Accepted Manuscript

Full Length Article

High, Low, and In Between: Self-Esteem Development from Middle Childhood

to Young Adulthood

Joanne M. Chung, Roos Hutteman, Marcel A.G. van Aken, Jaap J.A. Denissen

PII: S0092-6566(17)30069-7
DOI: http://dx.doi.org/10.1016/j.jrp.2017.07.001
Reference: YJRPE 3652

To appear in: Journal of Research in Personality

Received Date: 1 March 2017

Revised Date: 24 June 2017
Accepted Date: 3 July 2017

Please cite this article as: Chung, J.M., Hutteman, R., van Aken, M.A.G., Denissen, J.J.A., High, Low, and In
Between: Self-Esteem Development from Middle Childhood to Young Adulthood, Journal of Research in
Personality (2017), doi: http://dx.doi.org/10.1016/j.jrp.2017.07.001

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 1
ADULTHOOD

High, Low, and In Between: Self-Esteem Development from Middle Childhood to Young

Adulthood

Joanne M. Chung1

Roos Hutteman2

Marcel A.G. van Aken2

Jaap J.A. Denissen1

1
Tilburg University
2
Utrecht University

Author Note

Joanne M. Chung, Department of Developmental Psychology, Tilburg University,

Tilburg, Netherlands. Roos Hutteman, Department of Developmental Psychology, Utrecht

University, Utrecht, Netherlands. Marcel A.G. van Aken, Department of Developmental

Psychology, Utrecht University, Utrecht, Netherlands. Jaap J.A. Denissen, Department of

Developmental Psychology, Tilburg University, Tilburg, Netherlands.

Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 2
ADULTHOOD
Please address correspondence to Joanne M. Chung, Department of Developmental

Psychology, Tilburg University, PO Box 90153, 5000 LE, Tilburg, The Netherlands. Phone:

+31 13 466 2151. E-mail: [email protected].

High, Low, and In Between: Self-Esteem Development from Middle Childhood to Young

Adulthood
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 3
ADULTHOOD
Abstract

We describe self-esteem development in a German sample (N = 240, 48% female) followed

longitudinally from middle childhood to young adulthood, using data spanning 20 years. Data

from the Self-Perception Profile for Children (Harter, 1985) and the Self-Description

Questionnaire III (Marsh & O’Neill, 1985) were linked using item response theory methods.

Rank-order stability was high in middle childhood, low in adolescence, and highest in young

adulthood. Mean-levels were relatively high in middle childhood, decreased into adolescence,

but increased into young adulthood. Early childhood shyness and aggressiveness as rated by

parents, teachers, and observers did not influence the self-esteem trajectory. We provide the

first longitudinal evidence for the self-esteem trajectory from middle childhood to young

adulthood, replicating and extending previous findings.

Keywords: self-esteem, longitudinal, shyness, aggressiveness, item response theory

Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 4
ADULTHOOD
High, Low, and In Between: Self-Esteem Development from Middle Childhood to

Young Adulthood

Global self-esteem refers to a person’s subjective evaluation of his or her self-worth.

The extent to which a person holds positive self-views has been shown to be important for

promoting goals, coping mechanisms, and behaviors that enable success for one's

relationships, career, and well-being (Kuster, Orth, & Meier, 2013; Orth, Robins, &

Widaman, 2012; Steiger, Allemand, Robins, & Fend, 2014; Trzesniewski, Donnellan,

Moffitt, Robins, Moulton, & Caspi, 2006). Recently, there has been an influx of longitudinal

research characterizing the trajectory of self-esteem (Orth & Robins, 2014; Orth, Maes, &

Schmitt, 2015; von Soest, Wichstrøm, and Kvalem, 2015), offering much needed insights

into the normative development of self-esteem from adolescence to young adulthood, and

young adulthood to old age. Yet, to date, research on the lifespan trajectory of self-esteem

has not included the shift from childhood to adolescence. During this time, children begin to

mature in their cognitive abilities, encounter puberty, and become more concerned with their

peers (Robins & Trzesniewski, 2005), making this transition especially important for self-

esteem development. Furthermore, little research has examined the developmental

antecedents of self-esteem (but see Harris, Gruenenfelder-Steiger, Ferrer, Donnellan,

Allemand, Fend, Conger, Trzesniewski, 2015; Orth, 2017). Therefore, gaining more

understanding of the lifespan development of self-esteem from middle childhood onward,

and moreover, identifying childhood factors that may predict the trajectory of self-esteem

itself are important.

In the present study, we examined global self-esteem development from middle

childhood into young adulthood. We began by charting the trajectory of self-esteem, and

sought to extend previous cross-sectional and longitudinal findings regarding self-esteem

development by examining both stability and change during middle childhood, adolescence,
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 5
ADULTHOOD
and young adulthood. We used multiple informant reports (i.e., parent, teacher, and observer

ratings) of early childhood personality to predict who increased or decreased in self-esteem

during the transitions from childhood to adolescence and from adolescence to young

adulthood. Specifically, we focused on shyness and aggressiveness at the ages of 4 to 6 years

old as predictors of self-esteem because these traits have been linked to low self-esteem and

difficulties in the social realm, raising the possibility that they could affect the development

of self-esteem. Below, we review previous research on these topics, first turning to the

literature on self-esteem development.

The Development of Self-Esteem from Middle Childhood to Young Adulthood

Self-esteem is an evaluation of one’s own worthiness and competence, but theoretical

perspectives emphasize the importance of the social world in shaping self-esteem. Our self-

views are thought to develop from our interactions with others and how we believe others see

us (Cooley, 1902; Harter, 2012; Mead, 1934). Sociometer theory (Leary, Tambor, Terdal, &

Downs, 1995) highlights the social nature of self-esteem and posits that it is a sociometer, or

psychological gauge that signals the extent to which one is accepted by others, helping people

maintain their social ties. The stable component of self-esteem is seen as one’s judgment that

he or she is generally valued and accepted by others, and as the “resting state” of the

sociometer (Leary et al., 1995). Self-esteem is moderately stable across time and contexts, yet

it is also mutable, especially during developmental transitions such as the ones from

childhood to adolescence, and adolescence to young adulthood (Huang, 2010; Hutteman,

Nestler, Wagner, Egloff, & Back, 2015). Both rank-order stability and mean-levels of self-

esteem change across the lifespan. Rank-order stability refers where an individual stands on

the construct of interest, relative to others in the sample. Rank-order stability of self-esteem is

lowest in early childhood and old age, relatively low but increasing in adolescence, and

highest in adulthood (Donnellan, Kenney, Trzesniewski, Lucas, & Conger., 2012; Orth &
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 6
ADULTHOOD
Robins, 2014; Trzesniewski, Donnellan, & Robins, 2003). In childhood, stability is

considered to be low because self-esteem is emerging and not fully formed during this time

(Eccles, Wigfield, Harold, & Blumenfeld, 1993; Robins, Trzesniewski, Tracy, Gosling, &

Potter, 2002). In adolescence, stability is argued to be higher than in childhood due to an

increased awareness of self, but lower relative to young adulthood because of maturational

and social changes that are experienced during this time (Orth & Robins, 2014; Robins &

Trzesniewski, 2005). Accordingly, in the present study, we expected to find low rank-order

stability in childhood, low but increasing stability into adolescence, and high stability into

young adulthood.

Mean, or average levels of self-esteem, are relatively high in childhood, decrease

during adolescence, and then steadily increase into young adulthood (Orth, Maes, & Schmitt,

2015; Orth, Robins, & Widaman, 2012; Orth, Trzesniewski, & Robins, 2010; Robins,

Trzesniewski, Tracy, Gosling, & Potter, 2002, von Soest, Wichstrøm, & Kvalem, 2016). This

research indicates that individuals exhibit relatively high self-esteem in childhood, perhaps

because their self-views are unrealistically positive (Robins & Trzesniewski, 2005). As

children's cognitive skills mature, they begin to compare their skills and abilities to their

peers and consider feedback from close others, including the extent to which they are liked

and accepted by others, and begin to develop more accurate, and generally, less positive self-

views (Robins & Trzesniewski, 2005). Self-esteem continues to decrease into adolescence,

with pubertal changes, changing school contexts, and the increased capacity for self-

reflection being implicated in this decline (Orth & Robins, 2014; Robins & Trzesniewski,

2005). During the transition into young adulthood, self-esteem has been found to increase as

individuals gain increasing autonomy at school and work and continue to expand and deepen

their social ties with others (Orth & Robins, 2014). Accordingly, in the present study, we

expected to find self-esteem to be relatively high in mid-childhood, decreasing into

Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 7
ADULTHOOD
adolescence, and then increasing into young adulthood.

The majority of longitudinal research on self-esteem development has examined it

exclusively in childhood (Arunkumar, Midgley, & Urdan,1999; Rodriguez, Wigfield, &

Eccles, 2003), from childhood to adolescence (Hoge, Smit, & Hanson, 1990; Wigfield,

Eccles, Mac Iver, Reuman, & Midgley, 1991), or from adolescence to adulthood (Erol &

Orth, 2011; Orth & Robins, 2014; von Soest, Wichstrøm, & Kvalem, 2016), including in a

recent study that examined self-esteem in the LOGIC data from when youth were 17- to 29-

years-old (Luan, Hutteman, Denissen, Asendorpf, & van Aken, 2016). These studies have

contributed much toward our understanding of self-esteem development. Yet, when piecing

together the evidence from different developmental and time periods, we cannot rule out

cohort effects -- the possibility that age differences in self-esteem are confounded with

experiences of events that are not shared with other populations (Baltes, Cornelius, &

Nesselroade, 1979). For example, some have claimed that societal shifts in the focus on the

self have influenced the nature and developmental course of self-esteem (Twenge, 2006).

