulary tests compared to peers may be at risk of starting kindergarten
behind in self-regulation. Although a more thorough investigation of predictive traits is
warranted, even these few consistent predictors, in conjunction with delays in preschool self-
regulation progress, may be enough to indicate a need for support, particular given research
indicating that later behavioral self-regulation development can impact school readiness and
Author Manuscript

future academic skills and success for years to come (see McClelland et al., 2006; Wanless
et al., 2016).

This type of developmental work also could enhance progress monitoring by early childhood
professionals. Recent research has advocated that early education environments should
include a three-tiered system of service delivery that includes increasing instructional
intensity to the level of student needs (e.g., Fox, Dunlap, & Cushing, 2002). A better
understanding of how self-regulation development progresses, and what constitutes normal
versus delayed development (i.e., benchmarks related to student needs) gives teachers the
tools they need to customize level of instruction appropriately, including information for
when to seek outside support for children who may need intervention. Past research
indicates that early interventions targeted at increasing behavioral self-regulation skills can
Author Manuscript

result in significant behavioral self-regulatory gains as well as some academic gains,

particularly for individuals who are have lower initial levels of behavioral self-regulation
compared to peers (Schmitt, et al., 2015; Tominey & McClelland, 2011).

As preschool and kindergarten have evolved in focus across the last few decades (and
continue to evolve) from social-emotional skills to more academic skills (Kagan, Kauerz, &
Tarrant, 2007; Stipek, 2011), it is critical for researchers and policymakers alike to
remember that social-emotional development such as self-regulation development can have
long lasting impacts on children’s school readiness and success, including academic success.
Past research indicates that children who begin kindergarten with lower levels of self-
regulation skills also lag in math and literacy skills through sixth grade, with the gaps in
achievement widening through second grade between children with higher kindergarten self-
Author Manuscript

regulation and children with comparatively lower levels of self-regulation (McClelland et al.,
2006). Although future research linking academic achievement to different self-regulation
trajectory classifications is needed, we speculate that children classified in the current study
as ‘later developing’ may struggle with academic achievement given that these children did
not demonstrate the rapid gains their peers demonstrated in preschool with some not
showing those types of gains until nearly the end of kindergarten. It is possible that a lagged
trajectory in early childhood may have cascading effects on children’s future development as
many of the skills that rely in part on self-regulation skills may also be delayed until self-

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 21

regulation skills are more adequately acquired. Regardless, the fact that all ‘later developers’
Author Manuscript

across samples in the current study demonstrated markedly less developmental progress
compared to peers should be addressed as early as possible given that self-regulation in its
own right is an important developmental milestone (Bronson, 2000).

Limitations and Future Studies

The current study utilized one measure of self-regulation, which, although reliable and valid,
does not fully capture the multi-dimensional nature of self-regulation. The inclusion of other
measures of self-regulation that each focus on a different aspect of self-regulation would
certainly be informative in future studies for better understanding how these multiple aspects
come together to support behavioral self-regulation. Additional measurement work is also
needed, as it is not clear how these aspects integrate, or develop in relation to each other
across time (Allan & Lonigan, 2014; Miller, Giesbrecht, Müller, McInerney, & Kerns, 2012;
Author Manuscript

Willoughby et al., 2011; Zelazo & Carlson, 2008). Studying the development of multiple
aspects of self-regulation could provide a greater specificity of information, particularly
related to the subset of children who demonstrated later development.

In addition, caution must be taken regarding whether the functional form found within the
current study is specific to the measure used, or is indicative of the latent construct of
behavioral self-regulation. Past evidence and theory (e.g., Diamond, 2002; Vallotton &
Ayoub, 2011) support the current finding that the development of behavioral self-regulation
is non-linear across early childhood, yet more research with multiple measures is needed to
adequately confirm that this is the case and that the functional form of development is not an
artifact of the particular assessment(s) used. Likewise, additional research focused on
functional form differences across the trajectories is necessary. Consistent with typical
GMM practices, the current work used a common functional form (exponential), while
Author Manuscript

allowing rate of change to vary across trajectories. Our findings suggested that some
children may demonstrate a different functional form indicative of a more steady
progression of skill development. There is little past work or theory specifying trajectory
form differences or how mechanisms of development relate to those differences. In short, it
is still not clear whether children may acquire skills differently across trajectories with some
children ‘leaping’ in skills indicating a possible rapid integration of multiple skills into a
behavior, while others slowly build the skills that support self-regulation.

In the current study we focused on a core set of predictors to validate trajectory differences;
more comprehensive work is necessary to determine underlying mechanisms of regulatory
differences. As anticipated, the set of chosen indicators we focused on supported that the
individual differences captured by the three trajectories found per sample are meaningful,
Author Manuscript

yet a host of other indicators will likely provide more information regarding trajectory
differences, as well as indicating what child skills and environments are important to help
children optimize their regulatory development. For example, direct evaluation of the early
contexts provided to children by aspects of parental warmth and responsiveness and their
predictive association with developmental differences in self-regulation is a promising next
step (Grolnick & Farkas, 2002; Landry et al., 2002).

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 22

Likewise there is a need to evaluate genetic and neurological predictors of trajectory

Author Manuscript

differences, particularly given recent findings suggesting that self-regulation and many of
the processes, including neural processes, that underlie it may be genetic in nature (e.g., see
Friedman, et al., 2008; Friedman, Miyake, Robinson & Hewitt, 2011) although still open to
environmental influences via probabilistic epigenesis (see Blair & Raver, 2012; Deater-
Deckard, 2014). This must be considered given the remarkable regularity of patterns across
three diverse samples in the current study.

Conclusions
The present study is an important contribution to our understanding of how self-regulation
develops during early childhood. Specifically, findings indicate that the development of
behavioral self-regulation is exponential in nature. Likewise, there are differences in this
trajectory such that majority of children demonstrate rapid gains across the preschool time
Author Manuscript

period, while a subset of children demonstrated low levels of initial behavioral self-
regulation and later self-regulation development. Differences in what trajectory a child was
likely to follow were linked to different child attributes and background characteristics.
Based on these findings researchers and educators alike should consider carefully how best
to support children’s development of behavioral self-regulation during early childhood,
particularly for children who may be at-risk of making few gains during preschool as early
self-regulation development seemingly places children on a trajectory for later school,
economic, and health successes.

