The Family Environment in Late Childhood and Early Adolescence as a Predictor of Narcissism in Emerging Adulthood: Findings from a 17-year Longitudinal Study of Mexican-Origin Youth

Abstract

The development of narcissism has been theoretically tied to childhood experiences, yet little empirical research has examined the influence of the family environment on adult narcissism using longterm longitudinal data. The present study identified family factors in late childhood and early adolescence that are prospectively associated with narcissism (grandiosity, superiority, exploitativeness, vulnerability) in emerging adulthood using data from a longitudinal study of Mexican-origin youth (N=674) spanning from age 10 to 26. Using structural equation modeling of multimethod data (child report, parent & spouse report, and coded parent-child interactions), we found that grandiosity and superiority were significantly predicted by positive aspects of the family environment (e.g., parental warmth, parent-child relationship quality, parental relationship quality), whereas exploitativeness and vulnerability were significantly predicted by negative aspects of the family environment (e.g., parental hostility, parent-child conflict). These findings were much weaker, and often non-significant, after controlling for earlier levels of narcissism, suggesting that the family environment exerts its effect on adult narcissism via its concurrent effect on adolescent narcissism. Overall, our study highlights several underexplored family factors in childhood that contribute to the development and maintenance of narcissism in adulthood.

Introduction

Narcissism, a construct that includes characteristics like self-entitled vanity, superiority, and hypersensitivity, is associated with multiple aspects of psychological functioning and interpersonal outcomes (Brunell et al., 2008; Czarna et al., 2021; Kjærvik & Bushman, 2021; Rosenthal & Pittinsky, 2006; van Geel et al., 2017; Wetzel et al., 2021). Although there is growing evidence of the consequences of narcissism, we know little about the developmental antecedents of narcissism.

The present study addresses an important and largely unanswered question: Are children who grow up in a dysfunctional family environment more likely to become narcissistic in adulthood? It is especially important to identify the childhood factors that contribute to the development of adult narcissism given that it tends to remain relatively stable over the lifespan (Chopik & Grimm, 2019; Ma, Beck, & Robins, 2025; Orth & Luciano, 2015; Wetzel et al., 2020). The present study is the first to prospectively investigate numerous factors of the family environment across late childhood and adolescence (ages 10 to 16) as predictors of emerging adult narcissism (ages 19 to 26) using a long-term longitudinal study of Mexican-origin youth, an understudied and ethnically minoritized community. Understanding the precursors of narcissistic tendencies during this period of development can offer clarity into the processes that underlie narcissism and help us pinpoint why some individuals grow up to be arrogant, entitled, and emotionally vulnerable adults, whereas others grow up to be humble, compassionate, and emotionally secure.

Developmental Antecedents of Narcissism

The developmental roots of narcissism—composed of the grandiose (superior, manipulative, and attention-seeking) and the vulnerable (insecure, shameful, and hypersensitive) self—are believed to stem from familial influences in childhood. Psychodynamic perspectives theorized that parenting behaviors, such as hostility, coldness, and excessive criticism, can engender feelings of inadequacy that fuel the compensatory need for praise and admiration from others (Kernberg, 1975; Kohut, 1977).

In contrast, Millon’s (1981) social learning perspective hypothesized that parental permissiveness can lead children to develop feelings of entitlement. In other words, narcissistic tendencies arise from observing and internalizing parents’ adulation toward oneself and expecting others to conform to this behavior in the future. Cross-sectional research supports both perspectives, with associations between narcissism and hostile or invalidating parenting (Huxley et al., 2021) and overindulgent or lenient parenting (Horton et al., 2006; van Schie et al., 2020; Watson et al., 1992), suggesting that the development of narcissism may reflect an equifinal principle (multiple divergent trajectories converging on the same outcome; Cicchetti & Rogosch, 1996). Research has also linked narcissism to lower levels of parental monitoring (Barry et al., 2007; Horton et al., 2006; Mechanic & Barry, 2015; Miller & Campbell, 2008) and higher levels of parental psychological control (Miller & Campbell, 2008), and warmth (in connection with adaptive narcissism; Horton et al., 2006).

