HHS Public Access

Author manuscript
J Res Pers. Author manuscript; available in PMC 2017 August 01.
Author Manuscript

Published in final edited form as:

J Res Pers. 2016 August ; 63: 84–94. doi:10.1016/j.jrp.2016.05.005.

Are Parenting Practices Associated with the Development of

Narcissism? Findings from a Longitudinal Study of Mexican-
origin Youth
Eunike Wetzel1,2 and Richard W. Robins2
1Department of Psychology, University of Konstanz
Author Manuscript

2Department of Psychology, University of California, Davis

Abstract
Narcissism is an important and consequential aspect of personality, yet we know little about its
developmental origins. Using data from a longitudinal study of 674 Mexican-origin families, we
examined cross-lagged relations between parenting behaviors (warmth, hostility, monitoring) and
narcissism (superiority, exploitativeness). Parental hostility at age 12 was associated with higher
levels of exploitativeness at age 14, whereas parental monitoring at age 12 was associated with
lower levels of exploitativeness at age 14. These effects replicated across three different parenting
measures: child reports, spouse reports, and behavioral coding of parent-child interactions. None
of the parenting dimensions was related to superiority, suggesting that parenting practices are more
strongly related to the maladaptive than the adaptive component of narcissism.
Author Manuscript

Keywords
narcissism; exploitativeness; superiority; parental warmth; parental hostility; parental monitoring

Narcissism encompasses a wide range of characteristics, including feelings of superiority, a

sense of grandiosity, exhibitionism, exploitative behaviors in the interpersonal domain,
feelings of entitlement, fantasies of unlimited power, success, or beauty, and a lack of
empathy. Despite the abiding, and even growing, interest in narcissism, we know little about
its developmental origins and childhood correlates. Given the link between narcissism and
adjustment problems during childhood and adolescence (Barry, Frick, Adler, & Grafeman,
2007; Barry, Frick, & Killian, 2003; Washburn, McMahon, King, Reinecke, & Silver, 2004),
Author Manuscript

it is important to understand the socialization processes that contribute to the development of

narcissism. A large body of theoretical and empirical work suggests that parent socialization
practices play a central role in shaping children’s developmental trajectories (Bornstein,
2006; Parke & Buriel, 2006).

Corresponding Author: Eunike Wetzel, Department of Psychology, University of Konstanz, 78457 Konstanz, Germany,
eunike.wetzel@uni-konstanz.de.
Publisher's Disclaimer: 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 citable 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.
 Wetzel and Robins Page 2

The present study uses longitudinal data from 674 Mexican-origin families to examine
Author Manuscript

prospective effects of parenting on the development of narcissism. There are strong

theoretical reasons to expect that parenting practices play an important role in shaping
narcissistic tendencies. The first influential theory linking parenting to narcissism is based
on psychodynamic theory and was articulated by Kohut (1971, 1977) and Kernberg (1975),
although it can be traced back to Freud (1914). According to Kohut and Kernberg, parental
hostility and excessive criticism, along with a lack of warmth and responsiveness, lead to
feelings of inadequacy in children and impede the development of positive self-regard.
Children try to compensate for these feelings of inadequacy by inflating their self-worth and
constantly seeking approval and admiration from others. Narcissism can thus be seen as a
defensive reaction to parenting behaviors that convey disapproval and lack of acceptance and
support. The second influential theory, social learning theory, also posits that parenting
practices shape the development of child narcissism. In contrast to psychodynamic theory,
Author Manuscript

this perspective based on work by Millon (1969, 1981) posits that extremely permissive
parenting behavior, and in particular excessive parental indulgence and approval, are
responsible for the development of narcissistic tendencies. According to social learning
theory, children directly learn the behavior modeled by their parents and internalize their
parents’ beliefs that they are superior to others and entitled to special treatment and therefore
develop increased narcissism.

Despite the rich theoretical literature on parenting and narcissism, we know of only two
longitudinal studies on the topic. Cramer (2011) showed that children raised by authoritative
and permissive parents (high responsiveness) exhibited more adaptive narcissistic
tendencies, such as superiority and grandiosity, whereas children raised by authoritarian
parents (low responsiveness) were less likely to exhibit such traits. In contrast, Cramer did
not find any main effects of parenting on the more maladaptive components of narcissism,
Author Manuscript

such as exploitativeness and entitlement. However, this study relied on self-reports of

parenting and involved a very small sample (e.g., 89 mothers reported on their parenting).

In a more recent longitudinal study, Brummelman et al. (2015) examined the influence of
parental indulgence (termed “parental overvaluation” in their study) and parental warmth on
the development of narcissism in late childhood. Parental indulgence (assessed via parent
self-report) predicted increases in narcissism from age 10 to 12, whereas parental warmth
(assessed via child-report and parent self-report) was not related to narcissism. No reciprocal
effects of child narcissism on parenting were found, suggesting the association is
unidirectional, going from parenting to narcissism rather than vice versa. Note, however, that
Brummelman et al. examined overall narcissism levels and did not differentiate between
facets of narcissism. An emerging body of research suggests that different facets of
Author Manuscript

narcissism have highly divergent antecedents and consequences. Facets encompassing

feelings of superiority and having a grandiose self-concept appear to be at least partially
adaptive since they have been linked to positive outcomes such as high self-esteem and
emotional stability and low loneliness and depression (Rhodewalt & Morf, 1995; Sedikides,
Rudich, Gregg, Kumashiro, & Rusbult, 2004; Trzesniewski, Donnellan, & Robins, 2008). In
contrast, facets encompassing a sense of entitlement and exploitativeness appear to be
maladaptive since they have been linked to negative outcomes such as trait anger,
aggression, counterproductive work behaviors, and dysfunctional interpersonal relationships

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 3

(Bushman & Baumeister, 1998; Campbell & Campbell, 2009; Campbell, Foster, & Finkel,
Author Manuscript

2002; Miller et al., 2009). These findings highlight the need to distinguish the facets of
narcissism.

In addition to examining facets of narcissism, it is also important to consider self-esteem

when studying narcissism. Narcissism and self-esteem are conceptually related but distinct
traits. Although both entail a positive evaluation of the self, individuals with high self-
esteem are assumed to generally like and accept themselves, whereas narcissists are assumed
to have inflated views of their worth, and a compulsive need to be better than others,
presumably as a defense against underlying feelings of inadequacy. Consequently, whereas
high self-esteem involves seeing oneself as “a person of worth, at least on an equal basis
with others” (sample item from the Rosenberg Self-esteem Scale; Rosenberg, 1979),
narcissism involves feeling superior to others, and carries with it a pattern of interpersonally
toxic tendencies such as exploitativeness and contempt toward others. Not surprisingly,
Author Manuscript

although measures of the two constructs tend to be moderately correlated (e.g., Brown &
Zeigler-Hill, 2004; Paulhus, Robins, Trzesniewski, & Tracy, 2004), they have quite different
relations with other constructs; for example, antisocial behavior, aggression, and hostility are
positively related to narcissism, but negatively related to self-esteem (Tracy, Cheng, Robins,
& Trzesniewski, 2009). The conceptual and empirical overlap between narcissism and self-
esteem raises the possibility that prior findings concerning the association between parenting
and narcissism may have been confounded by self-esteem. Neither Cramer (2011) nor
Brummelman et al. (2015) controlled for self-esteem when they examined associations
between parenting and narcissism, raising the possibility that their findings were driven by
the variance in narcissism that reflects genuine self-esteem rather than narcissistic
tendencies1. Narcissism with self-esteem partialed out can be conceptualized as a more pure
measure of narcissistic tendencies, with any aspects of genuine self-esteem removed. Thus,
Author Manuscript

in our models investigating parenting practices and narcissism, we will include self-esteem
as a control variable. By doing so, we will ensure that any observed associations with
narcissism are due to narcissism itself, and not its overlap with self-esteem.