Supplementary Material
Refer to Web version on PubMed Central for supplementary material.
Author Manuscript

Acknowledgments
The research reported here was partially supported by the Institute of Education Sciences, U.S. Department of
Education, through Grant R305A100566 to Oregon State University. The opinions expressed are those of the
authors and do not represent views of the Institute of the U.S. Department of Education. This work was also
supported by the National Institute of Child Health and Human Development and the National Science Foundation
under grant numbers R01 HD27176 and 0111754, respectively. Additional funding was provided by the U.S.
Department of Education, Institute for Education Sciences, Cognition and Student Learning (R305H04013) and the
National Institute of Child Health and Human Development (R01 HD48539).

References
Allan NP, Lonigan CJ. Exploring dimensionality of effortful control using hot and cool tasks in a
sample of preschool children. Journal of Experimental Child Psychology. 2014; 122:33–47. DOI:
10.1016/j.jecp.2013.11.013 [PubMed: 24518050]
Author Manuscript

Asparouhov T, Muthén B. Auxiliary variables in mixture modeling: Three-step approaches using M

plus. Structural Equation Modeling: A Multidisciplinary Journal. 2014; 21:329–341. DOI:
10.1080/10705511.2014.915181
Barkley, R. Attention-Deficit/Hyperactivity disorder, self-regulation, and executive functioning. In:
Vohs, KD.; Baumeister, RF., editors. Handbook of self-regulation, second edition: Research, theory,
and applications. 2. New York, NY: Guilford Press; 2011. p. 551-564.
Benjamini Y, Hochberg Y. Controlling the false discovery rate: A practical and powerful approach to
multiple testing. Journal of the Royal Statistical Society. Series B (Methodological). 1995:289–300.

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 23

Bergman, LR.; Magnusson, D.; Khouri, BME. Studying individual development in an interindividual
context: A person-oriented approach. Mahwah, NJ: Psychology Press; 2002.
Author Manuscript

Best JR, Miller PH. A developmental perspective on executive function. Child Development. 2010;
81:1641–1660. DOI: 10.1111/j.1467-8624.2010.01499.x [PubMed: 21077853]
Bindman SW, Hindman AH, Bowles RP, Morrison FJ. The contributions of parental management
language to self-regulation in preschool children. Early Childhood Research Quarterly. 2013;
28:529–539. DOI: 10.1016/j.ecresq.2013.03.003 [PubMed: 23997425]
Blair C. School readiness: Integrating cognition and emotion in a neurobiological conceptualization of
children’s functioning at school entry. American Psychologist. 2002; 57:111–127. DOI:
10.1037/0003-066X.57.2.111 [PubMed: 11899554]
Blair C. Stress and the development of self-regulation in context. Child Development Perspectives.
2010; 4:181–188. DOI: 10.1111/j.1750-8606.2010.00145.x [PubMed: 21779305]
Blair C, Granger DA, Willoughby M, Mills-Koonce R, Cox M, Greenberg MT, … Fortunato CK.
Salivary cortisol mediates effects of poverty and parenting on executive functions in early
childhood. Child Development. 2011; 82:1970–1984. DOI: 10.1111/j.1467-8624.2011.01643.x
[PubMed: 22026915]
Author Manuscript

Blair C, Raver CC. Child development in the context of adversity: Experiential canalization of brain
and behavior. American Psychologist. 2012; 67:309–318. DOI: 10.1037/a0027493 [PubMed:
22390355]
Blair C, Raver CC. School readiness and self-regulation: A developmental psychobiological approach.
Annual Review of Psychology. 2015; 66:711–731. DOI: 10.1146/annurev-psych-010814-015221
Blandon AY, Calkins SD, Keane SP, O’Brien M. Individual differences in trajectories of emotion
regulation processes: The effects of maternal depressive symptomatology and children’s
physiological regulation. Developmental Psychology. 2008; 44:1110.doi:
10.1037/0012-1649.44.4.1110 [PubMed: 18605838]
Bodrova, E.; Leong, DJ. Tools of the mind: The Vygotskian approach to early childhood education.
Columbus, OH: Merrill/Prentice Hall; 2007.
Bohlmann NL, Maier MF, Palacios N. Bidirectionality in self-regulation and expressive vocabulary:
Comparisons between monolingual and dual language learners in preschool. Child Development.
2015; 86:1094–1111. DOI: 10.1111/cdev.12375
Author Manuscript

Bowles, RP.; Montroy, JJ. Latent growth curve modeling using structural equation modeling. In:
Petscher, Y.; Schatschneider, C.; Compton, DL., editors. Applied quantitative analysis in the social
sciences. New York, NY: Routledge; 2013. p. 265-303.
Bowles RP, Pentimonti JM, Gerde HK, Montroy JJ. Item response analysis of uppercase and lowercase
letter name knowledge. Journal of Psychoeducational Assessment. 2013; 32:146–156. DOI:
10.1177/0734282913490266
Bradley RH, Corwyn RF. Socioeconomic status and child development. Annual Review of Psychology.
2002; 53:371–399. DOI: 10.1146/annurev.psych.53.100901.135233
Brooks-Gunn J, Duncan GJ. The effects of poverty on children. Future Child. 1997; 7:55–71. DOI:
10.2307/1602387 [PubMed: 9299837]
Bronson, MB. Self-Regulation in early childhood: Nature and nurture. New York, NY: Guilford Press;
2000.
Brownell, R. Expressive one-word picture vocabulary test - 2000 edition. Novato, CA: Academic
Therapy Publications; 2000.
Calkins, SD. The emergence of self-regulation: Biological and behavioral control mechanisms
Author Manuscript

supporting toddler competencies. In: Brownell, CA.; Kopp, CB., editors. Socioemotional
development in the toddler years: Transitions and transformations. New York, NY: Guilford Press;
2007. p. 261-284.
Calkins SD, Dedmon SE, Gill KL, Lomax LE, Johnson LM. Frustration in infancy: Implications for
emotion regulation, physiological processes, and temperament. Infancy. 2002; 3:175–197. DOI:
10.1207/S15327078IN0302_4
Cameron Ponitz CE, McClelland MM, Jewkes AM, Connor CM, Farris CL, Morrison FJ. Touch your
toes! Developing a direct measure of behavioral regulation in early childhood. Early Childhood
Research Quarterly. 2008; 23:141–158. DOI: 10.1016/j.ecresq.2007.01.004