Studies linking childhood to adulthood are rare, understandably, because researchers are

often faced with the issue of heterotypical continuity, where manifestations of the same

underlying trait change as individuals mature. In the present study, we had the unique

opportunity to track self-esteem in the same group of individuals over two decades, allowing

us to examine normative changes in self-esteem from middle childhood to young adulthood.

Early Childhood Shyness and Aggressiveness as Antecedents of Self-Esteem

Development

Personality is thought to influence the ways individuals construe the world around

them (Caspi & Shiner, 2011; Rothbart, 2011) as well as themselves (Robins, Donnellan,

Widaman, & Conger, 2010; Robins, Tracy, Trzesniewski, Potter, & Gosling, 2001).

Specifically, personality can influence which features of a social interaction that one attends
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 8
ADULTHOOD
to, in turn, coloring one's perceptions and experience, including one's self-evaluations (Caspi

& Shiner, 2011). Shyness and aggressiveness are aspects of personality that reveal how

individuals relate to the social world, and reflect both temperament and social competence

(Caspi, Elder, &Bem, 1987; Caspi, Elder, & Bem, 1988; Denissen, Asendorpf, & van Aken,

2008; Horney, 1950; Rubin, Coplan, & Bowker, 2009). Shyness refers to a tendency towards

experiencing worry, tension, or awkwardness during social interactions, especially with

strangers (Cheek & Buss, 1981; Coplan & Rubin, 2010). Aggressiveness, in contrast, refers

to a proneness towards behaving with the intention of causing harm to another person

(Anderson & Bushman, 2002). Caspi and colleagues (1987; 1988) operationalized “moving

away from the world” and “moving against the world” as informant ratings of children's

tendencies towards exhibiting shy and aggressive behaviors respectively, and found that

being perceived as either shy or aggressive in childhood predicted poor outcomes in the

psychological, social, and work domains 30 years later. Shyness and aggressiveness in

childhood are thought to have such powerful consequences on individuals’ life trajectories

because of their consistency throughout the lifespan (Caspi et al., 1987; 1988).

The extant research indicates that shyness in childhood is concurrently associated

with low self-esteem (Crozier, 1995; Kemple, David, & Wang, 1996; Rubin, Coplan, &

Bowker, 2009). Studies also indicate that shyness in childhood shows negative prospective

effects for self-esteem. In a previous study that examined the LOGIC data, youth who were

seen as extremely shy towards familiar peers at 4- to 6-years-old were likely to report low

levels of self-esteem between the ages of 8- and 10-years-old (Asendorpf & van Aken, 1994).

In another, longitudinal study, Icelandic youth who were categorized as shy at 7-years-old

were likely to exhibit low levels of self-esteem in early and middle adolescence (Hart,

Hoffman, Edelstein, & Keller, 1997). Shyness in childhood is also predictive of a restricted

social life (Gest, 1997) and low self-esteem (Rubin, Chen, McDougall, Bowker, &
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 9
ADULTHOOD
McKinnon, 1995) in adulthood. Additionally, Swedish girls who were perceived as shy in

early childhood were likely to exhibit low levels of self-esteem 30 years later (Kerr, 2000).

These results suggest that we might expect that individuals perceived as shy in early

childhood might be negatively affected in their self-esteem development.

With regard to aggression, the extant research suggests that it is associated with low

self-worth (Coplan, Prakash, O'Neil, & Armer, 2004; Donnellan, Trzesniewski, Robins,

Moffitt, & Caspi, 2005; but see Denissen, Thomaes, & Bushman, in press). And, like

shyness, longitudinal studies that examine aggression in childhood suggests that it has

negative prospective effects on related constructs. In one longitudinal study, American

children who were rated by teachers as aggressive in kindergarten were less likely to be

accepted by their peers, more likely to have fewer friends, and more likely to be lonelier than

their less aggressive peers from kindergarten to the 2nd grade (Ladd & Burgess, 1999). In

another, longitudinal study, Norwegian adolescents who were seen by their mothers as

aggressive from infancy to mid-adolescence were likely to report low levels of life

satisfaction and high levels of depression and anxiety (Kjeldsen, Nilsen, Gustavson,

Skipstein, Melkevik, & Karevold, 2016). In yet another, longitudinal study, American

children who were seen as aggressive at 7-years-old were likely to experience sharply

increasing levels of loneliness from late childhood to mid-adolescence (Schinka, van

Dulmen, Mata, Bossarte, & Swahn, 2013).

Taken together, while much of the research on self-esteem has focused on its

prospective effects (e.g., Kuster, Orth, & Meier, 2013; Orth, Robins, & Widaman, 2012;

Trzesniewski, Donnellan, Moffitt, Robins, Poulton, & Caspi, 2006), including on aggression

(Denissen, Thomaes, & Bushman, in press; Donnellan, Trzesniewski, Robins, Moffitt, &

Caspi, 2005), conceptual and empirical links point to the possibility that children who possess

traits that might hinder successful social relations, such as high levels of shyness and
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 10
ADULTHOOD
aggressiveness, might be negatively impacted in their subsequent self-esteem development.

Examining this possibility not only provides a longitudinal consideration of the ideas posited

by sociometer theory (Leary et al., 1995), but can shed light on whether or not early

childhood is an appropriate time to target these ways of interacting with the world. To date,

however, there is no research examining how shyness and aggressiveness in early childhood

predicts changes in self-esteem from middle childhood to adulthood.

Present Study

In the present study, we examined global self-esteem development in a sample of

individuals tracked over 20 years. Self-esteem was assessed using two different measures of

global self-esteem. In order to chart the trajectory of self-esteem from middle childhood to

young adulthood, we linked scores from the two different measures of global self-esteem

using item response theory methods. Following previous cross-sectional and longitudinal

findings regarding self-esteem development (Robins et al., 2002; Orth & Robins, 2014), we

expected that self-esteem would be relatively high in childhood, decrease into adolescence,

and then increase into young adulthood. We then tested whether or not shyness and

aggressiveness in early childhood influenced the trajectory of self-esteem. We expected that

those who were seen as having shy or aggressive personalities in childhood would show a

less positive self-esteem trajectory relative to their peers.

In the present study, we adopted a multi-method approach in addressing our research

aims. Specifically, we used data from multiple informants, consisting of parent, teacher, and

observer ratings of the early childhood shyness and aggressiveness. These data encompassed

measures that differed in their objectivity -- ranging from the familiar perceptions of the

participant's shyness and aggressiveness made by parents and teachers to the more distal

judgments of observers in both laboratory and field settings. By using this approach to

assessing early childhood personality, we were able to obtain a more comprehensive

Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 11
ADULTHOOD
perspective of shyness and aggressiveness and their influences on participant ratings of their

own self-esteem, moving beyond previous research that has relied on either self- or single

informant ratings.

Method

Sample and Procedure

We used data from the Munich Longitudinal Study on the Genesis of Individual

Competencies (LOGIC)1, a longitudinal study that was, in part, designed to examine the

effects of personality and social experiences on adjustment (for further details about the

study, see Schneider & Bullock, 2009; Weinert & Schneider, 1999). Our sample consists of

240 individuals (48% female) who were born in 1980 or 1981 and who attended preschool in

Munich, Germany. Families were recruited from 20 preschools, and more than 90% allowed

their child to participate in the study. All children spoke German as their first language, and

according to parent reports of professional status (Wegener, 1988), 63% of families were of

middle SES, 28% were of low SES, and 9% were of high SES.

The present study used 9 waves of data, when youth were 4, 5, 6, 9, 10, 12, 17, 23,

and 29 years of age. Assessments for the first 7 assessments were conducted at the Max

Planck Institute for Psychological Research in Munich, and assessments for the last two

1
Data from the LOGIC study have been used in several studies. Studies focused on social
development have examined shyness (Asendorpf, 1992), as well as aggressiveness, including
their developmental trajectories (Denissen, Asendorpf, & van Aken, 2008), long-term
stability and effect on life outcomes (Asendorpf, Denissen, & van Aken, 2008), the interplay
between the two (Hutteman, Denissen, Asendorpf, & van Aken, 2009), and their associations
with the Big Five personality traits in childhood (Asendorpf & van Aken, 2003).
Additionally, some studies have focused on self-esteem during late adolescence and young
adulthood (Luan, Poorthuis, Hutteman, Asendorpf, Denissen, & van Aken, 2017; Sturaro,
Denissen, van Aken, & Asendorpf, 2008), as well as the extent to which shyness predicts
social self-esteem during childhood (Asendorpf & van Aken, 1994). However, no studies
have examined the trajectory of global self-esteem development from childhood to young
adulthood, and none have examined how early childhood shyness and aggressiveness predict
individual differences in this trajectory.
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 12
ADULTHOOD
waves were administered in an online questionnaire. Families were contacted several times

during the year to encourage participation in the study; retention rates were relatively high

from the Age 4 to Age 12 assessments, and decreased slightly from the Age 17 to Age 29

assessments (See Supplemental Information, p.1, Table 1 for assessment intervals and sample

sizes for each type of rating). To investigate the impact of attrition, we compared individuals

who provided self-esteem ratings at the Age 9 assessment to those who did not at the Age 29

assessment on the study variables. Those who provided self-esteem ratings at both waves (M

= 2.86), F(1, 95) = 8.51, p < .05, ηp2 = .08) were rated as higher on the aggregated Shyness

Towards Unfamiliar Peers subscale as rated by parents (described in the Measures section

below) than those who did not provide self-esteem ratings at the Age 29 assessment (M =

3.65). We did not find any additional effects of attrition on any of the other main study

variables.