The present study extends previous research on the development of narcissism by

investigating longitudinal relations between parenting practices and narcissism, using data
from a sample of 674 Mexican-origin youth followed from age 12 to 16. We extend previous
research in several important ways. First, we examined a more comprehensive set of
parenting dimensions, including the three most commonly studied components of parenting
(hostility, warmth, monitoring). Second, we assessed parenting using three different
methods: child report, spouse report, and behavioral coding of parent-child interactions.
Third, we examined effects separately for two important components of narcissism,
Author Manuscript

exploitativeness and superiority, as well as for overall narcissism levels. Fourth, we analyzed
data collected in a cohort-sequential longitudinal design with three waves of data spanning
five years. Fifth, we applied latent variable modeling techniques to investigate reciprocal
relations between parenting practices and narcissism, thereby controlling for measurement

1Although Brummelman et al. (2015) did not control for self-esteem when they examined the association between parenting and
narcissism, they did examine relations between parenting and self-esteem, and found that child-reported (but not parent-reported)
parental warmth was reciprocally related to self-esteem.

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 4

error. Sixth, we examined the effects of narcissism while controlling for its overlap with
Author Manuscript

self-esteem. Finally, we investigated the association between parenting and narcissism in an

important but understudied ethnic minority group, Mexican-origin families.

Hypotheses
We hypothesized that higher levels of parental hostility would be related to higher levels of
narcissism in adolescents, based on psychodynamic theories about the origins of narcissism
(Kernberg, 1975; Kohut, 1977). Although we know no empirical studies (longitudinal or
cross-sectional) that have investigated the effects of parental hostility, a study on parental
coldness, which is conceptually similar to parental hostility, found a positive association
with narcissism (Otway & Vignoles, 2006). Note, however, that if parental hostility is
assumed to be simply the converse of overvaluing a child, then we might expect hostility to
be related to lower levels of narcissism, based on Brummelman et al.’s (2015) finding that
Author Manuscript

overvaluation is related to higher narcissism. This is a situation where the social learning
perspective conflicts with the psychodynamic perspective. The former assumes a relatively
direct translation of feedback from one’s parents (“my parents think I’m perfect therefore I
think I’m perfect”; “my parents think I’m terrible therefore I think I’m terrible”), whereas
the latter assumes a defensive reaction (“my parents think I’m terrible, and that makes me
feel worthless so I am going to try to convince myself and others that I’m perfect”). Thus,
our hypothesis of a positive relationship between parental hostility and narcissism is derived
from psychodynamic perspective.

We also hypothesized that higher levels of parental monitoring would be related to lower
levels of narcissism. This hypothesis is based on concurrent studies documenting this
association (Barry et al., 2007; Horton, Bleau, & Drwecki, 2006; Miller & Campbell, 2008).
Author Manuscript

Furthermore, a lack of monitoring can be understood as part of a permissive, indulgent

parenting style (Horton, 2011). Thus, a negative relationship between parental monitoring
and narcissism can also be predicted from social learning theory and Brummelman et al.’s
finding that parental indulgence is positively associated with narcissism. A large body of
research has demonstrated that parental monitoring decreases risk for a wide range of
negative adolescent outcomes, and thus we would expect it to impede the development of
narcissistic tendencies (Barber, Olsen, & Shagle, 1994; Dishion & McMahon, 1998; Pettit,
Laird, Dodge, Bates, & Criss, 2001), at least the maladaptive component. Finally, we did not
expect to find a relation between parental warmth and narcissism, based on Brummelman et
al.

In sum, we hypothesized:
Author Manuscript

H1: Parental hostility at age 12 and 14 will be positively related to narcissism

at age 14 and 16, even after controlling for prior levels of narcissism and
concurrent associations with parental hostility.

H2: Parental monitoring at age 12 and 14 will be negatively related to

narcissism at age 14 and 16, even after controlling for prior levels of
narcissism and concurrent associations with parental monitoring.

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 5

H3: The effects specified in Hypotheses 1 and 2 will hold after controlling for
Author Manuscript

self-esteem and gender.

We did not make separate predictions for the exploitativeness and superiority facets because
prior research and theory concerning the effects of parenting has focused on the overall
narcissistic profile. Note that these hypotheses are based on the assumption that narcissism is
still undergoing development during adolescence. If narcissism emerges and stabilizes
earlier than the time period examined in the present study, we would not expect any of the
parenting measures to predict narcissism. However, it seems likely that the socialization
processes specified by both social learning theory and psychodynamic theory continue into
adolescence. For example, the socialization processes described by social learning theory
imply that narcissistic tendencies are likely to continue to develop across the lifespan, and
certainly within the late childhood to adolescence period covered by the present study.
Although classic psychodynamic theories imply that narcissism first emerges earlier, the
Author Manuscript

defensive processes that theoretically give rise to narcissistic tendencies are likely to be
particularly relevant during and following the transition into adolescence (the time period
covered by the present study), when youth face considerable socioemotional difficulties that
often trigger feelings of insecurity and inadequacy.

Method
Sample
Data come from the California Families Project, an ongoing longitudinal study of 674
Mexican-origin children (50% female) and their families. The children were drawn at
random from student rosters provided by two school districts in metropolitan areas of
Northern California. The focal child had to be in the 5th grade, of Mexican origin, and living
Author Manuscript

with his or her biological mother. Of the eligible families, 73% agreed to participate. The
sample has been assessed annually since 2006, when the children were 10.8 (SD = .61) years
of age on average. Trained interviewers visited the participants at home. The interviews were
conducted either in Spanish or in English, depending on the preference of the participant.

The present study used data collected when the children were 12, 14, and 16 years old, when
most of the key measures were available. The retention rate in the age 16 assessment was
90% (relative to the original sample of 674). Attrition analyses showed that families who
participated in the age 16 assessment did not differ significantly from nonparticipating
families in child gender, parent education, and family income, all ps > .05, or on any of the
constructs investigated in this study, namely narcissism (overall and facet-level), parental
warmth, parental hostility, parental monitoring, and self-esteem (all ps > .05).
Author Manuscript

Measures
Descriptive statistics and alpha reliability coefficients for all measures are shown in Table 1.
All measures were available at ages 12, 14, and 16, except for the narcissism scale, which
was available at ages 14 and 16, and the observational assessments of parenting, which were
available at age 12.

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 6

Narcissistic Personality Questionnaire for Children – Revised (NPQC-R)—The

Author Manuscript

NPQC-R (Ang & Raine, 2008) is a self-report instrument designed to assess narcissism in
children and adolescents. It includes two subscales, Superiority and Exploitativeness, which
are assessed with six items each. Superiority captures the grandiose aspects of narcissism
including feelings of superiority, vanity, and inflated self-views. A sample item on the
superiority scale is “I am going to be a great person.” Exploitativeness captures the
interpersonally maladaptive aspects of narcissism including interpersonal exploitativeness,
feelings of entitlement, and manipulativeness. A sample item on the exploitativeness scale is
“I am good at getting people to do things my way.” Participants responded on a five-point
rating scale ranging from not at all like me (1) to completely like me (5).

Rosenberg Self-esteem Scale (RSE)—The RSE (Rosenberg, 1965, 1979) consists of

10 items assessing global self-esteem such as “On the whole, I am satisfied with myself”.
Participants rated how well the items described them on a scale ranging from totally disagree
Author Manuscript

(1) to totally agree (4).