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 24

Cameron Ponitz CE, McClelland MM, Matthews JS, Morrison FJ. A structured observation of
behavioral self-regulation and its contribution to kindergarten outcomes. Developmental
Author Manuscript

Psychology. 2009; 45:605–619. DOI: 10.1037/a0015365 [PubMed: 19413419]

Cameron Ponitz CE, Rimm-Kaufman SE, Brock LL, Nathanson L. Early adjustment, gender
differences, and classroom organizational climate in first grade. The Elementary School Journal.
2009; 110:142–162. DOI: 10.1086/605470
Carlson SM, Moses LJ, Breton C. How specific is the relation between executive function and theory
of mind? Contributions of inhibitory control and working memory. Infant and Child Development.
2002; 11:73–92. DOI: 10.1002/icd.298
Chang H, Shaw DS, Dishion TJ, Gardner F, Wilson MN. Direct and indirect effects of the family
check-up on self-regulation from toddlerhood to early school-age. Journal of Abnormal Child
Psychology. 2014; 42:1117–1128. [PubMed: 24599383]
Clark CA, Sheffield TD, Chevalier N, Nelson JM, Wiebe SA, Espy KA. Charting early trajectories of
executive control with the shape school. Developmental Psychology. 2013; 49:1481–1493. DOI:
10.1037/a0030578 [PubMed: 23106846]
Cole, PM.; Armstrong, LM.; Pemberton, CK. The role of language in the development of emotion
Author Manuscript

regulation. In: Calkins, SD.; Bell, M., editors. Child development at the intersection of emotion
and cognition: Human brain development. Washington, DC: American Psychological Association;
2010. p. 59-77.
Connor CM, Ponitz CC, Phillips BM, Travis QM, Glasney S, Morrison FJ. First graders’ literacy and
self-regulation gains: The effect of individualizing student instruction. Journal of School
Psychology. 2010; 48:433–455. DOI: 10.1016/j.jsp.2010.06.003 [PubMed: 20728691]
Corrigan P. On moral regulation: Some preliminary remarks. The Sociological Review. 1981; 29:313–
337. DOI: 10.1111/j.1467-954X.1981.tb00176.x
Da Silva S, Moreira B, Da Costa N Jr. 2D: 4D digit ratio predicts delay of gratification in preschoolers.
PloS One. 2014; 9:e114394. [PubMed: 25490040]
Deater-Deckard K. Family matters intergenerational and interpersonal processes of executive function
and attentive behavior. Current Directions in Psychological Science. 2014; 23:230–236. DOI:
10.1177/0963721414531597 [PubMed: 25197171]
Diamond, A. Normal development of prefrontal cortex from birth to young adulthood: Cognitive
functions, anatomy, and biochemistry. In: Stuss, DT.; Knight, RT., editors. Principles of frontal
Author Manuscript

lobe function. New York, NY: Oxford University Press; 2002. p. 466-503.
Diamond A, Barnett WS, Thomas J, Munro S. Preschool program improves cognitive control. Science.
2007; 318:1387–1388. DOI: 10.1126/science.1151148 [PubMed: 18048670]
Diamond A, Prevor MB, Callender G, Druin DP. Prefrontal cortex cognitive deficits in children treated
early and continuously for PKU. Monographs of the Society for Research in Child Development.
1997; 62:1–205. DOI: 10.2307/1166208 [PubMed: 9353949]
Dickinson, DK.; McCabe, A.; Essex, MJ. A window of opportunity we must open to all: The case for
preschool with high-quality support for language and literacy. In: Dickinson, DK.; Neuman, SB.,
editors. Handbook of early literacy research. Vol. 2. New York, NY: Guilford Press; 2006. p.
11-28.
Enders, CK. Applied missing data analysis. New York, NY: Guilford Press; 2010.
Entwisle DR, Alexander KL, Olson LS. Early schooling: The handicap of being poor and male.
Sociology of Education. 2007; 80:114–138. DOI: 10.1177/003804070708000202
Fernald A, Marchman VA, Weisleder A. SES differences in language processing skill and vocabulary
Author Manuscript

are evident at 18 months. Developmental Science. 2013; 16:234–248. DOI: 10.1111/desc.12019

[PubMed: 23432833]
Fox L, Dunlap G, Cushing L. Early intervention, positive behavior support, and transition to school.
Journal of Emotional and Behavioral Disorders. 2002; 10:149–157. DOI:
10.1177/10634266020100030301
Friedman NP, Miyake A, Robinson JL, Hewitt JK. Developmental trajectories in toddlers’ self-restraint
predict individual differences in executive functions 14 years later: a behavioral genetic analysis.
Developmental Psychology. 2011; 47:1410–1430. DOI: 10.1037/a0023750 [PubMed: 21668099]

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 25

Friedman NP, Miyake A, Young SE, DeFries JC, Corley RP, Hewitt JK. Individual differences in
executive functions are almost entirely genetic in origin. Journal of Experimental Psychology:
Author Manuscript

General. 2008; 137:201–225. DOI: 10.1037/0096-3445.137.2.201 [PubMed: 18473654]

Garon N, Bryson SE, Smith IM. Executive function in preschoolers: A review using an integrative
framework. Psychological Bulletin. 2008; 134:31–60. DOI: 10.1037/0033-2909.134.1.31
[PubMed: 18193994]
Gerstadt CL, Hong YJ, Diamond A. The relationship between cognition and action: Performance of
children 3–7 years old on a stroop-like day-night test. Cognition. 1994; 53:129–153. DOI:
10.1016/0010-0277(94)90068-X [PubMed: 7805351]
Graham JW. Adding missing-data-relevant variables to FIML-based structural equation models.
Structural Equation Modeling. 2003; 10:80–100. DOI: 10.1207/S15328007SEM1001_4
Grimm, KJ.; McArdle, JJ.; Hamagami, F. Nonlinear growth mixture models in research on cognitive
aging. In: van Montfort, K.; Oud, J.; Satorra, A., editors. Longitudinal models in the behavioral
and related sciences. Mahwah, NJ: Erlbaum; 2007. p. 267-295.
Grimm KJ, Ram N, Estabrook R. Nonlinear structured growth mixture models in M plus and OpenMx.
Multivariate behavioral research. 2010; 45:887–909. DOI: 10.1080/00273171.2010.531230
Author Manuscript