Measures

Self-esteem. Two measures of self-esteem were administered at six assessments (i.e.,

Age 9, 10, 12, 17, 23, and 29), and did not overlap. In the Age 9 to Age 12 assessments, a 6-

item German version of the Global Self-Worth subscale from the Self-Perception Profile for

Children (SPPC; Harter, 1985) was administered. Children were first asked which part of the

statement was most like them (e.g., "Some children are pleased with themselves but other

children are not pleased with themselves"), and then asked how well it described them.

Responses were then coded from 1 (very low self-esteem) to 4 (very high self-esteem).

Coefficient alphas ranged between .76 to .77. In the Age 17 to Age 29 assessments, a 6-item

German version of the General Self-Concept subscale from the Self-Description

Questionnaire III (SDQ-III; Marsh & O’Neill, 1985) was administered (e.g., "All in all, I

accept myself as I am"). Response options for this measure ranged from 1 (not at all) to 5

(completely). Coefficient alphas ranged between .79 to .80.

Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 13
ADULTHOOD
Self-esteem-related items. In the Age 12 and Age 17 assessments, bipolar adjective

pairs assessing Big Five personality dimensions (Ostendorf, 1990) were administered. We

identified two items measuring the Emotional Stability dimension (“self-confident –

vulnerable”, and “insecure” – “self-confident”) that are related to self-esteem (rs among the

items and the self-esteem score were both .21 at Age 12, and .45 - .50 at Age 17; all ps < .05).

Participants evaluated each adjective pair in relation to each other, such that response options

ranged from 1 (very [first adjective], not [second adjective]), to 5 (not [first adjective], very

[second adjective]). The responses to the first item was reverse-coded.

Shyness and aggressiveness in early childhood. Six measures of shyness and three

measures of aggressiveness as rated by three informants were examined for the Age 4 to Age

6 assessments. As in previous research on the LOGIC study (Asendorpf & van Aken, 1999),

averages were computed across these three assessments to represent indices of shyness and

aggressiveness in early childhood, with the exception of observer ratings derived from the

laboratory task, for which Age 5 data were not available.

Parent ratings. Parents completed a questionnaire designed to assess their perceptions

of their child’s shyness and aggressiveness. The 4-item Shyness Towards Unfamiliar Adults

(SUA) subscale (e.g., “My child is shy toward unknown adults”) and the 4-item Shyness

Towards Unfamiliar Peers (SUP) subscale were examined. The items in the subscales were

identical, but referenced either other children or adults. Parents also completed the 4-item

Aggressiveness Towards Peers (AP) subscale (e.g., “My child is aggressive toward other

children”). Coefficient alphas showed ranges between .93-.95, .89-.94, and .82-.85 for SUA,

SUP, and AP, respectively.

Teacher ratings. Teachers completed a 54-item German version of the California

Child Q-Set (Göttert & Asendorpf, 1989) to assess the children’s personality. Teachers were

instructed to assign six items each into nine categories, ranging from 1 (extremely
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 14
ADULTHOOD
uncharacteristic) to 9 (extremely characteristic). Prototypicality scores were derived by

correlating each child’s Q-sort profile with a prototypic Q-sort profile for shyness (e.g., “The

child is inhibited and constricted”) and aggressiveness (e.g., “The child is aggressive”).

Coefficient alphas were above .80 and .78 for the three assessments of shyness and

aggressiveness, respectively.

Observer ratings. Children were observed interacting in a laboratory task meant to

elicit behavioral inhibition, and free play at preschool (see Asendorpf, 1990 for more

information). For shyness, Contact Initiation Latency (the number of seconds it took for the

child to approach the stranger) and Shyness Towards Strangers ratings (how shy child

seemed to be in presence of the stranger) were obtained from the laboratory task. For the

latter, observer ratings ranged from 1 (not at all) to 7 (extremely), and coefficient alphas

ranged between .89 to .97. Wait-and-Hover ratings (percentage of social interactions initiated

by child after gaining proximity to other children relative to total number of initiations) were

obtained from the free play task, and kappa reliability was above .90. For aggressiveness,

Aggressive Attacks among All Own Initiations ratings (percentage of verbal and physical

acts of aggression relative to total number of initiations) were obtained from the free play

task, and kappa reliability was above .90.

Power Considerations for the Current Study

The literature regarding adequate power for item response theory models does not

provide a definitive answer. Some researchers recommend 250 or 500 respondents for the

Graded Response Model (Embretson & Reise, 2000), while others say that this

recommendation may be relaxed if there are reasonable associations between the items with

the latent factor (K. J. Grimm, personal communication, May 16, 2017) (See Supplemental

Information, p.2, Table 2 for factor loadings). For structural equation models, some

researchers recommend a minimum sample size of 100 or 200 (Boomsma, 1982, 1985; Kline,
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 15
ADULTHOOD
2005), while others recommend 5 or 10 observations per estimated parameter (Bentler &

Chou, 1987). Thus, readers should interpret our findings under the caveat that our sample (N

= 195 for the unconditional model) is modestly powered depending on which rule of thumb

the reader favors. With regard to the structural equation models, we conducted Montecarlo

simulations to examine post-hoc power for the parameters estimated by the growth models

we conducted, and to take into account missing data. We did this by entering the estimates

provided by our models as population values, including the patterns of missingness we

observed in our data and the proportion of subjects within each pattern, and requested 10,000

replications for each of our models.

Description of Linking Procedure Using Item Response Theory Methods

The goals of using a linking procedure were to describe the trajectory of self-esteem

from when the child was 9-years-old to when they were 29-years-old (See Figure 1 for

original scale scores).

We employed item response theory methods to link two different measures of self-

esteem to each other (Curran, Hussong, Cai, Huang, Chassin, Sher, & Zucker, 2008; de

Ayala, 2009; Kolen & Brennan, 2004). Linking is a less restrictive form of equating, a group

of procedures that refer to the adjustment of person location estimates derived from different

forms to place them on a common metric (de Ayala, 2009; Kolen & Brennan, 2004).

Because there was no overlap in the administration of self-esteem measures in the

Age 12 and Age 17 assessments, we used external common items-- items that are closely

related to the constructs of interest and that were administered during both Age 12 and Age

17 assessments. External common items do not contribute directly to the final self-esteem

estimates, but are used to calibrate, or place the two different measures of a construct on the

same metric (de Ayala, 2009; Kolen & Brennan, 2004). In the present study, we linked the
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 16
ADULTHOOD
global self-esteem data from the SPPC (Harter, 1985) and SDQ-III (Marsh & O’Neill, 1985)

using the self-esteem related items described above.

We took a concurrent calibration approach (de Ayala, 2009; Kolen & Brennan, 2004)

to linking the data. This method treats item characteristics (i.e., difficulty and discrimination)

for the common items on one form (e.g., the data from the Age 12 assessment) as fixed, true

values, and then scales the items on the other form (e.g., the data from the Age 17

assessment) according to the estimates of the fixed common items (de Ayala, 2009; Kolen &

Brennan, 2004). Under this approach, scaling was done simultaneously across all times of

measurement (de Ayala, 2009; Kolen & Brennan, 2004).

For each construct, we began by fitting a Graded Response Model (GRM; Samejima,

1969) that included the items from the first measure of the construct at the Age 12

assessment, the two external common items at the Age 12 and Age 17 assessments, and the

items from the second measure of the construct at the Age 17 assessment. The GRM is an

item response model that estimates both discrimination (e.g., how well an item differentiates

people of different standing on the latent trait, theta) and location parameters (e.g., how

difficult an item is to endorse given one’s standing on theta) for items that have more than

two response categories (Samejima, 1969). We saved the estimated item parameters from this

GRM.

We continued by fitting another GRM to all of the longitudinal data available from

the Age 9 through Age 29 assessments in one step, using the saved parameters from the

previous GRM as fixed estimates in which to calibrate items from the two different measures

onto the same metric. By linking the data from the two different measures using this method,

the model provided estimates of theta --the underlying latent traits-- in our case, self-esteem--

for each individual at each wave, allowing us to examine the trajectory of self-esteem from

when individuals were 9-years-old to when they were 29-years-old. We then saved the theta
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 17
ADULTHOOD
estimates as individual factor scores, and used these values in subsequent analyses (See

Figure 2 for these estimates). As is generally the case in this procedure, theta was scaled so

that it had a mean of 0 and a standard deviation of 1 (de Ayala, 2009; Kolen & Brennan,

2004).

Results

We conducted our study analyses using the psych Version 1.5.8 (Revelle, 2015) and

mirt (Chalmers, 2012) packages in R Version 3.2.3 (R Core Team, 2015), and Mplus Version

7 (Muthén & Muthén, 1998-2012). We used the Root Mean Square Error of Approximation

(RMSEA) (≤ .08), the Comparative Fit Index (CFI) (≥ .90), and the Tucker-Lewis Index

(TLI) (≥ .90) as indices to determine adequate model fit (Browne & Cudeck, 1993; Hu &

Bentler, 1999; See Table 1 for fit statistics for all models and Table 2 for descriptive

statistics).