Behavioral Affect Rating Scale (BARS)—The BARS (Conger, 1989a) assesses

parental warmth (9 items) and parental hostility (13 items). The parental warmth scale
reflects how often the parent displayed warmth toward the child, for example by listening
carefully to the child’s point of view. The parental hostility scale reflects the frequency of
hostile behavior toward the child, for example by ignoring the child when the child tries to
talk to the parent. Responses to the BARS items were collected from two informants: the
child and the spouse. That is, the mother’s parenting behaviors were rated by the child and
father whereas the father’s parenting behaviors were rated by the child and mother. The
ratings were made with respect to behavior displayed during the past three months on a
response scale from almost never or never (1) to almost always or always (4).
Author Manuscript

Iowa Parenting Scale (IPS)—The IPS (Conger, 1989b) assesses parenting practices
mainly with respect to discipline. Parental warmth (9 items) was assessed by combining
items related to positive reinforcement (e.g., “When you have done something your mom
likes or approves of, how often does she let you know she is pleased about it?”) and
inductive reasoning(e.g., “How often does your mom give you reasons for her decisions?”).
The rating scale was the same as the BARS rating scale. As with the BARS, ratings by the
child and the spouse were collected. For the child-report data, the correlation of IPS
maternal (paternal) warmth with BARS maternal (paternal) warmth ranged from .72 to .75 (.
79 to .80) across assessments. For the parent-reports, the correlation of IPS maternal
(paternal) warmth with BARS maternal (paternal) warmth ranged from .65 to .76 (.78 to .81)
across assessments.
Author Manuscript

Parental Monitoring of Child (PMC)—The PMC (Small & Kerns, 1993) assesses the
extent to which the parents are informed about how and with whom their child spends their
time (e.g., “Your mom knew who your friends were.”, “Your mom knew where you were
and what you were doing.”). For 9 out of the 14 items, a four-point rating scale ranging from
almost never or never to almost always or always was used. For five items addressing the
parent’s monitoring of social plans such as “If you were going to get home late, you were

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 7

expected to call her”, a fifth response category was introduced to reflect that the child was
Author Manuscript

not allowed to show the described behavior (e.g., not allowed to stay out late). Parental
monitoring was assessed with respect to the past three months via child-report, parent self-
report, and spouse-report.

Observational assessments of parenting—During the interviewer visits, structured

interaction tasks (mother-child, father-child) were videotaped in the families’ homes and
later rated by trained coders on a wide range of behavioral dimensions adapted from the
Iowa Family Interaction Rating Scales (Melby et al., 1998). Different coders rated each
parent’s behavior.

To parallel the BARS, IPS, and PMC scales, we used the observational data to assess
parenting behaviors related to hostility, warmth, and monitoring. Hostility reflects the degree
to which the focal parent displayed hostile, angry, critical, or disapproving behavior toward
Author Manuscript

the child during the interaction. Warmth reflects the degree to which the focal parent
expressed liking, appreciation, praise, care, concern, or support for the child. Monitoring
reflects the degree to which the parent displayed knowledge of and pursued information
about the child’s life and daily activities during the interaction.

All scales were rated on a nine-point scale ranging from not at all characteristic (1) to
mainly characteristic (9). Most interactions were rated by a single coder. For a random
selection of 25% of the interaction tasks, ratings from two coders were available in which
case the averaged score across raters was analyzed. The intraclass correlation between
coders was .86 (.85) for maternal (paternal) hostility, .80 (.66) for maternal (paternal)
warmth, and .68 (.67) for maternal (paternal) monitoring.
Author Manuscript

Analyses
Overview—We first tested for longitudinal measurement invariance and then estimated
latent cross-lagged panel models to investigate reciprocal relations between narcissism (total
score, superiority, exploitativeness) and parenting practices. All analyses were conducted in
Mplus (Version 7.11; Muthén & Muthén, 1998–2014) using maximum likelihood estimation
robust to non-normality (denoted MLR in Mplus). To deal with missing data, we used full
information maximum likelihood estimation to fit models directly to the raw data (Schafer &
Graham, 2002). Model fit was evaluated by the comparative fit index (CFI; Bentler, 1990),
the Tucker-Lewis index (TLI; Tucker & Lewis, 1973), and the root mean square error of
approximation (RMSEA; Steiger, 1990). Good fit is indicated by values above .95 for the
CFI and TLI, and values below .05 for the RMSEA (Hu & Bentler, 1999).
Author Manuscript

Longitudinal measurement invariance—Measurement invariance was evaluated

following the procedure outlined in Meredith (1993) and Widaman and Reise (1997). That
is, starting with a baseline model of configural invariance where all parameters were free to
vary across waves of assessment, we first constrained the factor loadings to be equal across
ages (metric invariance) and then additionally constrained the indicator intercepts to be
equal across ages (scalar invariance). These nested models were tested using a χ2 difference
test based on rescaled log-likelihood values (due to the MLR estimator) and the difference in
CFI between models (ΔCFI). If the χ2 difference test indicated that the more restrictive

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 8

model fit significantly worse than the less restrictive model at α = .01 and ΔCFI was > .002
Author Manuscript

(Meade, Johnson, & Braddy, 2008), one parameter at a time was freed based on the
modification indices and the χ2 difference test and check of ΔCFI were repeated until both
criteria were fulfilled.

With respect to child-report data, full scalar invariance held for overall narcissism,
superiority, exploitativeness, BARS maternal warmth, BARS maternal hostility, BARS
paternal warmth, BARS paternal hostility, and maternal monitoring. Partial scalar invariance
held for self-esteem, IPS maternal warmth, IPS paternal warmth, and paternal monitoring.
With respect to the parent-report data, mother-report of paternal monitoring showed full
scalar invariance and all other constructs (BARS, IPS, and PMC parenting scales) showed
partial scalar invariance. Thus, overall, the same constructs were being measured over time,
although for some constructs one or two of the indicators were not fully invariant across
waves.
Author Manuscript

Latent cross-lagged panel models (LCLP)—Figure 1 shows the basic structure of the
models to be tested. Each model includes one of the three parenting dimensions at age 12,
14, and 16 and either overall narcissism or the two narcissism facets at age 14 and 16.

Three parcels consisting of two to five items (depending on the length of the scale) were
used as indicators for each latent variable. The items were randomly assigned to parcels.
Indicators based on the same items were correlated across waves (e.g., indicator 1 for
exploitativeness at age 14 was correlated with indicator 1 for exploitativeness at age 16).
First-order and second-order stability paths were included (e.g., maternal warmth at age 12
predicted maternal warmth at age 14 and maternal warmth at age 16). First-order cross-
lagged regression paths were modeled between constructs, for example narcissism at age 16
Author Manuscript

was regressed on maternal warmth at age 14 and maternal warmth at age 16 was regressed
on narcissism at age 14. Different constructs within a wave were correlated (e.g., narcissism
at age 14 with maternal warmth at age 14). The regression paths were allowed to vary over
time because model comparisons showed that the constraint of time-invariant paths
deteriorated model fit.

The analyses were conducted separately for child-report and parent-report of BARS, IPS,
and PMC scales as well as for observational assessments of parental warmth, parental
hostility, and parental monitoring. Gender and self-esteem were included as covariates in all
models. Latent variables for self-esteem at age 12, age 14, and age 16 were included in the
latent cross-lagged panel models. Thus, self-esteem is a time varying covariate.

Results
Author Manuscript

Gender Differences and Stability over Time

Narcissism did not show any gender differences, either at the overall or facet level. At all
ages, girls reported higher paternal warmth than boys on the BARS (but not IPS) scale and
higher maternal warmth on the IPS (but not BARS) scale; however, neither gender difference
held for parent reported warmth. Girls reported higher maternal monitoring than boys at all
ages and higher paternal monitoring at ages 14 and 16. For the mothers’ self-report of their

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 9

monitoring of the child, monitoring scores were also higher for girls at age 14 and 16. No
Author Manuscript

significant gender differences were found for mothers reporting on paternal monitoring or
fathers reporting on their own monitoring and maternal monitoring. Boys reported higher
self-esteem than girls at age 14 (t = −2.87, p = .004), but not at age 12 or at age 16.