[PubMed: 25419006]
Grolnick, WS.; Farkas, M. Parenting and the development of children’s self-regulation. In: Bornstein,
MH., editor. Handbook of parenting: Practical issues in parenting. 2. Vol. 5. Mahwah, NJ:
Lawrence Erlbaum Associates Publishers; 2002. p. 89-110.
Gross, JJ.; Thompson, RA. Emotion regulation: Conceptual foundations. In: Gross, JJ., editor.
Handbook of emotion regulation. New York, NY: Guilford Press; 2007. p. 3-24.
Guttentag CL, Pedrosa-Josic C, Landry SH, Smith KE, Swank PR. Individual variability in parenting
profiles and predictors of change: Effects of an intervention with disadvantaged mothers. Journal
of Applied Developmental Psychology. 2006; 27:349–369. DOI: 10.1016/j.appdev.2006.04.005
Hart, B.; Risley, T. Meaningful differences in the everyday experience of young American children.
Baltimore, MD: Brookes Publishing; 1995.
Hindman AH, Morrison FJ. Differential contributions of three parenting dimensions to preschool
literacy and social skills in a middle-income sample. Merrill-Palmer Quarterly. 2012; 58:191–223.
DOI: 10.1353/mpq.2012.0012
Author Manuscript

Hoff, E.; Laursen, B.; Tardif, T. Socioeconomic status and parenting. In: Bornstein’s, MH., editor.
Handbook of parenting: Biology and ecology of parenting. 2. Vol. 2. Mahwah, NJ: Erlbaum; 2002.
p. 231-252.
Howse RB, Calkins SD, Anastopoulos AD, Keane SP, Shelton TL. Regulatory contributors to
children’s kindergarten achievement. Early Education & Development. 2003; 14:101–120. DOI:
10.1207/s15566935eed1401_7
Kagan, SL.; Kauerz, K.; Tarrant, K. The early care and education teaching workforce at the fulcrum:
An agenda for reform. Washington, DC: Teachers College Press; 2007.
Kazak AE. Fathers’ and mothers’ parenting behavior and beliefs as predictors of children’s social
adjustment in the transition to school. Journal of Family Psychology. 2004; 18:628–638. DOI:
10.1037/0893-3200.18.4.628 [PubMed: 15598168]
Kim S, Nordling JK, Yoon JE, Boldt LJ, Kochanska G. Effortful control in “hot” and “cool” tasks
differentially predicts children’s behavior problems and academic performance. Journal of
Abnormal Child Psychology. 2013; 41:43–56. DOI: 10.1007/s10802-012-9661-4 [PubMed:
22798038]
Author Manuscript

Kochanska G, Coy KC, Murray KT. The development of self-regulation in the first four years of life.
Child Development. 2001; 72:1091–1111. DOI: 10.1111/1467-8624.00336 [PubMed: 11480936]
Kopp CB. Antecedents of self-regulation: A developmental perspective. Developmental Psychology.
1982; 18:199–214. DOI: 10.1037/0012-1649.18.2.199
Kopp CB. Regulation of distress and negative emotions: A developmental view. Developmental
Psychology. 1989; 25:343–354. DOI: 10.1037/0012-1649.25.3.343
Kreuter F, Muthén BO. Analyzing criminal trajectory profiles: Bridging multilevel and group-based
approaches using growth mixture modeling. Journal of Quantitative Criminology. 2008; 24:1–31.
DOI: 10.1007/s10940-007-9036-0

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 26

Lan X, Legare CH, Ponitz CC, Li S, Morrison FJ. Investigating the links between the subcomponents
of executive function and academic achievement: A cross-cultural analysis of Chinese and
Author Manuscript

American preschoolers. Journal of Experimental Child Psychology. 2011; 108:677–692. DOI:

10.1016/j.jecp.2010.11.001 [PubMed: 21238977]
Landry SH, Miller-Loncar CL, Smith KE, Swank PR. The role of early parenting in children’s
development of executive processes. Developmental Neuropsychology. 2002; 21:15–41. DOI:
10.1207/S15326942DN2101_2 [PubMed: 12058834]
Landry SH, Smith KE, Swank PR. Responsive parenting: Establishing early foundations for social,
communication, and independent problem-solving skills. Developmental Psychology. 2006;
42:627–642. DOI: 10.1037/0012-1649.42.4.627 [PubMed: 16802896]
Lengua LJ, Moran L, Zalewski M, Ruberry E, Kiff C, Thompson S. Relations of growth in effortful
control to family income, cumulative risk, and adjustment in preschool-age children. Journal of
Abnormal Child Psychology. 2015; 43:705–720. DOI: 10.1007/s10802-014-9941-2 [PubMed:
25253079]
Lerner RM, Overton WF. Exemplifying the integrations of the relational developmental system:
Synthesizing theory, research, and application to promote positive development and social justice.
Author Manuscript

Journal of Adolescent Research. 2008; 23:245–255.

Lipsey, MW.; Nesbitt, KT.; Farran, DC.; Dong, N.; Fuhs, MW.; Wilson, SJ. Learning-related cognitive
self-regulation measures for prekindergarten children with predictive validity for academic
achievement (Working Paper). Peabody Research Institute at Vanderbilt University website. 2014.
https://my.vanderbilt.edu/cogselfregulation/files/2012/11/Self-Reg-summary-paper-5-7-141.pdf
Lonigan, CJ.; Wagner, RK.; Torgesen, JK.; Rashotte, CA. TOPEL: Test of preschool early literacy.
Austin, TX: PRO-ED; 2007.
Lonigan CJ, Phillips BM, Clancy JL, Landry SH, Swank PR, Assel M, … Eisenberg N. Impacts of a
comprehensive school readiness curriculum for preschool children at risk for educational
difficulties. Child Development. 2015; 86:1773–1793. DOI: 10.1111/cdev.12460 [PubMed:
26510099]
Martel MM, Roberts BA. Prenatal testosterone increases sensitivity to prenatal stressors in males with
disruptive behavior disorders. Neurotoxicology and Teratology. 2014; 44:11–17. DOI: 10.1016/
j.ntt.2014.05.001 [PubMed: 24819590]
Matthews JS, Marulis LM, Williford AP. Gender processes in school functioning and the mediating
Author Manuscript

role of cognitive self-regulation. Journal of Applied Developmental Psychology. 2014; 35:128–