Rank-Order Stability of Self-Esteem from Middle Childhood to Young Adulthood

To examine the rank-order stability of self-esteem from childhood to young

adulthood, we fit a first-order autoregressive model (Model 1) to the theta estimates for the

six assessments, fixing the measurement error to be equal across assessments (see Table 3 for

model estimates and description of how estimates were corrected for time interval). First-

order autoregressive models are used on repeated measures data and are in contrast to

traditional OLS methods because they take into account the dependency of an earlier

assessment on later assessments so that each assessment consists of the previous assessment

and a random error component. Therefore, first-order autoregressive models provide

estimates of the latent variable separately from error, allowing for stability estimates that

have been corrected for unreliability of the measure.

Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 18
ADULTHOOD
Rank-order stability for self-esteem was lowest between the Age 10 and Age 12

assessments (b corrected for time interval = .755), and highest between the Age 23 and Age

29 assessments (b corrected for time interval = .994).

Mean-Level Change in Self-Esteem from Middle Childhood to Young Adulthood

Our next step was to examine mean-level change and individual differences in the

trajectory of self-esteem. A visual inspection of the factor scores associated with the theta

estimates provided by our linking procedure (See Figure 2 for theta estimates) suggested a

non-linear trajectory, with a decrease at the Age 12 assessment. Thus, we fit a piecewise

growth model. We first fit an unconditional model (Model 2) to determine whether there was

an average, or typical, trajectory of self-esteem. In addition to a latent intercept, we specified

two latent slopes that were centered at the Age 9 assessment, with these coefficients being 0.

For the first latent slope, the five remaining assessments were given values of 1, 3, 3, 3, 3,

corresponding to the difference in years between assessments for the first piece (i.e., the

number of years between the Age 9 and Age 10 assessments was 1, the number of years

between the Age 9 and Age 12 assessments was 3, and to accommodate the specification of

the second piece, the remaining assessments were fixed at 3). For the second latent slope, the

Age 10 and Age 12 assessments were both given a value of 0, reflecting the fact that the

specification of the second piece began at the Age 17 assessment. The Age 17 assessment

was allowed to be freely estimated, and the Age 23 and Age 29 assessments were given

values of 11 and 17, corresponding to the difference in years between assessments after

accounting for the fixed value of 3 for the same assessments in the first piece (i.e., the

number of years between the Age 9 and Age 23 assessments, and between the Age 9 and Age

29 assessments were 14 and 20 years, respectively)2.

2
To make sure that there were no other, simpler models that could adequately characterize
our data, we fit additional models: linear, latent basis, quadratic, cubic, and piecewise linear
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 19
ADULTHOOD
By specifying the latent slopes this way, we created two “pieces” that converged at

the self-esteem drop that occurred at the Age 12 assessment. Additionally, these values take

into account the uneven spacing of assessments (e.g., the length of time between assessments

become much greater in the later assessments than in the beginning) and the slight non-

linearity at the Age 17 assessment. Thus, we found that a model with one latent intercept, one

latent linear slope, and one latent basis slope that were allowed to covary fit the data well

(See Figure 3 for path diagram).

Below we refer to the latent intercept as the self-esteem intercept, the first latent slope

as the childhood slope (covering the Age 9 to Age 12 assessments, and the second latent

slope as the youth slope (covering the Age 12 to Age 29 assessments) (See Table 4 for model

estimates).

The mean (B = .199) and variance (B = .408) of the self-esteem intercept were

significant, indicating that on average, children started out with relatively high levels of self-

esteem, but that there were individual differences in initial levels of self-esteem. The mean of

the childhood slope was negative and also significant (B = -.153), but its variance was not

significant (B = .031), indicating that on average, children decreased in self-esteem into late

childhood, and also, that the majority of the youth in our sample experienced this decrease.

Due to the non-significant variance of the childhood slope, we do not interpret the

associations between this latent variable and others in this model and in subsequent models

that include it. The mean (B = .019) and variance (B = .002) of the young adulthood slope

were significant, indicating that on average, youth increased in self-esteem from adolescence

to young adulthood and also, that there were individual differences in this trajectory. The

models. Each of these models differ in the way the latent slope(s) were specified. See
Supplemental Information, pp. 3-4, Table 3 for model specifications and fit. The results of
these models indicate that the piecewise latent basis model was the most reasonable model
for our data.
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 20
ADULTHOOD
self-esteem intercept was not significantly associated with the young adulthood slope (r = -

.286), indicating that the increase in self-esteem into young adulthood was not associated

with self-esteem levels when children were 9-years-old.

Early Childhood Shyness and Aggressiveness as Antecedents of the Self-Esteem

Trajectory

Do shyness and aggressiveness as rated by parents, teachers, and observers of youth

in early childhood predict the trajectory of self-esteem from middle childhood to young

adulthood? To address this question, we specified separate conditional models (Model 3 and

Model 4) to examine the potential moderating effects of shyness and aggression on the self-

esteem trajectory (See Table 5 for model estimates).

Parent, teacher, and observer ratings for shyness and aggressiveness were first

standardized and then entered into each model as manifest indicators of the latent variable. In

the case of more than one variable for each informant, the average of the standardized

variables was used. Specifically, to test the effects of shyness (Model 3), we used the average

of parent ratings as assessed by the SUA and SUP subscales, teacher ratings of shyness, and

the average of observer ratings as measured by Contact Initiation Latency, Shyness with

Strangers, and Wait-and-Hover as manifest indicators of a latent Shyness factor. To test the

effects of aggressiveness (Model 4), we used the parent ratings as assessed by the AP

subscale, teacher ratings of aggressiveness, and observer ratings as assessed by Aggressive

Attacks among All Own Initiations as manifest indicators of a latent Aggressiveness factor.

For both models, our results show that there were no significant effects on the self-

esteem trajectory, indicating that neither shyness nor aggressiveness as rated by parents,

teachers and observers when individuals were 4- to 6-years-old predicted self-esteem

development when individuals were 9-years-old to 29-years-old.

Discussion
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 21
ADULTHOOD
In the present research, we examined global self-esteem development in a sample of

individuals followed longitudinally from age 9 to age 29, and then tested the possibility that

children’s styles of relating to the social world in early life – as perceived by parents,

teachers, and observers – were predictive of individual differences in the trajectory of self-

esteem. Our findings contribute to the literature on self-esteem development by describing its

longitudinal trajectory from middle childhood, in contrast to previous studies that have

investigated self-esteem development using adolescence as its starting point. Furthermore, by

considering the extent to which the ways that “moving away from” or “moving against” the

world (Caspi, Elder, & Bem, 1987; Caspi, Elder, & Bem, 1988; Horney, 1950) might allow

for the development of positive self-views, we provide tentative evidence that shyness and

aggressiveness in early childhood may not be key factors for later self-esteem development,

adding to the sparse literature on the origins of self-esteem. Below we discuss the

implications of our findings with regards to our research aims.

Rank-Order Stability of Self-Esteem from Middle Childhood to Young Adulthood

One aim of the present study was to examine the stability in individual differences in

self-esteem from middle childhood to young adulthood. These findings are in line with

previous research that shows that stability is lower in adolescence (Alsaker & Olweus, 1992;

Block & Robins, 1993) than it is in young adulthood (Trzesniewski, Donnellan, & Robins,

2003), but differ markedly from estimates that have been previously found for childhood, in

which stability is found to be substantially lower. One likely reason for this is that we used a

first-order autoregressive model to examine rank-order stability in the present study, allowing

for estimates that are free of measurement error and that also take into account the

dependency of each assessment on a prior assessment.

What else could explain the high levels of stability we found in middle childhood?

Research suggests that self-esteem during this time is influenced by self-perceived

Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 22
ADULTHOOD
competence in domains that the individual finds to be important to the individual, as well as

social support from peers (Harter, 2012). At the ages of 9 and 10, the individuals in our

sample were nearing the end of elementary school. Most German children attend elementary

school for four years from the age of 6, and it could be that such high levels of stability

manifested because the social environment during this time was relatively stable, especially

in comparison to the impending transition into junior high as well as pubertal changes to

come, two factors that have been demonstrated to influence self-esteem stability (Eccles,

Midgley, & Adler, 1984; Simmons, Rosenberg, & Rosenberg, 1973).

Mean-Level Change in Self-Esteem from Middle Childhood to Young Adulthood

Another aim of the present study was to examine the trajectory of self-esteem through

the important developmental transitions between childhood and adolescence, and adolescence

and young adulthood. The findings from the present study contribute to the literature by

being the first to track self-esteem in the same group of individuals during middle childhood,

adolescence, and young adulthood by linking two measures of self-esteem using item

response theory methods. Additionally, by examining how self-esteem changes starting from

middle childhood, when young children’s global self-esteem is thought to first develop

(Harter, 2012), we were able to track self-esteem during a period that has been understudied

within the lifespan literature. Our results show that there is a normative, or average trajectory

of self-esteem from middle childhood to young adulthood, such that self-esteem is relatively

high in middle childhood, decreases into late childhood, and then steadily increases into

young adulthood.

The findings from the present study provide longitudinal support for previous cross-

sectional accounts that self-esteem is high in childhood (Harter, 2012; Marsh, Craven, &

Debus, 1991; Robins et al., 2002), and decreases into adolescence (Robins et al., 2002).

These results indicate that the self-esteem of individuals in our sample was most impacted at
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 23
ADULTHOOD
12-years-old, as it showed a mean-level decrease that did not vary much across individuals.