Overall narcissism and the superiority facet were moderately stable from age 14 to 16, with
average stability coefficients of .58 (range .57 – .59) for the total scale and .61 (range .61 – .
62) for the superiority facet. In contrast, the exploitativeness facet was less stable with an
average stability coefficient of .46 (range .46 – .47). The first-order (age 12 to 14 and 14 to
16) stability coefficients for the parenting dimensions ranged (across waves and child- and
parent-reports) from .44 to .75 for paternal hostility, .31 to .63 for maternal hostility, .45 to .
75 for paternal warmth, .42 to .53 for maternal warmth, .46 to .72 for paternal monitoring,
and .41 to .63 for maternal monitoring. The second-order stability coefficients, spanning age
12 to 16, were, as expected given that they control for the 1st order stabilities, much lower,
Author Manuscript

with an average of .20 for parental hostility, .24 for parental warmth, and .19 for parental
monitoring.

Reciprocal Relations between Parenting and Narcissism

As shown in Table 2, the overall fit of all models tested was good (CFI ≥ .96, TLI ≥ .95,
RMSEA ≤ .04). Standardized regression coefficients for cross-lagged effects from all
models are depicted in Tables 3–5. Below we organize the results by parenting dimension;
within each section, we describe the results separately for the narcissism total score,
exploitativeness, and superiority.

Parental hostility—Parental hostility at age 12 predicted narcissism at age 14 for both

mothers and fathers. Importantly, this effect was consistent across child- and parent-report as
Author Manuscript

well as observational data, although in the observational data the effect was only significant
for maternal hostility. Thus, higher parental hostility at age 12 was associated with children
reporting higher narcissism at age 14. Thus, H1 was partly confirmed since this relationship
held for age 12 to age 14, but not age 14 to age 16. Furthermore, H3 was confirmed since
this effect was significant with both gender and self-esteem entered as control variables.
Facet-level analyses revealed that this effect was mainly attributable to the exploitativeness
facet of narcissism: Consistently a cross-lagged effect between parental hostility at age 12
and exploitativeness at age 14 appeared in models differentiating the two narcissism facets
(see Table 3). Thus, for both mothers and fathers, a higher degree of hostility towards their
child at age 12 was related to their child reporting being more exploitative at age 14. As for
overall narcissism, this finding held not only across maternal and paternal hostility, but also
across different informants on the BARS (child and spouse) as well as for the observational
Author Manuscript

data, though the effect was only marginally significant for observational data on paternal
hostility (β = .10, p = .12). For example, for maternal hostility, the child-report BARS
yielded a beta of .29, the parent-report BARS yielded a beta of .18 and the observational
rating yielded a beta of .17. Parental hostility at age 14 did not significantly predict
exploitativeness or overall narcissism at age 16. Furthermore, the cross-lagged regressions
between parental hostility and superiority were non-significant in all cases. There were no
notable differences, in either effect sizes or significance levels, between the models

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 10

including self-esteem as a control variable and the models not including self-esteem as a
Author Manuscript

control variable.

Parental warmth—Parental warmth was not related to overall narcissism, as expected. At

the facet-level, there were also no consistent cross-lagged effects on either exploitativeness
or superiority. One effect that replicated across the BARS and IPS scales was a negative
relationship between paternal warmth at age 12 and exploitativeness at age 14 (β = −.12, p
= .015 for BARS and β = −.11, p = .036 for IPS). However, this effect was only significant
for the spouse-report of paternal warmth and not for the child-report and it was also not
found in the observational data. Parental warmth at age 12 predicted superiority at age 14 for
IPS maternal warmth (β = .14, p = .035), for IPS paternal warmth (β = .19, p = .001), and
for BARS paternal warmth (β = .12, p = .031). However, in this case, the effect was only
present for child-report data and not for parent-report data and it was also not present in the
observational ratings of parental warmth (see Table 4). In some (but not all) of the models
Author Manuscript

without self-esteem as a control variable, parental warmth at age 12 positively predicted

superiority at age 14 (e.g., β = .15, p = .019 for BARS maternal warmth in the spouse-report
data).

Parental monitoring—Our prediction (H2) that parental monitoring would be associated

with overall narcissism levels was not confirmed. However, when narcissism was
differentiated into the two facets exploitativeness and superiority, we found that both
maternal and paternal monitoring at age 12 were negatively related to exploitativeness at age
14. For example, the standardized regression coefficient for the child-report of parental
monitoring was −.16 for paternal monitoring and −.14 for maternal monitoring. In the
models without self-esteem as a control variable, the effect was slightly smaller for maternal
monitoring (β = −.10) and did not reach significance (p = .072). This result also held for
Author Manuscript

parent-report data (e.g., β = −.10, p = .049 for mother-report of paternal monitoring),

although not all effects were large enough to reach significance (e.g., β = −.07, p = .28 for
father-report of maternal monitoring). For the data from observational assessments this
effect was confirmed for paternal monitoring, but not maternal monitoring (see Table 5). The
superiority facet was not related to parental monitoring in models including self-esteem. In
models without self-esteem, a positive association between superiority and parental
monitoring was found when monitoring was assessed via parent and child reports (e.g., β = .
15, p = .006 for mother-report of paternal monitoring), but not using the observational data.
Thus, it appears that adolescents who were monitored less closely at age 12 tended to have
higher exploitativeness levels over time compared to adolescents who were monitored more
closely at age 12.
Author Manuscript

Supplemental Analyses
In supplemental analyses, we examined reciprocal cross-lagged relations between the
narcissism facets and self-esteem. Higher self-esteem at age 12 was related to higher
superiority at age 14 (β = .43, p < .001), whereas superiority at age 12 was not related to
self-esteem at age 14. Exploitativeness and self-esteem were not related. These effects
consistently held when each of the parenting dimensions was added to the model. We also
tested whether the parenting dimensions had interactive effects on narcissism. We found that

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 11

the association between paternal hostility and exploitativeness was stronger for higher levels
Author Manuscript

of monitoring (interaction effect b = .17, p<.001) and for higher levels of warmth
(interaction effect b = .14, p < .001). In addition, we found that the association between
paternal hostility and superiority was stronger for higher levels of monitoring, (interaction
effect, b = .10, p = .041). Thus, it appears that monitoring tends to magnify the effects of
parenting on narcissism, at least for fathers.

Lastly, we tested whether nativity (born in Mexico vs. the U.S.) moderated the relations
between parenting and narcissism. We found two significant effects: nativity moderated the
association between observed maternal hostility and superiority and between observed
maternal warmth and exploitativeness. However, we do not believe these interaction effects
merit interpretation because: (a) we conducted a total of 72 interactions tests and two
significant effects is less than would be expected by chance; (b) the interaction effects held
only for observational assessments of parenting and did not replicate for either child or
Author Manuscript

parent reports, and (c) the effects did not replicate for any measure of paternal hostility and
warmth. Overall, then, the results suggest that our basic findings hold for youth who were
born in Mexico as well as those born in the United States. Moreover, all of the significant
findings remained significant after controlling for nativity; that is, after nativity was entered
as a covariate in the models.

Discussion
The present study investigated the association between parenting practices and the
development of narcissism using data from a large longitudinal study of Mexican-origin
children and their parents. The two most robust predictors were parental hostility and
parental monitoring, with hostility associated with higher exploitativeness from age 12 to 14
Author Manuscript

and parental monitoring associated with lower exploitativeness from age 12 to 14. These
effects replicated across three different parenting measures (child reports, spouse reports,
and behavioral coding of parent-child interactions), and held for youth born in Mexico and
the United States. Surprisingly, none of the parenting dimensions was related to superiority,
suggesting that parenting practices are more strongly related to the maladaptive than the
adaptive component of narcissism. Below we discuss the implications of the findings for
theory, research, and practice.