137. DOI: 10.1016/j.appdev.2014.02.003
Matthews JS, Ponitz CC, Morrison FJ. Early gender differences in self-regulation and academic
achievement. Journal of Educational Psychology. 2009; 101:689–704. DOI: 10.1037/a0014240
McArdle JJ. Latent variable growth within behavior genetic models. Behavior Genetics. 1986; 16:163–
200. DOI: 10.1007/BF01065485 [PubMed: 3707483]
McArdle, JJ.; Bell, RQ. An introduction to latent growth models for developmental data analysis. In:
Little, TD.; Schnabel, KU., editors. Modeling longitudinal and multilevel data: Practical issues,
applied approaches and specific examples. Mahwah, NJ: Lawrence Erlbaum Associates; 2000. p.
69-107.
McArdle, JJ.; Nesselroade, JR. Growth curve analysis in contemporary psychological research. In:
Schinka, JA.; Velicer, WF., editors. Handbook of psychology: Research methods in psychology.
New York, NY: Wiley; 2003. p. 447-480.
McClelland MM, Acock AC, Morrison FJ. The impact of kindergarten learning-related skills on
Author Manuscript

academic trajectories at the end of elementary school. Early Childhood Research Quarterly. 2006;
21:471–490. DOI: 10.1016/j.ecresq.2006.09.003
McClelland MM, Acock AC, Piccinin A, Rhea SA, Stallings MC. Relations between preschool
attention span-persistence and age 25 educational outcomes. Early Childhood Research Quarterly.
2013; 28:314–324. DOI: 10.1016/j.ecresq.2012.07.008 [PubMed: 23543916]
McClelland MM, Cameron CE. Self-regulation in early childhood: Improving conceptual clarity and
developing ecologically valid measures. Child Development Perspectives. 2012; 6:136–142. DOI:
10.1111/j.1750-8606.2011.00191.x

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 27

McClelland MM, Cameron CE, Connor CM, Farris CL, Jewkes AM, Morrison FJ. Links between
behavioral regulation and preschoolers’ literacy, vocabulary, and math skills. Developmental
Author Manuscript

Psychology. 2007; 43:947–959. DOI: 10.1037/0012-1649.43.4.947 [PubMed: 17605527]

McClelland MM, Cameron CE, Duncan R, Bowles RP, Acock AC, Miao A, Pratt ME. Predictors of
early growth in academic achievement: The head-toes-knees-shoulders task. Frontiers in
Psychology. 2014; 5:1–14. DOI: 10.3389/fpsyg.2014.00599 [PubMed: 24474945]
McClelland, MM.; Cameron Ponitz, CE.; Messersmith, E.; Tominey, SL. The integration of cognition
and emotion. In: Lerner, RM.; Overton, WF., editors. Handbook of lifespan human development.
4. Hoboken, NJ: Wiley & Sons; 2010. p. 509-553.
McClelland MM, Morrison FJ, Holmes DL. Children at risk for early academic problems: The role of
learning-related social skills. Early Childhood Research Quarterly. 2000; 15:307–329. DOI:
10.1016/S0885-2006(00)00069-7
McClelland MM, Wanless SB. Growing up with assets and risks: The importance of self-regulation for
academic achievement. Research in Human Development. 2012; 9:278–297. DOI:
10.1080/15427609.2012.729907
Meredith W, Tisak J. Latent curve analysis. Psychometrika. 1990; 55:107–122. DOI: 10.1007/
Author Manuscript

BF02294746
Miech R, Essex MJ, Goldsmith HH. Socioeconomic status and the adjustment to school: The role of
self-regulation during early childhood. Sociology of Education. 2001; 74:102–120.
Miller MR, Giesbrecht GF, Müller U, McInerney RJ, Kerns KA. A latent variable approach to
determining the structure of executive function in preschool children. Journal of Cognition and
Development. 2012; 13:395–423. DOI: 10.1080/15248372.2011.585478
Mischel W, Ayduk O, Berman MG, Casey BJ, Gotlib IH, Jonides J, … Shoda Y. “Willpower” over the
life span: Decomposing self-regulation. Social Cognitive and Affective Neuroscience. 2011;
6:252–256. DOI: 10.1093/scan/nsq081 [PubMed: 20855294]
Moffitt TE, Arseneault L, Belsky D, Dickson N, Hancox RJ, Harrington H, … Caspi A. A gradient of
childhood self-control predicts health, wealth, and public safety. Proceedings of the National
Academy of Sciences. 2011; 108:2693–2698. DOI: 10.1073/pnas.1010076108
Montroy JJ, Bowles RP, Skibbe LE, Foster TD. Social skills and problem behaviors as mediators of the
relationship between behavioral self-regulation and academic achievement. Early Childhood
Research Quarterly. 2014; 29:298–309. DOI: 10.1016/j.ecresq.2014.03.002
Author Manuscript

Muthén, BO. Latent variable mixture modeling. In: Marcoulides, GA.; Schumacker, RE., editors. New
developments and techniques in structural equation modeling. Mahwah, NJ: Lawrence Erlbaum
Associates; 2001. p. 1-33.
Muthén BO, Muthén LK. Integrating person-centered and variable-centered analyses: Growth mixture
modeling with latent trajectory classes. Alcoholism: Clinical and Experimental Research. 2000;
24:882–891. DOI: 10.1111/j.1530-0277.2000.tb02070.x
Muthén, LK.; Muthén, BO. Mplus user’s guide. 6. Los Angeles, CA: Muthén & Muthén; 1998–2010.
Nagin DS. Analyzing developmental trajectories: A semiparametric, group-based approach.
Psychological Methods. 1999; 4:139–157. DOI: 10.1037/1082-989X.4.2.139
Posner MI, Rothbart MK. Developing mechanisms of self-regulation. Development and
Psychopathology. 2000; 12:427–441. [PubMed: 11014746]
Ready DD, LoGerfo LF, Burkam DT, Lee VE. Explaining girls’ advantage in kindergarten literacy
learning: Do classroom behaviors make a difference? The Elementary School Journal. 2005;
106:21–38. DOI: 10.1086/496905
Author Manuscript