There are at least two gaps in our understanding of the developmental antecedents of adult narcissism. First, identifying developmental antecedents is not possible without longitudinal work, which remains scarce in narcissism research. Cramer (2011) found that both authoritative and indulgent parenting positively predicted adaptive forms of narcissism, whereas authoritarian parenting negatively predicted adaptive traits. Although this study provided the first prospective support for both psychodynamic and social learning theories, it sampled only 102 individuals across two waves from age 3 to age 23, used a broad-based personality measure that was not designed to assess narcissism, and relied solely on self-report data for parenting behaviors. Brummelman et al. (2015) found evidence using parent- and child-reports of parenting that parental indulgence (but not lack of warmth) prospectively predicted narcissism; however, lack of parental warmth is conceptually different from parental hostility and coldness. Moreover, this study utilized a global measure of narcissism, leaving unanswered questions about whether the grandiosity and vulnerability facets of narcissism are differentially impacted by parenting. Another longitudinal study found that parental hostility led to increased levels of exploitativeness whereas parental monitoring led to decreased levels of exploitativeness (Wetzel & Robins, 2016), but only examined prospective effects of parenting from age 12 to 14 and from age 14 to 16, leaving open the questions of whether parenting behaviors continue to influence narcissism development beyond adolescence and into adulthood. Longterm longitudinal research is needed to better understand the development of adult narcissism.

Second, most prior work on the antecedents of narcissism has focused on parenting behaviors, leaving other aspects of the family environment understudied. The family environment has been documented in longitudinal studies as a significant and enduring predictor of self-esteem (Krauss et al., 2020; Orth, 2018), a conceptual sibling of narcissism (Lawson & Robins, 2021). If narcissism is interpreted as dysfunctional self-esteem regulation (Morf & Rhodewalt, 2001; Thomaes et al., 2009; Tracy et al., 2009), then there is reason to suspect that the home environment would contribute to the cultivation of narcissism as well. For instance, a strained relationship between a parent and a child may have deleterious effects on their normative development and foster a fragile ego. The quality as well as the degree of conflict between parents might impact their offspring, since a fulfilling and satisfying romantic relationship can lay the foundation for a nurturing approach to parenting (Krishnakumar & Buehler, 2000; Horton et al., 2006; Rothstein, 1979). Sibling relationships may also have a profound impact on an individual’s narcissistic development, given that siblings serve as one of the earliest forms of social comparison and self-evaluation (Ferencz et al., 2023). Beyond parenting behaviors and relationship quality/conflict, parenting goals that are oriented toward either agency (gaining status) or communion (enhancing social bonds) may shape the internalization of their children’s values and engender or suppress narcissistic tendencies, especially since narcissism is associated with more agentic and less communal goals (Findley & Ojanen, 2013).

Despite these conceptual reasons to expect that family factors beyond parenting behaviors might influence the development of narcissism, only a handful of studies (and only one longitudinal study) have reported associations between the family environment and child narcissism (see Table 1). Thus, our research takes a first look at how various unexplored features of the family environment in childhood and adolescence predict the development of narcissism in adulthood in an understudied, minoritized population.

Table 1.

Summary of Research on the Family Environment and Narcissism

Authors	Family environment	Type of study	Sample size	Average age	Findings
Jonason et al. (2014)	Parent-child relationship quality	Cross-sectional	352	M = 25.1, SD = 9.8	Low maternal care was associated with entitlement, exploitativeness, and leadership; high paternal care was associated with entitlement and exploitativeness
Tajmirriyahi et al. (2021)	Parent-child relationship quality	Cross-sectional	262	M = 23.2, SD = 3.9	Parental support, conflict, and depth were not associated with child narcissism
Jung & Schröder-Abé (2025)	Parent-child relationship quality	Cross-sectional, longitudinal	2,008	M = 10.5, SD = 2.7 at each timepoint (range between 5–16 years old)	Decreased intimacy and admiration was cross-sectionally (but not longitudinally) associated with child antagonistic narcissism
Finzi-Dottan & Cohen (2010)	Sibling relationship quality	Cross-sectional	202	M = 26.7, SD = 6.3	Child narcissism was associated with greater conflict between siblings
Ferencz et al. (2023)	Sibling relationship quality	Cross-sectional	117	M = 30.4, SD = 10.6	Sibling relationship quality was not associated with child narcissism
Kauten et al. (2015)	Parent-child conflict, parental conflict	Cross-sectional	153	M = 17.1, SD = 1.0	Marital discord and parent-child conflict moderated the relationship between child narcissism and aggression

The Present Study

The present study examined prospective associations between the family environment and narcissism using data from the California Families Project, an ongoing 17-year longitudinal study of Mexican-origin youth (N = 674) assessed 13 times from late childhood to emerging adulthood. Specifically, we examined the prospective effects of family factors (assessed at ages 10, 12, 14, and 16) on grandiose narcissism (assessed at ages 14, 16, 19, 21, 23, and 26) and vulnerable narcissism (assessed at ages 19, 21, 23, and 26). We addressed the following research questions (see Figure 1 for a visual depiction):

Figure 1. Visual depiction of research questions.