The tendency for youth raised by hostile parents to develop higher levels of exploitativeness
is consistent with psychodynamic theory. Specifically, psychodynamic theory posits that
children exposed to hostile and overly critical parents will develop feelings of inadequacy
that they will try to compensate for by exploiting and seeking admiration from others
(Kernberg, 1975; Kohut, 1977). However, psychodynamic theory also posits that a lack of
Author Manuscript

parental warmth plays an important role in the development of narcissism. This prediction
was not confirmed in our study since we found no relationship between parental warmth and
narcissism, neither at the level of overall narcissism nor at the level of the exploitativeness
and superiority facets. Interestingly, when we explored interactions among the parenting
dimensions, we found that the effect of paternal hostility on exploitativeness was stronger
for higher levels of warmth, as well as for higher levels of monitoring. Especially the
interaction between parental hostility and parental warmth is consistent with the idea

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 12

grounded in psychodynamic theory that parental behavior characterized by first elevating

Author Manuscript

children with love and support and then tearing them down with excessive criticism fosters
the development of narcissism. Although these results should be interpreted with caution,
since neither effect replicated for parent-report or observational assessments of parenting,
they do suggest that the association between parenting practices and narcissism is likely to
be more complex than what can be captured by unidimensional measures of parenting.
Finally, it is important to note that most theories of the origins of narcissism do not specify
distinct developmental pathways to the specific facets of narcissism. The moderate stability
coefficients at the level of overall narcissism and at the facet level indicate that narcissism
and in particular the exploitativeness facet have not fully stabilized by adolescence. It is
mainly the maladaptive components of narcissism such as exploitativeness that are related to
a number of adjustment problems in adolescence (e.g., Washburn et al., 2004). Thus, more
conceptual work is needed to extend current theories to the development of the maladaptive
Author Manuscript

and adaptive components of narcissism, taking into account different developmental periods
from childhood to adolescence.

Most of our findings are consistent with prior research. For example, our finding that overall
narcissism was not related to parental warmth is consistent with Brummelman et al. (2015).
In some of our models without self-esteem as a control variable, parental warmth was related
to the adaptive superiority facet. This is in line with Cramer (2011) who found that parenting
styles characterized by high responsiveness (an aspect of warmth) were related to higher
levels of adaptive narcissistic traits, although we did not find this relationship when we
controlled for self-esteem. We found that high parental monitoring was related to lower
levels of exploitativeness, but not to higher levels of superiority. This is consistent with
cross-sectional research showing a negative relationship between parental monitoring and
overall narcissism (Barry et al., 2007; Horton et al., 2006; Miller & Campbell, 2008). The
Author Manuscript

convergence of our results with previous research suggests that the relationship between
parenting behaviors and narcissism is similar across different cultural contexts, such as
Mexican-origin and European background youth living in the United States and Dutch
youth. This is substantiated by our finding that nativity did not influence the relations
between parenting and narcissism. Nevertheless, the degree to which the cultural context
plays a role in the developmental processes involved in narcissism is an empirical question
that merits further research.

Controlling for self-esteem was an important extension of the analyses conducted in

previous studies. Superiority and self-esteem were positively related, with self-esteem at age
12 predicting superiority at age 14. This overlap between superiority and self-esteem
produced differences between the effects found in models that controlled vs. did not control
Author Manuscript

for self-esteem. For example, in models investigating the narcissism facets and parental
monitoring without controlling for self-esteem, parental monitoring at age 12 predicted
superiority at age 14. When self-esteem was added to the model, this relationship was
diminished greatly and no longer statistically significant, illustrating that the initial effect
was only due to the confounding of superiority and self-esteem. The same pattern occurred
in some models examining the association between parental warmth and superiority. Thus,
relationships between positive parenting dimensions and superiority appear to be attributable
to the shared variance between superiority and self-esteem; when the variance associated

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 13

with self-esteem is statistically removed, the resulting more pure measure of narcissistic
Author Manuscript

self-aggrandizement is not related to the positive parenting dimensions of warmth and

monitoring. This finding highlights the importance of controlling for self-esteem when
examining the effects of narcissism, and vice versa. Note that as exploitativeness was not
correlated with self-esteem, all observed associations between parenting and exploitativeness
were robust across models that included or did not include self-esteem as a control.

The present study also extended previous research by investigating the codevelopment of the
two facets of narcissism. With regard to general patterns of change, both facets showed
virtually no mean-level change from age 14 to 16 and moderately high rank-order stability,
with superiority showing somewhat higher stability than exploitativeness. Interestingly,
when we examined prospective cross-lagged associations between the two facets, we found
that superiority at age 14 predicted exploitativeness at age 16, but exploitativeness at age 14
did not predict superiority at age 16. It is possible that a certain degree of superiority is
Author Manuscript

needed in order to exhibit exploitative behaviors; that is, a sense of superiority can engender
feelings of entitlement and help rationalize exploiting others. Thus, superiority may be a
developmental precursor of exploitativeness, with both facets merging into a coherent
narcissistic style later in development. In line with this possibility, the concurrent correlation
between the two facets of narcissism was significantly higher at age 16 (latent r = .55)
compared to age 14 (latent r = .38), implying that superiority and exploitativeness levels
were more aligned at age 16 than at age 14. In general, the divergent pattern of findings for
superiority and exploitativeness, and the fact that superiority is prospectively associated with
exploitativeness, illustrate the importance of investigating narcissism at the facet level.

The present findings have important practical implications. In particular, they highlight two
promising targets of intervention: parental hostility and monitoring. Specifically, parenting
Author Manuscript

interventions that reduce hostility and improve monitoring could potentially disrupt the
developmental pathways that lead to exploitativeness, a maladaptive component of
narcissism that has been linked to adjustment and behavioral problems in childhood and
adolescence (Barry et al., 2007; Barry et al., 2003; Washburn et al., 2004).

Limitations of the Study

Several limitations merit attention. First, narcissism was assessed on two occasions during
adolescence. Both narcissism facets showed moderate rank-order stability during this time
period, raising the possibility that some of the processes that contribute to the development
of narcissism have already unfolded by adolescence. Stronger links to parenting practices
could be expected earlier in development when narcissism levels are presumably undergoing
greater change. Nevertheless, our results indicate that the socialization processes involved in
Author Manuscript

the development of narcissism continue into adolescence. That is, an adolescent

experiencing hostile feedback from his/her parents is likely to exhibit the same feelings of
inadequacy, and compensate for these feelings by inflating his/her self-worth, in the same
way as a child receiving such feedback, and consequently the socialization processes
through which hostile parenting influences narcissism are likely to persist across
adolescence. Similarly, from a social learning perspective, it seems likely that the relatively
straightforward process of learning about oneself via feedback from parents is likely to

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 14

persist across adolescence. It would be important to investigate the effects of parenting on

Author Manuscript

narcissism in a longitudinal study where both constructs are assessed at younger ages,
preferably using informant assessments of narcissism given the problems with self-reports
by young children. Even if parent socialization processes persist into adolescence, as we
believe they do, they may be more consequential earlier in development; that is, there may
be a sensitive period in the development of narcissism.

Second, the study design does not allow for strong conclusions regarding the causal effect of
parenting on narcissism. As in all passive observational designs, effects between factors may
be caused by third variables that were not assessed (Finkel, 1995). Nevertheless, longitudinal
analyses are useful because they can indicate whether the data are consistent with a causal
model of the relation between the variables.

Third, the results do not allow for firm conclusions with regard to the clinical category of
Author Manuscript

narcissistic personality disorder (NPD). The narcissism measure used in the present research
was designed to assess individual differences in the normal range of narcissistic tendencies,
but conclusions about the antecedents of NPD should be based on diagnoses from clinical
interviews. Moreover, our analyses are based on a nonclinical sample, which do not allow
for valid conclusions about narcissistic tendencies in clinical populations.

Conclusion
Considering the detrimental effects of narcissism in the interpersonal domain, investigating
the development of narcissism is an important endeavor. Our study showed that parental
hostility and parental monitoring were related to adolescents’ exploitativeness two years
later, indicating that parenting practices play a central role in the development of narcissism
during adolescence. Nevertheless, there are many open questions, including identifying the
Author Manuscript

developmental precursors of the superiority facet of narcissism and understanding how

parenting practices shape the development of narcissism in younger children.