Rimm-Kaufman SE, Pianta RC, Cox MJ. Teachers’ judgments of problems in the transition to
kindergarten. Early Childhood Research Quarterly. 2000; 15:147–166. DOI: 10.1016/
S0885-2006(00)00049-1
Roben CK, Cole PM, Armstrong LM. Longitudinal relations among language skills, anger expression,
and regulatory strategies in early childhood. Child Development. 2013; 84:891–905. DOI:
10.1111/cdev.12027 [PubMed: 23278601]
Rogosa, D. Myths about longitudinal research. In: Schaie, KW.; Campbell, RT.; Meredith, W.;
Rawlings, SC., editors. Methodological issues in aging research. New York, NY: Springer
Publishing Co; 1988. p. 171-209.

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 28

Rothbart, MK.; Posner, MI.; Kieras, J. Temperament, attention, and the development of self-regulation.
In: McCartney, K.; Phillips, D., editors. Blackwell handbook of early childhood development.
Author Manuscript

Malden: Blackwell Publishing; 2006. p. 338-357.

Rothbart MK, Ellis LK, Rosario Rueda M, Posner MI. Developing mechanisms of temperamental
effortful control. Journal of Personality. 2003; 71:1113–1144. DOI: 10.1111/1467-6494.7106009
[PubMed: 14633060]
Rueda, MR.; Posner, MI.; Rothbart, MK. Attentional control and self-regulation. In: Vohs, KD.;
Baumeister, RF., editors. Handbook of self-regulation: Research, theory, and applications. 2. New
York NY: Guilford Press; 2004. p. 284-299.
Schmitt SA, McClelland MM, Tominey SL, Acock AC. Strengthening school readiness for Head Start
children: Evaluation of a self-regulation intervention. Early Childhood Research Quarterly. 2015;
30:20–31. DOI: 10.1016/j.ecresq.2014.08.001
Schmitt SA, Pratt ME, McClelland MM. Examining the validity of behavioral self-regulation tools in
predicting preschoolers’ academic achievement. Early Education and Development. 2014;
25:641–660. DOI: 10.1080/10409289.2014.850397
Schunk DH, Zimmerman BJ. Social origins of self-regulatory competence. Educational Psychologist.
Author Manuscript

1997; 32:195–208. DOI: 10.1207/s15326985ep3204_1

Singer, JD.; Willett, JB. Applied longitudinal data analysis: Modeling change and event occurrence.
New York, NY: Oxford University Press; 2003.
Skibbe LE, Grimm KJ, Bowles RP, Morrison FJ. Literacy growth in the academic year versus summer
from preschool through second grade: Differential effects of schooling across four skills.
Scientific Studies of Reading. 2012; 16:141–165. DOI: 10.1080/10888438.2010.543446
[PubMed: 23645979]
Stifter CA, Braungart JM. The regulation of negative reactivity in infancy: Function and development.
Developmental Psychology. 1995; 31:448–455. DOI: 10.1037/0012-1649.31.3.448
Stipek, D. Classroom practices and children’s motivation to learn. In: Zigler, E.; Gillian, WS.; Barnett,
SW., editors. The pre-K debates: Current controversies and issues. Vol. 2011. Baltimore, MD,
US: Paul H Brookes Publishing; 2011. p. 98-103.
Tofighi, D.; Enders, CK. Identifying the correct number of classes in growth mixture models. In:
Hancock, GR., editor. Mixture models in latent variable research. Greenwich, CT: Information
Age; 2007. p. 317-341.
Author Manuscript

Tominey SL, McClelland MM. Red light, purple light: Findings from a randomized trial using circle
time games to improve behavioral self-regulation in preschool. Early Education & Development.
2011; 22:489–519. DOI: 10.1080/10409289.2011.574258
Vallotton C, Ayoub C. Use your words: The role of language in the development of toddlers’ self-
regulation. Early Childhood Research Quarterly. 2011; 26:169–181. DOI: 10.1016/j.ecresq.
2010.09.002 [PubMed: 21969766]
Vohs, KD.; Baumeister, RF. Handbook of self-regulation, second edition: Research, theory, and
applications. New York, NY: Guilford Press; 2011.
von Eye A, Bergman LR. Research strategies in developmental psychopathology: Dimensional identity
and the person-oriented approach. Development and Psychopathology. 2003; 15:553–580. DOI:
10.1017/S0954579403000294 [PubMed: 14582932]
von Suchodoletz A, Gestsdottir S, Wanless SB, McClelland MM, Birgisdottir F, Gunzenhauser C,
Ragnarsdottir H. Behavioral self-regulation and relations to emergent academic skills among
children in Germany and Iceland. Early Childhood Research Quarterly. 2013; 28:62–73. DOI:
Author Manuscript

10.1016/j.ecresq.2012.05.003
Vygotsky, LS. Thought and language. Kozulin, A., translator. Cambridge, MA: The MIT Press;
1934/1986.
Wachs, TD. The nature of nurture. Vol. 3. Newbury Park, CA: Sage Publications; 1992.
Wachs TD, Gurkas P, Kontos S. Predictors of preschool children’s compliance behavior in early
childhood classroom settings. Journal of Applied Developmental Psychology. 2004; 25:439–457.
Wang L, Zhang Z, McArdle JJ, Salthouse TA. Investigating ceiling effects in longitudinal data
analysis. Multivariate behavioral research. 2008; 43:476–496. DOI:
10.1080/00273170802285941

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 29

Wanless SB, Kim KH, Zhang C, Degol JL, Chen JL, Chen FM. Trajectories of behavioral regulation
for Taiwanese children from 3.5 to 6 years and relations to math and vocabulary outcomes. Early
Author Manuscript

Childhood Research Quarterly. 2016; 34:104–114. DOI: 10.1016/j.ecresq.2015.10.001