Biological (e.g., pubertal), psychological (e.g., increased capacity for abstract thought) and

social (e.g., changing schools, shifts in social relationships) changes are likely to be

implicated in this decrease.

We also found that self-esteem increased from the point the individuals in our sample

were 12-years-old, and our results differ somewhat from previous cross-sectional findings

that show that self-esteem continues to decrease throughout the adolescent years, starting to

increase only around the time individuals are 18-years-old (Robins et al., 2002). Instead, our

findings are more consistent with extant longitudinal work that show increases in self-esteem

beginning at 13- and 14- years-old (Orth, Maes, & Schmitt, 2015; von Soest, Wichstrøm, &

Kvalem, 2016). Our findings also correspond with previous longitudinal research that shows

self-esteem increases from adolescence into young adulthood (Hutteman et al., 2015; Orth,

Robins, & Widaman, 2012). Additionally, the effect size estimate from our age 13

assessment to our age 29 assessment (Cohen’s d = .29) is similar to the effect size estimates

shown in previous research (Cohen’s d ~ .25 between age 15 and age 30 in Orth, Maes, &

Schmitt, 2015; Cohen’s d = .30 between age 13 to age 31 in von Soest, Wichstrøm, &

Kvalem, 2016). The increases that we found in self-esteem from adolescence to young

adulthood suggest that many of the individuals in our sample were successful in negotiating

the challenges that accompany these important transitions (Hutteman, Hennecke, Orth, Reitz,

& Specht, 2014).

Early Childhood Shyness and Aggressiveness as Antecedents of Self-Esteem

Development

In using piecewise growth models, we were able to identify a normative trajectory of

self-esteem from middle childhood to young adulthood, and also ascertain that there were

individual differences in self-esteem levels when children were 9-years old and subsequent
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 24
ADULTHOOD
changes in self-esteem from age 12 to 29. Given that we were able to model self-esteem

development from an age in which self-esteem is first thought to take shape (Harter, 2012), a

secondary aim of the present study was to examine whether individual differences in shyness

and aggressiveness in early childhood predicted individual differences in the self-esteem

trajectory.

Sociometer theory (Leary et al., 1995) identifies social acceptance as a key

contributor to self-esteem, and we reasoned that acting in ways that hinder successful social

interactions (i.e, being high in shyness and/or aggressiveness) might negatively affect self-

esteem development. Thus, we examined whether individual differences in shyness and

aggressiveness in early childhood predicted individual differences in the self-esteem

trajectory. We found that these effects were not significant, indicating that children’s shyness

and aggressiveness in early childhood did not appear to have an impact on self-esteem levels

at 9-years-old, as well as their subsequent development. By taking a multi-informant

approach to assessing shyness and aggressiveness, the present study provides a

multidimensional view of these constructs, moving beyond limitations of previous research

that may suffer from shared method variance.

Why is it that these crucial ways of interacting to the world, shown in previous

literature to be associated with self-esteem (Donnellan et al., 2005; Kemple, David, & Wang,

1996; Rosenberg, Schooler, & Schoenbach, 1989; Rubin, Coplan, & Bowker, 2009), do not

affect its developmental course? It could be that certain types of shyness and aggressiveness,

may not always be detrimental for social acceptance, and ultimately for self-esteem

development. Although the current study reveals non-significant effects that should be

examined in future research, our results lend support to studies that find that being socially

withdrawn and aggressive is not necessarily consequential for peer relationships in childhood
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 25
ADULTHOOD
(Ladd & Burgess, 1999; Richardson, Hitti, Mulvey, & Killen, 2014; Vaughn, Vollenweider,

Bost, Azria-Evans, & Snider, 2003).

Furthermore, although we examined shyness and aggressiveness as predictors of the

self-esteem trajectory because we were interested in the effect of early childhood personality

on self-esteem development, it is likely that individual differences in the self-esteem

trajectory predict later shyness and aggressiveness. Previous research examining self-esteem

as a predictor of aggression shows that low self-esteem in late childhood indeed predicts

greater likelihood of aggression in early adolescence (Donnellan et al., 2005), suggesting that

self-esteem might be better modeled as a cause than as a consequence of shyness and

aggressiveness (Orth & Robins, 2014).

Limitations and Future Directions

The present study is not without limitation. First, our sample size (N = 240) is

relatively small. Yet, our analyses were conducted on data from 9 assessments collected over

25 years, and are modestly powered. Nevertheless, the robustness of findings and

generalizability of results should be interpreted with these considerations in mind. Although

we do not imagine that it is likely that researchers will conduct direct replications of the

current research given the long timespan, we encourage researchers to examine self-esteem

development from childhood to adolescence, because this is where the data are most lacking,

probably due to the complexity of measuring self-esteem adequately through a developmental

transition (e.g., in many longitudinal studies, like ours, self-esteem is measured by one

measure in childhood and another in adolescence). Recently, a lifespan measure of self-

esteem (Harris, Donnellan, & Trzesniewski, 2017) has been introduced that allows for the

comparison of the same measure across the lifespan.

Second, it is important to note that we used a linking procedure that assumes that the

measures used in our study are assessing the same constructs. Longitudinal studies that
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 26
ADULTHOOD
examine developmental processes from childhood to adulthood typically employ different

measures of the psychological construct of interest to ensure that the assessments are

developmentally appropriate. The costs of changing measures make statistical inferences

about how the construct of interest develops difficult. In the present study, it was important

that our measures change to match the developmental shift that occurred at the cusp of

adolescence, resulting in two different measures of each construct being used. For example,

the SPPC (Harter, 1985) uses simplified language, explicitly references children, and

employs a procedure designed to obtain self-ratings from youth. In contrast, the SDQ-III

(Marsh & O’Neill, 1985) references complex thoughts and feelings towards the self,

explicitly references the self, and uses a Likert-type response scale, making it appropriate for

adolescents and adults. However, because we made use of item response theory methods, we

were able to overcome the limitations of previous studies and examine the trajectories of self-

esteem and social acceptance from childhood to young adulthood, but under the assumption

that the different measures were assessing the same underlying construct.

Third, we did not differentiate types of shyness and aggressiveness in the current

study. In another study that examined the LOGIC data, shyness in response to strangers in

early childhood was not predictive of self-esteem change in middle childhood, but shyness in

the presence of peers did (Asendorpf & van Aken, 1994). Extant research also suggests that

when aggression is differentiated, relational aggression is sometimes predictive of greater

social acceptance, while physical aggression is generally predictive of lesser social

acceptance (Cillessen & Mayeux, 2004; Nelson, Robinson, & Hart, 2005).

Fourth, we used a piecewise growth model to characterize our data, which led us to

specify the childhood slope (i.e., ages 9, 10, and 12) with assessments that were closer

together than for the youth slope (i.e., ages 17, 23, and 29). On the one hand, it is possible

that a more precise picture of development is obtained when using more assessments during a
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 27
ADULTHOOD
shorter time span, and that a less precise picture of development is obtained when using

fewer assessments during a longer time span. On the other hand, it could be that a longer time

span allows for a longitudinal trend to be shown in a less noisy, more reliable manner.

However, previous theory and research regarding self-esteem development indicates that

normative changes in self-esteem occur the most in childhood and adolescence, and is more

uniform into young adulthood. Nevertheless, we encourage future research to include more

assessments during adolescence and young adulthood whenever possible.

Conclusion

In the present study, we examined the development of self-esteem in a sample of

individuals followed longitudinally from middle childhood to young adulthood. We applied

item response theory methods to link data from two different self-esteem measures, allowing

the present study to track self-esteem longitudinally from a developmental period that has

been neglected in the lifespan literature –middle childhood. We found that rank-order

stability of self-esteem was high in middle childhood, low in adolescence, and highest in

young adulthood. On average, self-esteem levels were relatively high in middle childhood,

decreased into adolescence, but increased into young adulthood. Furthermore, shyness and

aggressiveness as rated by parents, teachers, and observers in early childhood did not

influence self-esteem development.

Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 28
ADULTHOOD
References

Alsaker, F. D., & Olweus, D. (1992). Stability of global self-evaluations in early adolescence:

A cohort longitudinal study. Journal of Research on Adolescence, 2, 123–145.

doi: 10.1207/s15327795jra0202_2

Anderson, C.A., & Bushman, B.J. (2002). Human aggression. Annual Review of Psychology,

53, 27-51. doi: 10.1146/annurev.psych.53.100901.135231

Arunkumar, R., Midgley, C., & Urdan, T. (1999). Perceiving high or low home-school

dissonance: Longitudinal effects on adolescent emotional and academic well-being.

Journal of Research on Adolescence, 9, 441–466. doi: 10.1207/s15327795jra0904_4

Asendorpf, J. B. (1990). Development of inhibition during childhood: Evidence for

situational specificity and a two-factor model. Developmental Psychology, 26, 721–

730. doi: 10.1037/0012-1649.26.5.721

Asendorpf, J. B., & van Aken, M. G. (1994). Traits and relationship status: Stranger versus

peer group inhibition and test intelligence versus peer group competence as early

predictors of later self-esteem. Child Development, 65(6), 1786-1798.

doi:10.2307/1131294

Asendorpf, J. B., & van Aken, M. A. (1999). Resilient, overcontrolled, and undercontrolled

personality prototypes in childhood: Replicability, predictive power, and the trait-type

issue. Journal of Personality and Social Psychology, 77, 815–832.

doi: 10.1037/0022-3514.77.4.815

Baltes, P. B., Cornelius, S. W., & Nesselroade, J. R. (1979). Cohort effects in developmental

psychology. In J. R. Nesselroade & P. B. Baltes (Eds.), Longitudinal research in the

study of behavior and development (pp. 61–87). New York: Academic Press.