Acknowledgments
This research was supported by a post-doc fellowship awarded to Eunike Wetzel by the German Academic
Exchange Service (DAAD) and a grant from the National Institute on Drug Abuse and the National Institute on
Alcohol Abuse and Alcoholism (DA017902) to Richard W. Robins and Rand D. Conger. We thank the participating
families, staff, and research assistants who took part in this study.

References
Ang RP, Raine A. Reliability, validity and invariance of the Narcissistic Personality Questionnaire for
Children-Revised (NPQC-R). Journal of Psychopathology and Behavioral Assessment. 2008;
31:143–151.
Author Manuscript

Barber BK, Olsen JE, Shagle SC. Associations between parental psychological and behavioral control
and youth internalized and externalized behaviors. Child Development. 1994; 65(4):1120–1136.
[PubMed: 7956469]
Barry CT, Frick PJ, Adler KK, Grafeman SJ. The Predictive Utility of Narcissism among Children and
Adolescents: Evidence for a Distinction between Adaptive and Maladaptive Narcissism. Journal of
Child and Family Studies. 2007; 16(4):508–521.

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 15

Barry CT, Frick PJ, Killian AL. The relation of narcissism and self-esteem to conduct problems in
children: a preliminary investigation. Journal of Clinical Child and Adolescent Psychology. 2003;
Author Manuscript

32(1):139–152. [PubMed: 12573939]

Bentler PM. Comparative fit indexes in structural models. Psychological Bulletin. 1990; 107(2):238–
246. [PubMed: 2320703]
Bornstein, MH. Parenting science and practice. In: Renninger, KA., editor. Handbook of child
psychology. Vol. 4. Hoboken, NJ: Wiley; 2006. p. 893-949.
Brown RP, Zeigler-Hill V. Narcissism and the non-equivalence of self-esteem measures: A matter of
dominance? Journal of Research in Personality. 2004; 38(6):585–592.
Bushman BJ, Baumeister RF. Threatened egotism, narcissism, self-esteem, and direct and displaced
aggression: Does self-love or self-hate lead to violence? Journal of Personality and Social
Psychology. 1998; 75(1):219–229. [PubMed: 9686460]
Campbell WK, Campbell SM. On the self-regulatory dynamics created by the peculiar benefits and
costs of narcissism: A contextual reinforcement model and examination of leadership. Self and
Identity. 2009; 8(2–3):214–232.
Campbell WK, Foster CA, Finkel EJ. Does self-love lead to love for others? A story of narcissistic
Author Manuscript

game playing. Journal of Personality and Social Psychology. 2002; 83(2):340–354. [PubMed:
12150232]
Conger, RD. Behavioral Affect Rating Scale (BARS). Ames, IA: Iowa State University; 1989a.
Conger, RD. Iowa Parenting Scale. Ames, IA: Iowa State University; 1989b.
Cramer P. Young adult narcissism: A 20 year longitudinal study of the contribution of parenting styles,
preschool precursors of narcissism, and denial. Journal of Research in Personality. 2011; 45(1):19–
28.
Dishion TJ, McMahon RJ. Parental monitoring and the prevention of child and adolescent problem
behavior: A conceptual and empirical formulation. Clinical Child and Family Psychological
Review. 1998; 1(1):61–75.
Finkel, SE. Causal analysis with panel data. Thousand Oaks, CA: Sage; 1995.
Freud, S. On narcissism: An introduction. In: Strachey, J., editor. Standard edition. Vol. 14. London,
England: Hogarth Press; 1914.
Horton, RS. Parenting as a cause of narcissism: Empirical support for psychodynamic and social
Author Manuscript

learning theories. In: Campbell, WK.; Miller, JD., editors. The handbook of narcissism and
narcissistic personality disorder. Hoboken, NJ: John Wiley & Sons; 2011. p. 181-190.
Horton RS, Bleau G, Drwecki B. Parenting narcissus: What are the links between parenting and
narcissism? Journal of Personality. 2006; 74(2):345–376. [PubMed: 16529580]
Hu LT, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: Conventional
criteria versus new alternatives. Structural Equation Modeling-a Multidisciplinary Journal. 1999;
6(1):1–55.
Kernberg, OF. Borderline conditions and pathological narcissism. New York: Jason Aronson; 1975.
Kohut, H. The analysis of the self: A systematic approach to the psychoanalytic treatment of narcisstic
personality disorders. New York: International Universities Press; 1971.
Kohut, H. The restoration of the self. New York: International Universities Press; 1977.
Meade AW, Johnson EC, Braddy PW. Power and sensitivity of alternative fit indices in tests of
measurement invariance. Journal of Applied Psychology. 2008; 93(3):568–592. [PubMed:
18457487]
Author Manuscript

Melby J, Conger RD, Book R, Rueter M, Lucy L, Repinski D, Scaramella L. The Iowa Family
Interaction Rating Scales. 1998 Retrieved from http://www.scribd.com/doc/66440091/The-Iowa-
Family-Interaction-Rating-Scales.
Meredith W. Measurement invariance, factor-analysis and factorial invariance. Psychometrika. 1993;
58(4):525–543.
Miller JD, Campbell WK. Comparing clinical and social-personality conceptualizations of narcissism.
Journal of Personality. 2008; 76(3):449–476. [PubMed: 18399956]

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 16

Miller JD, Campbell WK, Young DL, Lakey CE, Reidy DE, Zeichner A, Goodie AS. Examining the
relations among narcissism, impulsivity, and self-defeating behaviors. Journal of Personality. 2009;
Author Manuscript

77(3):761–794. [PubMed: 20078737]

Millon, T. Modern psychopathology: A biosocial approach to maladaptive learning and functioning.
Philadelphia, PA: Saunders; 1969.
Millon, T. Disorders of personality. New York: Wiley; 1981.
Muthén, LK.; Muthén, BO. Los Angeles, CA: Muthén & Muthén; 1998–2014. Mplus [Computer
software]. Retrieved from http://www.statmodel.com
Otway LJ, Vignoles VL. Narcissism and childhood recollections: A quantitative test of psychoanalytic
predictions. Personality and Social Psychology Bulletin. 2006; 32(1):104–116. [PubMed:
16317192]
Parke, RD.; Buriel, R. Socialization in the family: Ethnic and ecological perspectives. In: Eisenberg,
N., editor. Handbook of Child Psychology. Vol. 3. Hoboken, NJ: Wiley; 2006. p. 429-504.
Paulhus DL, Robins RW, Trzesniewski KH, Tracy JL. Two replicable suppressor situations in
personality research. Multivariate Behavioral Research. 2004; 39(2):303–328. [PubMed:
26804578]
Author Manuscript

Pettit GS, Laird RD, Dodge KA, Bates JE, Criss MM. Antecedents and behavior-problem outcomes of
parental monitoring and psychological control in early adolescence. Child Development. 2001;
72(2):583–598. [PubMed: 11333086]
Rhodewalt F, Morf CC. Self and interpersonal correlates of the Narcissistic Personality Inventory - A
review and new findings. Journal of Research in Personality. 1995; 29(1):1–23.
Rosenberg, M. Society and the adolescent self-image. Princeton, NJ: Princeton Universiy Press; 1965.
Rosenberg, M. Conceiving the self. New York: Basic Books; 1979.
Sedikides C, Rudich EA, Gregg AP, Kumashiro M, Rusbult C. Are normal narcissists psychologically
healthy?: Self-esteem matters. Journal of Personality and Social Psychology. 2004; 87(3):400–416.
[PubMed: 15382988]
Small SA, Kerns D. Unwanted sexual activity among peers during early and middle adolescence -
Incidence and risk factors. Journal of Marriage and the Family. 1993; 55(4):941–952.
Steiger JH. Structural model evaluation and modification - an interval estimation approach.
Multivariate Behavioral Research. 1990; 25(2):173–180. [PubMed: 26794479]
Author Manuscript