Note. Fam = Family Environment Factors; Narc = Narcissism (Grandiose and Vulnerable). We examined RQ1-3 separately for grandiose and vulnerable narcissism. For RQ4, we controlled for prior levels of grandiose narcissism (vulnerable narcissism was not assessed at earlier waves). For RQ5, we compared the strength of the Fam→Narc associations for boys vs. girls and for youth born in the U.S. vs. Mexico.

1. Are family factors in adolescence prospectively associated with narcissism in emerging adulthood? Specifically, are parenting behaviors, family relationship quality (parent-child, inter-parental/marital, sibling-child), family conflict, and parenting goals averaged across ages 10 to 16 associated with grandiose and vulnerable narcissism averaged across ages 19 to 26?

2. Is late childhood a sensitive period for the development of adult narcissism or are the associations stronger for more proximal (i.e., late adolescent) family experiences? Specifically, do family experiences in late childhood (age 10-12) predict grandiose and vulnerable narcissism (averaged across ages 19 to 26) as strongly as family experiences later in adolescence (age 14-16)?

3. How long-term are the prospective associations between family experiences and emerging adult narcissism? Specifically, do family experiences (averaged across age 10-16) predict grandiose and vulnerable narcissism 3 years (age 19), 5 years (age 21), 7 years (age 23), and 10 years (age 26) later?

4. Do any observed associations between family experiences and emerging adult grandiose narcissism hold after controlling for grandiose narcissism at age 14 or 16?

5. Do the findings for RQ1-4 differ for boys versus girls (sex) or for youth born in the U.S. versus Mexico (nativity)?

With regard to RQ1, prior research and theory did not allow us to form nuanced hypotheses about how each particular family factor would be associated with adult narcissism. However, in general, we predicted that positive aspects of the family environment (parental warmth/support and the quality of parent-child, parental/marital, and sibling-child relationships) would be associated with lower levels of narcissism, whereas toxic aspects of the family environment (parental hostility, parent-child conflict, parental/marital conflict, and child defiance of parents) would be associated with higher levels of narcissism, especially factors directly related to the emotional bond between parents and children (i.e., parenting behaviors, parent-child relationship quality, parent-child conflict, child defiance of parents). In addition, we predicted that parenting goals related to higher agency and lower communion would be associated with higher levels of narcissism.

With regard to RQ2, we had two competing hypotheses. Consistent with the idea of a sensitive window during the transition into adolescence, we expected that family experiences in late childhood (age 10 & 12) would be more strongly associated with adult narcissism than family experiences at late adolescence (age 14 & 16). However, consistent with the idea that proximal experiences are more consequential than distal experiences (i.e., more distal experiences dissipate over time), we expected that family experiences assessed in late adolescence would have a stronger effect than those assessed earlier in adolescence, with the associations decreasing in strength from the most proximal (age 16) to the most distal (age 10) assessments.

With regard to RQ3, we expected that family factors would be more strongly associated with narcissism earlier in emerging adulthood (e.g., age 19) compared to later in emerging adulthood (e.g., age 26), based on the proximity principle and the fact that narcissism later in emerging adulthood is likely influenced by myriad factors related to the transition into adult social roles, such as launching one’s career, getting married, having children, and other life experiences that typically occur later (vs. earlier) in emerging adulthood.

We did not have hypotheses for RQ4 or RQ5 and considered these analyses exploratory.

The present study extends prior research on the development of narcissism in several ways. First, we used longitudinal data from late childhood (age 10) to emerging adulthood (age 26), which allowed us to examine the long-term prospective effects of family factors on later narcissism. Second, we examined a wide range of family factors, including three different kinds of family relationships (parent-child, sibling-child, and mom-dad). Third, we used multimethod assessments of parenting behaviors (including child report, spouse report, and behavioral coding of parent-child interactions) and multi-informant assessments of most other family factors to reduce the effects of response biases and shared method variance. Fourth, we applied latent variable modeling techniques to investigate relations between the family environment and narcissism, thereby accounting for measurement error. Fifth, instead of treating narcissism as a unidimensional construct, we examined two facets of grandiosity (superiority and exploitativeness) and vulnerability, allowing us to explore how family factors are associated with three different manifestations of narcissism. Lastly, we examined our research questions using data from a sample of Mexican-origin youth, a population that has been historically underrepresented in psychological science, especially in research on narcissism.