Wanless SB, McClelland MM, Acock AC, Cameron Ponitz C, Son S-H, Lan X, Morrison FJ, Chen J-
L, Chen F-M, Lee K, Sung M, Li S. Measuring behavioral regulation in four cultures.
Psychological Assessment. 2011; 23:364–378. DOI: 10.1037/a0021768 [PubMed: 21381840]
Wiebe SA, Sheffield T, Nelson JM, Clark CAC, Chevalier N, Espy KA. The structure of executive
function in 3-year-olds. Journal of Experimental Child Psychology. 2011; 108:436–452. DOI:
10.1016/j.jecp.2010.08.008 [PubMed: 20884004]
Willoughby M, Kupersmidt J, Voegler-Lee M, Bryant D. Contributions of hot and cool self-regulation
to preschool disruptive behavior and academic achievement. Developmental Neuropsychology.
2011; 36:162–180. DOI: 10.1080/87565641.2010.549980 [PubMed: 21347919]
Willoughby MT, Magnus B, Vernon-Feagans L, Blair CB. Family Life Project Investigators.
Developmental delays in executive function from 3 to 5 years of age predict kindergarten
academic readiness. Journal of Learning Disabilities. 2016; Advamce online publication. doi:
10.1177/0022219415619754
Author Manuscript

Wilson SB, Lonigan CJ. Identifying preschool children at risk of later reading difficulties: Evaluation
of two emergent literacy screening tools. Journal of Learning Disabilities. 2009; 43:62–76. DOI:
10.1177/0022219409345007 [PubMed: 19822699]
Winsler, A.; Fernyhough, C.; Montero, I., editors. Private speech, executive functioning, and the
development of verbal self-regulation. Cambridge: Cambridge University Press; 2009.
Woodcock, RW.; Mather, N. WJ-III: Woodcock Johnson III tests of achievement: Examiner’s manual.
Itasca, IL: Riverside; 2001.
Zelazo PD, Carlson SM. Hot and cool executive function in childhood and adolescence: Development
and plasticity. Child Development Perspectives. 2012; 6:354–360. DOI: 10.1111/j.
1750-8606.2012.00246.x
Zelazo, PD.; Carlson, SM.; Kesek, A. The development of executive function in childhood. In: Nelson,
CA.; Luciana, M., editors. Handbook of developmental cognitive neuroscience. 2. Cambridge,
MA: MIT Press; 2008. p. 553-574.
Zelazo, PD.; Müller, U. Executive function in typical and atypical development. In: Goswami’s, U.,
editor. The Wiley-Blackwell handbook of childhood cognitive development. 2. Hoboken, NJ:
Author Manuscript

Wiley & Sons; 2010. p. 574-603.

Zelazo P, Müller U, Frye D, Marcovitch S, Argitis G, Boseovski J, … Sutherland A. The development
of executive function in early childhood. Monographs of the Society for Research in Child
Development. 2003; 68:11–43. DOI: 10.1111/j.0037-976X.2003.00260.x
Author Manuscript

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 30
Author Manuscript
Author Manuscript

Figure 1.
Random subset of 25 children per sample’s (75 total) smoothed behavioral self-regulation
trajectories
Author Manuscript
Author Manuscript

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 31
Author Manuscript
Author Manuscript

Figure 2.
Latent growth curve model of the developmental trajectory of behavioral self-regulation by
sample.
Author Manuscript
Author Manuscript

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Montroy et al. Page 32
Author Manuscript
Author Manuscript

Figure 3.
Predicted developmental trajectories for behavioral self-regulation by sample. A. MLSELD
preschool sample, B. Oregon sample, C. Michigan longitudinal sample, D. all samples
picture together; MI = Michigan.
Author Manuscript
Author Manuscript

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 1
Descriptive Statistics for Self-regulation by Sample and Age

Variable Michigan longitudinal MLSELD preschool Oregon

Montroy et al.

Self-regulation N M SD N M SD N M SD
-HTKS age 39 mos. or less - - - 103 3.66 7.70 - - -
-HTKS age 40 – 42 mos. 98 4.98 8.92 128 5.50 9.95 - - -
-HTKS age 43 – 45 mos. 101 8.92 12.26 144 7.25 11.38 - - -
-HTKS age 46 – 48 mos. 111 12.95 14.81 185 9.15 11.32 - - -
-HTKS age 49 – 51 mos. 111 18.21 15.16 241 11.95 13.49 - - -
-HTKS age 52 – 54 mos. 153 20.48 15.22 276 16.73 14.27 157 12.44 13.11
-HTKS age 55 – 57 mos. 151 21.67 14.58 293 17.63 14.86 161 16.53 13.32
-HTKS age 58 – 60 mos. 152 25.19 13.46 217 22.78 14.28 201 18.74 14.20
-HTKS age 61 – 63 mos. 128 28.29 10.61 165 25.28 14.01 186 21.78 14.40
-HTKS age 64 – 66 mos. 140 29.51 10.78 - - - 174 23.15 13.54
-HTKS age 67 – 69 mos. 154 31.20 8.90 - - - 140 26.61 13.01
-HTKS age 70 – 72 mos. 110 33.49 6.97 - - - 162 28.85 10.47
-HTKS age 73 – 75 mos. 108 35.58 4.85 - - - 97 30.32 11.84
-HTKS age 76 – 78 mos. 118 35.37 5.31 - - - 95 30.47 10.61
-HTKS age 79 – 81 mos. 113 37.05 3.08 - - - - - -
-HTKS age 82 – 84 mos. 80 36.60 4.56 - - - - - -
-HTKS age 85 mos. or more 61 37.55 3.30 - - - - - -

Note. HTKS refers to the Head-Toes-Knees-Shoulders task. Mos. refers to months. Dashes represent ages that data were not collected by sample.

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

Page 33
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 2
Summary of Latent Growth Model Fit Statistics

Models -2 Log Likelihood Free parameters AIC aBIC

Michigan longitudinal sample
Montroy et al.