Tracy JL, Cheng JT, Robins RW, Trzesniewski KH. Authentic and hubristic pride: The affective core
of self-esteem and narcissism. Self and Identity. 2009; 8(2–3):196–213.
Trzesniewski KH, Donnellan MB, Robins RW. Is "Generation Me" Really More Narcissistic Than
Previous Generations? Journal of Personality. 2008; 76(4):903–918.
Tucker LR, Lewis C. Reliability coefficient for maximum likelihood factor-analysis. Psychometrika.
1973; 38(1):1–10.
Washburn JJ, McMahon SD, King CA, Reinecke MA, Silver C. Narcissistic features in young
adolescents: Relations to aggression and internalizing symptoms. Journal of Youth and
Adolescence. 2004; 33(3):247–260.
Widaman KF, Reise SP. Exploring the measurement invariance of psychological instruments:
Applications in the substance use domain. The science of prevention: Methodological advances
from alcohol and substance abuse research. 1997:281–324.
Author Manuscript

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 17

Highlights
Author Manuscript

- We investigated the longitudinal relations between parenting and

narcissism.

- Data from a sample of Mexican-origin youth spanning ages 12 to 16 were

analyzed.

- Parental hostility was associated with higher levels of exploitativeness at

age 14.

- Parental monitoring was associated with lower levels of exploitativeness at

age 14.

- None of the parenting dimensions was related to superiority.

Author Manuscript

- Findings replicated across child-report, spouse-report, and behavioral

codings
Author Manuscript
Author Manuscript

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Wetzel and Robins Page 18
Author Manuscript
Author Manuscript

Figure 1.
Latent cross-lagged panel model with narcissism and a parenting scale (parental hostility,
parental warmth, or parental monitoring). In the facet-level analyses, the overall narcissism
construct was replaced by exploitativeness and superiority.
Author Manuscript
Author Manuscript

J Res Pers. Author manuscript; available in PMC 2017 August 01.

 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 1

Descriptive Statistics for all Study Variables

Age 12 Age 14 Age 16

Construct Report M (SD) α M (SD) α M (SD) α

Narcissism
Wetzel and Robins

Total Child - - 2.88 (.57) .78 2.87 (.62) .83

Exploitativeness Child - - 2.42 (.72) .74 2.42 (.76) .78
Superiority Child - - 3.36 (.71) .77 3.33 (.71) .80
Hostility
BARS maternal Child 1.41 (.31) .77 1.51 (.41) .86 1.53 (.39) .86
BARS maternal Parent 1.54 (.31) .75 1.54 (.29) .73 1.55 (.29) .71
Maternal Observational 2.18 (1.47)
BARS paternal Child 1.35 (.32) .81 1.53 (.35) .80 1.56 (.39) .83
BARS paternal Parent 1.53 (.33) .77 1.64 (.32) .81 1.59 (.30) .80
Paternal Observational 1.75 (1.21)
Warmth
BARS maternal Child 3.12 (.59) .88 3.02 (.65) .91 2.89 (.66) .91
BARS maternal Parent 3.36 (.48) .85 3.27 (.50) .86 3.17 (.55) .87
IPS maternal Child 2.70 (.63) .87 2.56 (.62) .86 2.48 (.62) .88
IPS maternal Parent 3.03 (.51) .80 3.00 (.54) .84 2.90 (.57) .85
Observational maternal observational 3.58 (1.45)
BARS paternal child 3.01 (.72) .92 2.77 (.74) .91 2.61 (.75) .92

J Res Pers. Author manuscript; available in PMC 2017 August 01.

BARS paternal parent 3.10 (.66) .91 2.97 (.68) .90 2.88 (.68) .91
IPS paternal child 2.60 (.68) .89 2.39 (.66) .88 2.26 (.66) .89
IPS paternal parent 2.71 (.66) .88 2.65 (.68) .88 2.60 (.66) .88
Observational paternal observational 3.25 (1.27)
Monitoring
PMC maternal child 3.32 (.57) .91 3.22 (.60) .92 3.11 (.62) .92
PMC maternal father 3.68 (.43) .87 3.52 (.50) .89 3.38 (.60) .92
PMC maternal mother 3.70 (.43) .85 3.68 (.46) .87 3.58 (.53) .91
Observational maternal observational 5.20 (1.14)
Page 19
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Age 12 Age 14 Age 16

Construct Report M (SD) α M (SD) α M (SD) α
PMC paternal child 3.09 (.75) .94 2.85 (.78) .95 2.73 (.81) .95
PMC paternal father 3.51 (.49) .85 3.46 (.54) .90 3.35 (.57) .91
PMC paternal mother 3.34 (.74) .92 3.23 (.81) .94 3.17 (.78) .94
Wetzel and Robins

Observational paternal observational 4.77 (1.10)

Self-esteem child 3.19 (.42) .78 3.13 (.42) .85 3.11 (.43) .85

Note. BARS = Behavioral Affect Rating Scale, IPS = Iowa Parenting Scale, PMC = Parental Monitoring of Child.

J Res Pers. Author manuscript; available in PMC 2017 August 01.

Page 20
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 2

Fit Indices for Latent Cross-lagged Panel Models

Fit indices

Model Re- χ2 Df CFI TLI RMSEA 90% CI

port
Wetzel and Robins

Overallnarcissism
Hostility
Maternal child 335.36 246 0.99 0.98 .02 [.02, .03]
Maternal parent 322.24 244 0.99 0.98 .02 [.02, .03]
Maternal obs 176.89 104 0.98 0.98 .03 [.03, .04]
Paternal child 295.78 246 0.99 0.99 .02 [.01, .03]
Paternal parent 347.66 245 0.98 0.98 .03 [.02, .03]
Paternal obs 168.52 104 0.98 0.98 .03 [.02, .04]
Warmth
BARS maternal child 375.19 246 0.98 0.98 .03 [.02, .03]
BARS maternal parent 332.63 245 0.99 0.98 .02 [.02, .03]
IPS maternal child 326.44 245 0.99 0.99 .02 [.02, .03]
IPS maternal parent 305.75 245 0.99 0.99 .02 [.01, .03]
maternal obs 168.65 104 0.98 0.98 .03 [.02, .04]
BARS paternal child 339.13 246 0.99 0.99 .02 [.02, .03]
BARS paternal parent 306.29 245 0.99 0.99 .02 [.01, .03]
IPS paternal child 314.71 244 0.99 0.99 .02 [.01, .03]

J Res Pers. Author manuscript; available in PMC 2017 August 01.