Bentler, P. M., & Chou, C. (1987). Practical issues in structural equation modeling.

Sociological Methods and Research, 16, 78-117. doi: 10.1177/0049124187016001004

Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 29
ADULTHOOD

Block, J., & Robins, R. W. (1993). A longitudinal study of consistency and change in self‐

esteem from early adolescence to early adulthood. Child Development, 64, 909–923.

doi: 10.1111/j.1467-8624.1993.tb02951.x

Boomsma, A. (1985). Nonconvergence, improper solutions, and starting values in LISREL

maximum likelihood estimation. Psychometrika, 50, 229-242.

Boomsma, A. (1982). Robustness of LISREL against small sample sizes in factor analysis

models. In Jöreskog, K. G. & Wold, H. [Eds.] Systems under indirection observation:

Causality, structure, prediction (Part 1). Amsterdam, Netherlands: North Holland,

149-173.

Browne, M. W. & Cudeck, R. (1993). Alternative ways of assessing model fit. In Bollen, K.

A. & Long, J. S. [Eds.] Testing structural equation models. Newbury Park, CA:

Sage, 136–162.

Caspi, A., Elder, G. H., & Bem, D. J. (1987). Moving against the world: Life-course patterns

of explosive children. Developmental Psychology, 23, 308–313. doi: 10.1037/0012-

1649.23.2.308

Caspi, A., Elder, G. H., & Bem, D. J. (1988). Moving away from the world: Life-course

patterns of shy children. Developmental Psychology, 24, 824–831. doi: 10.1037/0012-

1649.24.6.824

Caspi, A., & Shiner, R. (2011). Temperament and personality. In M. Rutter, D. Bishop, D.

Pine, S. Scott, J. Stevenson, E. Taylor & A. Thapar (Eds.), Rutter’s child and

adolescent psychiatry (5th ed., pp. 182). Malden, Massachusetts, USA: Blackwell

Publishing Limited.

Chalmers, R. P. (2012). mirt: A multidimensional item response theory package for the R

environment. Journal of Statistical Software, 48, 1–29.

Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 30
ADULTHOOD
Cheek, J. M., & Buss, A. H. (1981). Shyness and sociability. Journal of Personality and

Social Psychology, 41, 330–339. doi: 10.1037/0022-3514.41.2.330

Cillessen, A. H., & Mayeux, L. (2004). From censure to reinforcement: Developmental

changes in the association between aggression and social status. Child

Development, 75(1), 147-163. doi: 10.1111/j.1467-8624.2004.00660.x

Cooley, C. (1902). Human Nature and the Social Order.

Coplan, R. J., Prakash, K., O'Neil, K., & Armer, M. (2004). Do You 'Want' to Play?

Distinguishing Between Conflicted Shyness and Social Disinterest in Early

Childhood. Developmental Psychology, 40(2), 244-258. doi:10.1037/0012-

1649.40.2.244

Coplan, R. J., & Rubin, K. H. (2010). Social Withdrawal and shyness in childhood. In K. H.

Rubin & R. J. Coplan (Eds.), The Development of shyness and social withdrawal.

New York, NY: Guilford Press.

Crozier, W. (Ed.) (1995). Shyness and self‐esteem in middle childhood. British Journal of

Educational Psychology, 65, 85–95. doi: 10.1111/j.2044-8279.1995.tb01133.x

Curran, P. J., Hussong, A. M., Cai, L., Huang, W., Chassin, L., Sher, K. J., & Zucker, R. A.

(2008). Pooling data from multiple longitudinal studies: The role of item response

theory in integrative data analysis. Developmental Psychology, 44(2), 365-380.

doi:10.1037/0012-1649.44.2.365

de Ayala, R. J. (Ed.) (2009). Theory and practice of item response theory (Methodology in

the social sciences). New York, NY: Guilford Press.

Denissen, J. J. A., Asendorpf, J. B., & Van Aken, M. A. G. (2008). Childhood personality

predicts long-term trajectories of shyness and aggressiveness in the context of

demographic transitions in emerging adulthood. Journal of Personality, 76, 67–99.

doi: 10.111/j.1467-6494.2007.00480.x
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 31
ADULTHOOD
Denissen, J. J. A, Thomaes, S., & Bushman, B.J. (in press). Self-regulation and aggression:

Aggression-provoking cues, individual differences, and self-control strategies.

Donnellan, M. B., Kenny, D. A., Trzesniewski, K. H., Lucas, R. E., & Conger, R. D. (2012).

Using trait‐state models to evaluate the longitudinal consistency of global self‐esteem

from adolescence to adulthood. Journal of Research in Personality, 46(6), 634-645.

doi:10.1016/j.jrp.2012.07.005

Donnellan, M. B., Trzesniewski, K. H., Robins, R. W., Moffitt, T. E., & Caspi, A. (2005).

Low self-esteem is related to aggression, antisocial behavior, and delinquency.

Psychological Science, 16, 328–335. doi: 10.1111/j.0956-7976.2005.01535.x

Eccles, J., Wigfield, A., Harold, R. D., & Blumenfeld, P. (1993). Age and gender

differences in children's self- and task perceptions during elementary school. Child

Development, 64(3), 830-847. doi:10.2307/1131221

Eccles, J. S., Midgley, C., & Adler, T. (1984). Grade-related changes in the school

environment. Advances in Motivation and Achievement, 3, 238–331.

Embretson, S., & Reise, S. P. (2000). Item Response Theory for Psychologists. Mahwah,

N.J.: L. Erlbaum Associates.

Erol, R. Y., & Orth, U. (2011). Self-esteem development from age 14 to 30 years: A

longitudinal study. Journal of Personality and Social Psychology, 101, 607–619.

doi: 10.1037/a0024299.

Gest, S. D. (1997). Behavioral inhibition: Stability and associations with adaptation from

childhood to early adulthood. Journal of Personality and Social Psychology, 72(2),

467-475. doi:10.1037/0022-3514.72.2.467

Göttert, R., & Asendorpf, J. (1989). Eine deutsche Version des California- Child-Q-Sort,

Kurzform [A German short version of the California Child Q-set]. Zeitschrift für

Entwicklungspsychologie und Pädagogische Psychologie, 21, 70–82.

Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 32
ADULTHOOD
Harris, M. A., Donnellan, M. B., & Trzesniewski, K. H. (2017). Lifespan Self-Esteem Scale:

Initial Validation of a New Measure of Global Self-Esteem. Journal of Personality

Assessment.

Harris, M. A., Gruenenfelder-Steiger, A. E., Ferrer, E., Donnellan, M. B., Allemand, M.,

Fend, H., Conger, R. D., & Trzesniewski, K. H. (2015). Do parents foster self-

esteem? Testing the prospective impact of parent closeness on adolescent self-esteem.

Child Development, 86(4), 995-1013. doi: 10.1111/cdev.12356

Hart, D., Hofmann, V., Edelstein, W., & Keller, M. (1997). The relation of childhood

personality types to adolescent behavior and development: A longitudinal study of

Icelandic children. Developmental Psychology, 33(2), 195. doi: 10.1037/0012-

1649.33.2.195

Harter, S. (Ed.). (1985). The self-perception profile for children: Revision of the perceived

competence scale for children. Denver, CO: University of Denver.

Harter, S. (Ed.). (2012). Construction of the self: Developmental and sociocultural

foundations. New York, NY: Guilford.

Hoge, D. R., Smit, E. K., & Hanson, S. L. (1990). School experiences predicting changes in

self-esteem of sixth- and seventh-grade students. Journal of Educational Psychology,

82, 117–127. doi: 10.1037/0022-0663.82.1.117

Horney, K. (1950). Neurosis and human growth. New York: Norton.

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure

analysis: Conventional criteria versus new alternatives. Structural Equation

Modeling, 6(1), 1-55. doi:10.1080/10705519909540118

Huang, C. (2010). Mean-level change in self-esteem from childhood through adulthood:

Meta-analysis of longitudinal studies. Review of General Psychology, 14, 251–260.

doi: 10.1037/a0020543
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 33
ADULTHOOD
Hutteman, R., Hennecke, M., Orth, U., Reitz, A. K., & Specht, J. (2014). Developmental

tasks as a framework to study personality development in adulthood and old age.

European Journal of Personality, 28(3), 267-278.

Hutteman, R., Nestler, S., Wagner, J., Egloff, B., & Back, M. D. (2015). Wherever I may

roam: Processes of self-esteem development from adolescence to emerging adulthood

in the context of international student exchange. Journal of Personality and Social

Psychology, 108, 767–783. doi: 10.1037/pspp0000015

Kemple, K. M., David, G. M., & Wang, Y. (1996). Preschoolers’ creativity, shyness, and

self-esteem. Creativity Research Journal, 9, 317–326. doi:

10.1207/s15326934crj0904_3

Kerr, M. (2000). Childhood and adolescent shyness in long-term perspective: Does it matter?

in Shyness: Development, consolidation and change, Ed. W. Ray Crozier, Routledge:

NY. pp. 64-87.

Kjeldsen, A., Nilsen, W., Gustavson, K., Skipstein, A., Melkevik, O., & Karevold, E. B.