- Linear 7071.40 6 14154.81 14158.94

- Quadratic 6961.99 10 13943.97 13950.86
- Modified logistic 7024.69 8 14065.38 14070.88
- Latent basis 7184.79 19 14407.57 14420.65
- Exponential 6934.36 10 13888.71 13895.59
MLSELD preschool sample
- Linear 6688.04 6 13588.08 13595.82
- Quadratic 6760.81 10 13541.63 13554.52
- Modified Logistic 6783.97 8 13583.95 13594.26
- Latent Basis 6779.69 13 13585.38 13602.14
- Exponential 6760.21 10 13540.42 13553.32
Oregon sample
- Linear 5216.08 6 10444.16 10448.97
- Quadratic 5192.96 10 10405.91 10413.92
- Modified logistic 5208.42 8 10432.85 10439.25
- Latent basis 5202.44 13 10430.88 10441.29
- Exponential 5189.12 10 10398.23 10406.24

Note. AIC refers to the Akaike information criterion, aBIC refers to the adjusted Bayesian information criterion. Bolded values indicate best fit.

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

Page 34
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 3
Summary of Growth Mixture Model Fit Statistics

Models Michigan longitudinal MLSELD preschool Oregon

Montroy et al.

AIC aBIC Ent. BLRT AIC aBIC Ent. BLRT AIC aBIC Ent. BLRT
2 Trajectory
- Means only 13897 13902 .81 <.01 13558 13567 .68 <.01 10445 10451 .83a <.01

- Means + shape 13948 13954 .79 <.01 13521 13531 .70 <.01 10218 10229 .83a <.01

3 Trajectory
- Means only 14032 14039 .65 .99 13428 13441 .63 <.01 10260 10270 .80 0.99
- Means + shape 13832 13840 .74 <.01 13378 13393 .62 <.01 10212 10221 .82 <.01
- Within level free 13833 13842 .74 <.01 13380 13396 .61 <.01b 10218 10230 .68 <.01

- Within level + slope free - - - - 13613 13640 1.00a 1.00 10455 10469 1.00a 1.00b
- All within free 13905 13917 .99a <.01 - - - - - - - -

4 Trajectory
- Means only 13994 14003 .67 .38 13430 13444 .71a <.01 - - - -

Note. Dashes (-) denote that the model did not converge, an indicator of poor fit. Means only model: only level and growth means were allowed to vary between trajectories; means + shape: basis
coefficients were allowed to vary (i.e., shape of curve) between trajectories. The “3 trajectory” models also included extra models to test whether relaxing within trajectory constraints resulted in better
model fit. AIC refers to the Akaike information criterion, aBIC refers to the adjusted Bayesian information criterion. Bolded values indicate best fit. Ent. refers to entropy values, and BLRT indicates the p-
value from the bootstrapped likelihood ratio tests.
a
Included non-positive definite warnings that affected model trustworthiness, class specification and related entropy values.
b
Bootstrap LRT p value may not be trustworthy

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

Page 35
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 4
Summary of Behavioral Self-regulation Trajectory Descriptives

Variables Gender Language Mother education

Mean (SD) Range Mean (SD) Range Mean (SD) Range
Montroy et al.

MLSELD preschool
- Later 0.44 (.50) 0–1 41.72 (15.09) 1–67 5.74 (1.95) 2–10
- Intermediate 0.55 (.50) 0–1 44.04 (15.39) 1–69 6.63 (2.00) 2–11
- Early 0.51 (.50) 0–1 52.52 (10.20) 23–69 7.49 (2.00) 3–11
Oregon
- Later 0.49 (.40) 0–1 455.42 (18.14) 384–495 3.23 (1.96) 1–11
- Intermediate 0.41 (.49) 0–1 465.54 (11.89) 425–498 5.13 (2.81) 1–11
- Early 0.57 (.50) 0–1 473.51 (9.39) 450–501 5.99 (2.91) 1–11
Michigan longitudinal
- Later 0.42 (.50) 0–1 466.12 (15.94) 398–498 7.16 (1.74) 2–9
- Intermediate 0.57 (.50) 0–1 467.47 (13.13) 418–498 7.37 (1.51) 2–9
- Early 0.54 (.50) 0–1 471.12 (15.03) 418–513 7.43 (1.51) 2–9

Note. For gender, girls are coded 1, boys are coded 0. MLSELD preschool language scores are from TOPEL vocabulary subtest; Michigan longitudinal and Oregon language scores are from the Woodcock
Johnson Picture Vocabulary Subtest (w-scores).

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

Page 36
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 5
Predictors of Behavioral Self-regulation Trajectories

b (s.e.) β Raw P-value B-H crit. value Odds ratio

MLSELD preschool
Montroy et al.

-Later vs. Preschool*

• Gender 0.38 (0.21) 0.09 .06 .06 1.46
• Language 0.05 (0.01) 0.34 <.01 <.01 1.40
• Mother ed. 0.26 (0.06) 0.26 <.01 <.01 1.30
-Intermediate vs. Early
• Gender −0.29 (0.23) 0.07 .21 .21 0.75
• Language 0.08 (0.02) 0.52 <.01 <.01 1.68
• Mother ed. 0.23 (0.07) 0.22 <.01 <.01 1.26
Oregon
-Later vs. Preschool*
• Gender −0.08 (0.31) −0.02 .79 .79 0.92
• Language 0.05 (0.02) 0.19 <.01 <.01 1.21
• Mother ed. 0.21 (0.08) 0.26 <.01 .01 1.23
-Intermediate vs. Early
• Gender 0.85 (0.26) 0.19 <.01 <.01 2.34
• Language 0.08 (0.02) 0.49 <.01 <.01 1.63
• Mother Ed. 0.00 (0.05) 0.001 .99 .99 1.00
Michigan longitudinal
-Later vs. Preschool*

Dev Psychol. Author manuscript; available in PMC 2017 November 01.

• Gender 0.58 (0.24) 0.15 .02 .04 1.79
• Language 0.02 (0.01) 0.14 .05 .07 1.15
• Mother ed. 0.09 (0.08) 0.07 .30 .30 1.09
-Intermediate vs. Early
• Gender −0.20 (0.31) −0.05 .50 .76 0.82
• Language 0.02 (0.01) 0.16 .13 .38 1.17
• Mother ed. 0.02 (0.11) 0.01 .89 .89 1.02

Note. B-H crit. Value refers the Benjamini-Hochberg critical value for a given predictor. Mother ed. refers to mother reported education level. Bolded values indicate significant findings.
Page 37