IPS paternal parent 348.10 245 0.99 0.98 .03 [.02, .03]
paternal obs 162.57 104 0.99 0.98 .03 [.02, .04]
Monitoring
Maternal child 374.28 246 0.98 0.98 0.03 [.02, .03]
Maternal mother 339.81 244 0.99 0.98 0.03 [.02, .03]
Maternal father 321.51 246 0.99 0.99 0.02 [.02, .03]
Maternal obs 166.33 104 0.98 0.98 0.03 [.02, .04]
Paternal child 360.52 244 0.99 0.98 0.03 [.02, .03]
Paternal moth 345.46 246 0.99 0.99 0.03 [.02, .03]
Page 21
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Fit indices

Model Re- χ2 Df CFI TLI RMSEA 90% CI

port
Paternal father 329.51 245 0.99 0.98 0.02 [.02, .03]
Paternal obs 176.30 104 0.98 0.98 0.03 [.02, .04]
Wetzel and Robins

Facets
Hostility
Maternal child 620.52 390 0.97 0.97 .03 [.03, .04]
Maternal parent 617.56 388 0.97 0.96 .03 [.03, .04]
Maternal obs 409.84 204 0.96 0.95 .04 [.03, .05]
Paternal child 598.75 390 0.97 0.97 .03 [.02, .03]
Paternal parent 626.09 389 0.97 0.96 .03 [.03, .04]
Paternal obs 401.76 204 0.96 0.95 .04 [.03, .05]
Warmth
BARS maternal child 660.50 390 0.97 0.97 .03 [.03, .04]
BARS maternal parent 604.96 389 0.97 0.96 .03 [.03, .03]
IPS maternal child 611.22 389 0.97 0.97 .03 [.03, .03]
IPS maternal parent 584.96 389 0.97 0.97 .03 [.02, .03]
Maternal obs 399.92 204 0.96 0.95 .04 [.03, .04]
BARS paternal child 624.78 390 0.98 0.97 .03 [.03, .04]
BARS paternal parent 590.63 389 0.98 0.97 .03 [.02, .03]
IPS paternal child 608.68 388 0.98 0.97 .03 [.03, .03]
IPS paternal parent 643.11 389 0.97 0.97 .03 [.03, .04]

J Res Pers. Author manuscript; available in PMC 2017 August 01.

Paternal obs 398.17 204 0.96 0.95 .04 [.03, .04]
Monitoring
Maternal child 650.38 390 0.97 0.97 0.03 [.03, .04]
Maternal mother 641.22 388 0.97 0.97 0.03 [.03, .04]
Maternal father 609.93 390 0.97 0.97 0.03 [.03, .03]
Maternal obs 400.90 204 0.96 0.95 0.04 [.03, .05]
Paternal child 624.79 388 0.98 0.97 0.03 [.03, .04]
Paternal mother 614.67 390 0.98 0.97 0.03 [.03, .04]
Paternal father 613.88 389 0.97 0.97 0.03 [.03, .04]
Page 22
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Fit indices

Model Re- χ2 Df CFI TLI RMSEA 90% CI

port
Paternal obs 415.46 204 0.96 0.95 0.04 [.04, .05]

Note. BARS = Behavioral Affect Rating Scale, IPS = Iowa Parenting Scale, obs = observerational, CFI = comparative fit index, TLI = Tucker-Lewis index, RMSEA = root mean square error of
Wetzel and Robins

approximation. Differing numbers of degrees of freedom for models with same data structure are due to parameters freed to account for partial scalar invariance in some constructs. Models in the upper half
of the table contain overall narcissism, self-esteem and one parenting scale. Models in the lower half of the table contain exploitativeness, superiority, self-esteem, and one parenting scale.

J Res Pers. Author manuscript; available in PMC 2017 August 01.

Page 23
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 3

Parameter Estimates for Latent Cross-lagged Panel Models Involving Parental Hostility

Model Report Standardized regression coefficients

Host12 → Host14 → Narc14 →
Overall narcissism Narc14 Narc16 Host16
Wetzel and Robins

Maternal hostility child .24 .02 .01

Maternal hostility parent .13 .02 .08
Maternal hostility obs .12 - -
Paternal hostility child .14 −.02 −.02
Paternal hostility parent .14 −.06 .04
Paternal hostility obs .07 - -

Host12 → Host14 → Exp14 → Host12 → Host14 → Sup14 →

Narcissism facets Exp14 Exp16 Host16 Sup14 Sup16 Host16

Maternal hostility child .29 .02 .04 .08 .05 −.03

Maternal hostility parent .18 .03 .01 .04 .02 .11
Maternal hostility obs .17 - - .03 - -
Paternal hostility child .21 .01 .04 .02 −.01 −.08
Paternal hostility parent .19 −.07 .09 .04 −.01 −.06
Paternal hostility obs .10 - - .02 - -

Note. Host = hostility, narc = narcissism, exp = exploitativeness, obs = observational, sup = superiority. Only standardized regression coefficients from cross-lagged paths are depicted. For stability estimates
see text. Beta coefficients significant at α = .05 are depicted in bold and italics. All models include self-esteem and gender as covariates.

J Res Pers. Author manuscript; available in PMC 2017 August 01.

Page 24
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 4

Parameter Estimates for Latent Cross-lagged Panel Models Involving Parental Warmth

Model Report Standardized regression coefficients

Warm12 → Warm14 → Narc14 →
Overall narcissism Narc14 Narc16 Warm16
Wetzel and Robins

BARS maternal warmth child .03 .01 .02

BARS maternal warmth parent .08 .03 .05
IPS maternal warmth child .06 .07 .02
IPS maternal warmth parent .06 .03 .09
Maternal warmth obs .01 - -
BARS paternal warmth child .02 .01 .06
BARS paternal warmth parent −.04 .03 −.07
IPS paternal warmth child .09 .05 .05
IPS paternal warmth parent −.06 .04 −.07
Paternal warmth obs −.00 - -

Warm12 → Warm14 → Exp14 → Warm12 → Warm14 → Sup14 →

Narcissism facets Exp14 Exp16 Warm16 Sup14 Sup16 Warm16

BARS maternal warmth child −.05 −.04 −.04 .09 .01 .08
BARS maternal warmth parent .03 .07 −.06 .08 −.01 .16
IPS maternal warmth child −.05 .03 −.02 .14 .07 .04
IPS maternal warmth parent .03 .04 −.04 .07 −.00 .17
Maternal warmth obs .01 - - .02 - -

J Res Pers. Author manuscript; available in PMC 2017 August 01.

BARS paternal warmth child −.08 −.04 −.05 .12 −.00 .14
BARS paternal warmth parent −.12 .02 −.09 .06 −.01 .01
IPS paternal warmth child −.04 −.01 −.04 .19 .06 .12
IPS paternal warmth parent −.11 .02 −.12 .01 .03 .06
Paternal warmth obs −.02 - - .01 - -

Note. Warm = warmth, narc = narcissism, exp = exploitativeness, obs = observational, sup = superiority. Only standardized regression coefficients from cross-lagged paths are depicted. For stability
estimates see text. Beta coefficients significant at α = .05 are depicted in bold and italics. All models include self-esteem and gender as covariates.
Page 25
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 5

Parameter Estimates for Latent Cross-lagged Panel Models Involving Parental Monitoring

Model Report Standardized regression coefficients

Overall narcissism Mon12 → Mon14 → Narc14 →

Narc14 Narc16 Mon16
Wetzel and Robins

Maternal monitoring child −.03 .00 −.03

Maternal monitoring moth .00 .05 −.01
Maternal monitoring father −.02 .02 −.00
Maternal monitoring obs −.01 - -
Paternal monitoring child −.04 .02 .04
Paternal monitoring moth −.02 .03 −.09
Paternal monitoring father −.01 .04 .03
Paternal monitoring obs −.03 - -

Mon12 → Mon14 → Exp14 → Mon12 → Mon14 → Sup14 →

Narcissism facets Exp14 Exp16 Mon16 Sup14 Sup16 Mon16

Maternal monitoring child −.14 −.08 −.05 .08 .02 .02

Maternal monitoring moth −.04 .03 −.10 .04 .04 .12
Maternal monitoring father −.07 −.00 −.08 .04 .02 .09
Maternal monitoring obs .02 - - −.03 - -
Paternal monitoring child −.16 −.04 −.06 .09 .05 .13
Paternal monitoring moth −.10 .03 −.09 .07 .00 −.00
Paternal monitoring father −.06 .05 −.10 .06 .00 .16

J Res Pers. Author manuscript; available in PMC 2017 August 01.

Paternal monitoring obs −.08 - - .03 - -

Note. Mon = monitoring, narc = narcissism, exp = exploitativeness, obs = observational, sup = superiority, moth = mother. Only standardized regression coefficients from cross-lagged paths are depicted.
For stability estimates see text. Beta coefficients significant at α = .05 are depicted in bold and italics. All models include self-esteem and gender as covariates.
Page 26