(2016). Predicting well‐being and internalizing symptoms in late adolescence from

trajectories of externalizing behavior starting in infancy. Journal of Research on

Adolescence, doi:10.1111/jora.12252

Kline, R. B. (2005). Principles and practice of structural equation modeling. (2nd ed.). New

York, NY: Guilford.

Kolen, M. J., & Brennan, R. L. (Eds.) (2004). Test equating, scaling, and linking. New York,

NY: Springer.

Kuster, F., Orth, U., & Meier, L. L. (2013). High self-esteem prospectively predicts better

work conditions and outcomes. Social Psychological and Personality Science, 4,

668–675. doi: 10.1177/1948550613479806

Ladd, G. W., & Burgess, K. B. (1999). Charting the relationship trajectories of aggressive,
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 34
ADULTHOOD
withdrawn, and aggressive/withdrawn children during early grade school. Child

Development, 70(4), 910-929. doi:10.1111/1467-8624.00066

Leary, M. R., Tambor, E. S., Terdal, S. K., & Downs, D. L. (1995). Self-esteem as

interpersonal monitor: The sociometer hypothesis. Journal of Personality and Social

Psychology, 68, 518–530. doi: 10.1037/0022-3514.68.3.518

Luan, Z., Hutteman, R. H., Denissen, J. J. A., Asendorpf, J. B., & van Aken, M. A. G. (2016).

See Me Through My Eyes: Adolescent-Parent Agreement in Personality Predicts

Later Self-Esteem Development. Manuscript submitted for publication.

Marsh, H. W., Craven, R. G., & Debus, R. (1991). Self-concepts of young children 5 to 8

years of age: Measurement and multidimensional structure. Journal of Educational

Psychology, 83, 377–392. doi: 10.1037/0022-0663.83.3.377

Marsh, H. W., & O'Neill, R. (1985). Self-Description Questionnaire III (SDQ III): The

construct validity of multidimensional self-concept ratings by adolescents. Journal of

Educational Measurement, 21, 153–174.

Mead, G. (1934). Mind, Self, and Society from the Standpoint of a Social Behaviorist.

University of Chicago Press: Chicago.

Muthén, L. K., & Muthén, B. O. (1998-2012). Mplus User’s Guide. Seventh Edition. Los

Angeles, CA: Muthén & Muthén.

Nelson, D. A., Robinson, C. C., & Hart, C. H. (2005). Relational and physical aggression of

preschool-age children: Peer status linkages across informants. Early Education &

Development, 16(2), 115-140. doi: 10.1207/s15566935eed1602_2

Orth, U. (2017). The family environment in early childhood has a long-term effect on self-

esteem: A longitudinal study from birth to age 27 years. Journal of Personality and

Social Psychology. doi: 10.1037/pspp0000143

Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 35
ADULTHOOD
Orth, U., Maes, J., & Schmitt, M. (2015). Self-esteem development across the life span: A

longitudinal study with a large sample from Germany. Developmental Psychology,

51, 248–259. doi: 10.1037/a0038481

Orth, U., & Robins, R. W. (2014). The development of self-esteem. Current Directions in

Psychological Science, 23, 381–387. doi: 10.1177/0963721414547414

Orth, U., Robins, R. W., & Widaman, K. F. (2012). Life-span development of self-esteem

and its effects on important life outcomes. Journal of Personality and Social

Psychology, 102, 1271–1288. doi: 10.1037/a0025558

Orth, U., Trzesniewski, K. H., & Robins, R. W. (2010). Self-esteem development from young

adulthood to old age: A cohort-sequential longitudinal study. Journal of Personality

and Social Psychology, 98, 645–658. doi: 10.1037/a0018769

Ostendorf, F. (1990). Sprache und Persönlichkeitsstruktur: Zur Validität des Fünf-

Faktoren-Modells der Persönlichkeit (language and personality structure: Validity of

the five-factor model of personality). Regensburg, Germany: S. Roderer.

R Core Team (2015). R: A language and environment for statistical computing. R Foundation

for Statistical Computing, Vienna, Austria.

Revelle, W. (2015). Psych: Procedures for Personality and Psychological Research,

Northwestern University, Evanston, Illinois, USA.

Richardson, C. B., Hitti, A., Mulvey, K. L., & Killen, M. (2014). Social exclusion: The

interplay of group goals and individual characteristics. Journal of Youth and

Adolescence, 43(8), 1281-1294. doi:10.1007/s10964-013-9967-8

Robins, R. W., Donnellan, M. B., Widaman, K. F., & Conger, R. D. (2010). Evaluating the

link between self-esteem and temperament in Mexican origin early adolescents.

Journal of Adolescence, 33, 403-410. doi: 10.1016/j.adolescence.2009.07.009

Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 36
ADULTHOOD
Robins, R. W., Trzesniewski, K. H., Tracy, J. L., Gosling, S. D., & Potter, J. (2002). Global

self-esteem across the life span. Psychology and Aging, 17, 423–434.

doi: 10.1037///0882-7974.17.3.423

Robins, R. W., & Trzesniewski, K. H. (2005). Self-esteem development across the lifespan.

Current Directions in Psychological Science, 14, 158–162.

doi: 10.1111/j.0963-7214.2005.00353.x

Rodriguez, D., Wigfield, A., & Eccles, J. S. (2003). Changing competence perceptions,

changing values: Implications for youth sport. Journal of Applied Sport Psychology,

15, 67–81. doi: 10.1080/10413200305403

Rosenberg, M., Schooler, C., & Schoenbach, C. (1989). Self-esteem and adolescent

problems: Modeling reciprocal effects. American Sociological Review, 1004–1018.

Rothbart, M. K. (2011). Becoming who we are. New York, NY: Guilford Press.

Rubin, K. H., Chen, X., McDougall, P., Bowker, A., & McKinnon, J. (1995). The Waterloo

Longitudinal Project: Predicting internalizing and externalizing problems in

adolescence. Development and Psychopathology, 7, 751–764.

doi: 10.1017/S0954579400006829

Rubin, K. H., Coplan, R. J., & Bowker, J. C. (2009). Social withdrawal in childhood. Annual

Review of Psychology, 60, 141–171. doi: 10.1146/annurev.psych.60.110707.163642

Samejima, F. (1969). Estimation of latent ability using a response pattern of graded

scores. Psychometrika Monograph Supplement, 34(4), 100.

Schinka, K. C., van Dulmen, M. M., Mata, A. D., Bossarte, R., & Swahn, M. (2013).

Psychosocial predictors and outcomes of loneliness trajectories from childhood to

early adolescence. Journal of Adolescence, 36(6), 1251-1260.

doi:10.1016/j.adolescence.2013.08.002
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 37
ADULTHOOD
Schneider, W., & Bullock, M. (Eds.). (2009). Human development from early childhood to

early adulthood: Findings from a 20 year longitudinal study. New York, NY:

Psychology Press.

Simmons, R. G., Rosenberg, F., & Rosenberg, M. (1973). Disturbance in the self-image at

adolescence. American Sociological Review, 553–568.

Steiger, A. E., Allemand, M., Robins, R. W., & Fend, H. A. (2014). Low and decreasing self-

esteem during adolescence predict adult depression two decades later. Journal of

Personality and Social Psychology, 106, 325–338. doi: 10.1037/a0035133

Trzesniewski, K. H., Donnellan, M. B., Moffitt, T. E., Robins, R. W., Poulton, R., & Caspi,

A. (2006). Low self-esteem during adolescence predicts poor health, criminal

behavior, and limited economic prospects during adulthood. Developmental

Psychology, 42, 381–390. doi: 10.1037/0012-1649.42.2.381

Trzesniewski, K. H., Donnellan, M. B., Robins, R. W. (2003). Stability of self-esteem across

the life span. Journal of Personality and Social Psychology, 84, 205–220.

doi: 10.1037/0022-3514.84.1.205

Twenge, J. M. (2006). Generation me: Why today’s young Americans are more confident,

assertive, entitled–and more miserable than ever before. New York, NY: Free Press.

Vaughn, B. E., Vollenweider, M., Bost, K. K., Azria-Evans, M. R., Snider, J. B. (2003).

Negative interactions and social competence for preschool children in two samples:

Reconsidering the interpretation of aggressive behavior for young children. Merrill-

Palmer Quarterly, 49, 245-278.

von Soest, T., Wichstrøm, L., & Kvalem, I. L. (2016). The development of global and

domain-specific self-esteem from age 13 to 31. Journal of Personality and Social

Psychology, 110, 592–608. doi: 10.1037/pspp0000060

Wegener, W. (1988). Kritik des Prestiges [Critique of prestige]. Opladen,

Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 38
ADULTHOOD
Germany: Westdeutscher Verlag.

Weinert, F. E., & Schneider, W. (Eds.) (1999). Individual development from 3 to 12:

Findings from the Munich longitudinal study. Cambridge, UK: Cambridge University

Press.

Wigfield, A., Eccles, J. S., Mac Iver, D., Reuman, D. A., & Midgley, C. (1991). Transitions

during early adolescence: Changes in children’s domain-specific self-perceptions and

general self-esteem across the transition to junior high school. Developmental

Psychology, 27, 552–565. doi: 10.1037/0012-1649.27.4.552

Figure 1

Self-Esteem from Age 9 to 29 as Assessed by the Self-Perception Profile for Children and
Self-Description Questionnaire-III

Note. N = 195. Effect sizes (Cohen’s d) are shown in the figure above. Light grey axis
denotes range of response categories for the Self-Perception Profile for Children (SPPC;
Harter, 1985); dark grey axis denotes additional response category for the Self-Description
Questionnaire III (SDQ-III; Marsh, 1985).
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 39
ADULTHOOD
Figure 2

Self-Esteem from Age 9 to 29 as Estimated by Linking Using Item Response Theory Methods

Note. N = 195. Theta represents latent self-esteem and is scaled so that it has a mean of 0 and
a standard deviation of 1. Effect sizes (Cohen’s d) are shown in the figure above.
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 40
ADULTHOOD

Figure 3

Path Diagram for Model of Self-Esteem Trajectory (Model 2)

.199* -.153* .019*

-.008 (-.286)
Self-
Esteem -.078* (-.693) Child hood .000 (.059) Youth
Intercep t Slope Slop e

.408* .031 .002*

0 0
1 3 0
3 3 0
3 *9.297
11
17

Age 9 Age 10 Age 12 Age 17 Age 23 Age 29

.382* .532* .451* .614* .497* .258*

Note. N = 195. We specified two latent slopes centered at the Age 9 self-esteem assessment,
and assigned a value of 0 to these coefficients. For the first latent slope, the five remaining
assessments were given values of 1, 3, 3, 3, 3, corresponding to the difference in years
between assessments for the first piece (i.e., the number of years between the Age 9 and 10
assessments was 1, the number of years between Age 9 and Age 12 was 3, and to
accommodate the specification of the second piece, the remaining assessments were fixed at
3). For the second latent slope, the Age 10 and Age 12 assessments were both given a value
of 0, reflecting the fact that the second piece began at the Age 17 assessment. The Age 17
assessment was allowed to be freely estimated, and the Age 23 and Age 29 assessments were
given values of 11 and 17, respectively, corresponding to the difference in years between
assessments after accounting for the fixed value of 3 for the same assessments in the first
piece (i.e., the number of years between the Age 9 and Age 23 assessments, and between the
Age 9 and Age 29 assessments were 14 and 20 years, respectively). Standardized estimates
are presented in parentheses. * p < .05, two-tailed
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG ADULTHOOD 41

Table 1

Fit Statistics for All Models

RMSEA
No. Description N χ2 df p est. 90% C.I. p CFI TLI

First-Order Autoregressive for Self-

1 195 6.325 9 .707 .000 .000 .061 .907 1.000 1.036
Esteem
2 Unconditional for Self-Esteem 195 12.307 11 .341 .025 .000 .081 .705 .990 .986
3 Shyness and Self-Esteem 232 28.149 28 .457 .005 .000 .051 .944 .999 .999
4 Aggressiveness and Self-Esteem 233 37.478 28 .109 .038 .000 .067 .717 .945 .929
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG ADULTHOOD 42

Table 2

Descriptive Statistics for Predictor Variables

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

1. Female 1.000
2. Shyness Towards Adults (Parent) .365* 1.000
3. Shyness Towards Peers (Parent) .280* .822* 1.000
4. Shyness Q-Sort (Teacher) .344* .486* .505* 1.000
5. Shyness Towards Strangers (Observer) .321* .645* .542* .634* 1.000
6. Wait-and-Hover (Observer) .258* .245* .294* .538* .305* 1.000
7. Contact Initiation Latency (Observer) .191 .574* .570* .333* .387* .090 1.000
8. Aggressiveness Towards Peers (Parent) -.317* -.189 -.103 -.220 -.073 -.090 -.071 1.000
9. Aggressiveness Q-Sort (Teacher) -.388* -.335* -.253 -.748* -.613* -.305* -.106 .252 1.000
10. Aggressive Attacks among
All Own Initiations (Observer) -.422* -.254* -.291* -.347* -.171 -.401* -.124 .177 .351* 1.000
Mean .458 3.723 3.058 .027 3.346 21.073 308.283 2.833 -.230 2.344
SD .505 1.224 1.039 .268 1.159 11.842 246.876 .713 .344 3.114

Note. Correlation table computed using all available data in Mplus (N = 192). All variables are aggregated across Age 4, Age 5, Age 6
assessments, with the exception of observer ratings of Shyness Towards Strangers and Contact Initiation Latency, which did not have an Age 5
assessment. Wait-and-Hover and Contact Initiation Latency are in seconds. * p < .05.
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG ADULTHOOD 43

Table 3

Estimates for First-order Autoregressive Model (Model 1)

Parameter Estimates

Original Annualized

Regressions Age Range Interval B SE B b p **Power B b

S2 θ on S1 θ 9 to 10 1 .820 .196 .841 < .001 .992 .820 .841

S3 θ on S2 θ 10 to 12 2 .494 .144 .570 .001 .981 .703 .755
S4 θ on S3 θ* 12 to 17 5 .531 .184 .427 .004 .837 .881 .844
S5 θ on S4 θ 17 to 23 6 .779 .178 .742 < .001 .998 .959 .951
S6 θ on S5 θ 23 to 29 6 .834 .157 .965 < .001 1.000 .970 .994

Note. N = 195. S denotes the latent status indicated by the manifest theta estimate. Age range and interval are in years. Parameter estimates are
annualized for time interval by taking the nth root of the estimate, where n is equal to the length of the interval (i.e., number of years between
assessments). * The self-esteem measure changed between S3 and S4. Parameter estimates for full model are shown in Supplemental
Information, Table 4. **A Montecarlo simulation was conducted to determine the power of each estimated parameter model output entered as
population values; it is possible to obtain a non-positive definite model implied variance-covariance matrix, which means that the standard errors
of the parameter estimates are not estimable (a common example of this would be if predictors in a regression are too highly correlated), and this
occurred .15 percent of the time. For significant effects, we expected power to be high (e.g., the regression of S2 on S1); for non-significant
effects, we expected power to be low (e.g., the covariance between the childhood and youth slopes).
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 44
ADULTHOOD

Table 4

Estimates for Model of Self-Esteem Trajectory (Model 2)

B SE B b p Power*

Model
Childhood slope (C S) by
θ Age 9 0 -- -- -- --
θ Age 10 1 -- -- -- --
θ Age 12 3 -- -- -- --
θ Age 17 3 -- -- -- --
θ Age 23 3 -- -- -- --
θ Age 29 3 -- -- -- --
Youth slope (Y S) by
θ Age 9 0 -- -- -- --
θ Age 10 0 -- -- -- --
θ Age 12 0 -- -- -- --
θ Age 17 9.279 2.214 -- < .001 .991
θ Age 23 11 -- -- -- --
θ Age 29 17 -- -- -- --
Covariances
C S with S-EI -.078 .032 -.693 .014 .693
Y S with S-EI -.008 .004 -.286 .070 .441
Y S with C S .000 .003 .059 .864 .059
Means
S-E I .199 .060 -- .001 .913
CS -.153 .025 -- < .001 1.000
YS .019 .005 -- < .001 .963
Variances
S-E I .408 .088 -- < .001 .999
CS .031 .017 -- .064 .422
YS .002 .001 -- .008 .847

Note. N = 195. θ represents latent trait of self-esteem as estimated by linking procedure.

Parameter estimates for full model are shown in Supplemental Information, Table 5. *A
Montecarlo simulation was conducted to determine the power of each estimated parameter
model output entered as population values; it is possible to obtain a non-positive definite
model implied variance-covariance matrix, which means that the standard errors of the
parameter estimates are not estimable (a common example of this would be if predictors in a
regression are too highly correlated), and this occurred .01 percent of the time. For significant
effects, we expected power to be high (e.g., the mean of the self-esteem intercept); for non-
significant effects, we expected power to be low (e.g., the covariance between the childhood
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG 45
ADULTHOOD

and youth slopes). Dashed lines indicate that parameters were not estimated or that
standardized estimates are not presented.
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG ADULTHOOD 46

Table 5

Shyness and Aggressiveness as Predictors of Self-Esteem Trajectory (Model 3 and Model 4)

Shyness (Model 3) Aggressiveness (Model 4)

Parameter B SE B b p Power* Parameter B SE B b p Power*

Regressions Regressions
Self-esteem intercept on Self-esteem intercept on
Shyness -.062 .182 -.044 .732 .060 Aggressiveness .231 .193 .176 .231 .238
Childhood slope on Childhood slope on
Shyness -.046 .070 -.120 .513 .101 Aggressiveness -.156 .097 -.420 .109 .548
Youth slope on Youth slope on
Shyness .001 .015 .014 .931 .048 Aggressiveness .012 .015 .137 .428 .126

Note. Ns = 232 and 233 for Model 3 and Model 4, respectively. Models were estimated separately. Parameter estimates for full model are shown
in Supplemental Information, Table 6 and Table 7. *Montecarlo simulations were conducted to determine the power of each estimated parameter
model output entered as population values; it is possible to obtain a non-positive definite model implied variance-covariance matrix, which
means that the standard errors of the parameter estimates are not estimable (a common example of this would be if predictors in a regression are
too highly correlated), and this occurred .1 percent and 3.1 percent of the time for Model 3 and Model 4, respectively. For significant effects, we
expected power to be high; for non-significant effects, we expected power to be low.
Running head: SELF-ESTEEM FROM MID-CHILDHOOD TO YOUNG ADULTHOOD 47

• We examined self-esteem in a German sample (N = 240) followed over two decades.

• We linked data from two self-esteem measures using item response theory methods.
• Self-esteem was most stable in young adulthood, and least stable in adolescence.
• Self-esteem declined from childhood to adolescence, and increased into young adulthood.
• Early childhood shyness and aggressiveness was not linked to self-esteem development.