Methods

Ethics Statement

This study was approved by the University of California, Davis Internal Review Board (IRB #217484-35; “California Families Project”).

Participants

We used data from the California Families Project (CFP), an ongoing longitudinal study of 674 Mexican-origin children (50% female) and their parents living in the United States¹. Existing published work has utilized the CFP to examine the stability and change of narcissism using latent variables of grandiosity and vulnerability (Ma et al., 2025) and to examine the effects of parenting practices on narcissism (Wetzel et al., 2016). Other than parceling grandiose and vulnerable narcissism indicators in the same manner as Ma et al. (2025), this study’s use of structural equation modeling to model the influence of family environment factors on narcissistic development across a longer 17-year period has not been examined with CFP data until now. A complete list of CFP publications can be found here: https://osf.io/ky7cw/.

The children (M_age = 10.8, SD = .61) were drawn at random from student rosters provided by two school districts in Northern California with the following criteria: the child had to be in the 5th grade, of Mexican origin, and living with their biological mother. The interviews were conducted by trained interviewers at the participants’ homes in either Spanish or English, depending on the participant’s preference.

The mean ages of the focal children from Waves 1 to 13 are 10.8, 11.8, 12.8, 13.8, 14.8, 15.8, 16.8, 17.7, 18.8, 19., 21.7, 23.2, and 26.1, respectively. Retention rates compared with the original sample of 674 youth are as follows: 84% (Wave 2), 86% (Wave 3), 88% (Wave 4), 90% (Wave 5), 88% (Wave 6), 89% (Wave 7), 89% (Wave 8), 87% (Wave 9), 87% (Wave 10), 80% (Wave 11), 81% (Wave 12), and 75% (Wave 13). As for the families, 549 were two-parent households and 124 were single-parent households (mothers only). Sixty-three percent of mothers and 65% of fathers had less than a high school education (median education level = 9th grade for both mothers and fathers). At baseline, the median total household income was $32,500. With regard to generational status, 84% of mothers and 89% of fathers were first-generation immigrants.

As for attrition analyses, we compared individuals who participated versus did not participate in the last wave of data (Wave 13) on their baseline scores of narcissism (Wave 5) as well as demographic variables (gender and nativity). Individuals who dropped out at Wave 13 did not have significantly different baseline narcissism scores from those who participated in Wave 13, t(195) = −.27, p = .79, nor did these groups differ significantly in whether they were born in the U.S. or in Mexico, ${\chi }^2$(1) = .66, p = .42. Participation in the last wave differed significantly by gender, such that boys were more likely to drop out by Wave 13 compared to girls, ${\chi }^2$(1) = 9.20, p = .002.

Measures

Narcissism

Narcissism was assessed using the Narcissistic Personality Questionnaire for Children–Revised (NPQC-R; Ang & Raine, 2009), a self-report questionnaire designed for children and adolescents, at ages 14, 16, 19, 21, 23, and 26. The NPQC-R measures two facets of narcissism, superiority and exploitativeness. Superiority (6 items) captures the grandiose aspects of narcissism, such as 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 (6 items) captures the interpersonally toxic 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 to all items on a five-point rating scale ranging from 1 (not at all like me) to 5 (completely like me).

Vulnerable narcissism (7 items) was assessed at ages 19, 21, 23, and 26 using items from the vulnerability subscale of the Brief Pathological Narcissism Inventory (B-PNI; Schoenleber et al., 2015). This scale includes items related to four facets of vulnerability: Contingent Self-Esteem (“It’s hard to feel good about yourself unless you know other people admire you”), Hiding the Self (“When others realize how needy you are, you feel anxious and ashamed”), Devaluing (“Sometimes you avoid people because you’re concerned that they will disappoint you”), and Entitlement (“You will never be satisfied until you get all that you deserve”). All items were rated on a scale ranging from 1 (not at all like me) to 5 (completely like me).

To address our research questions, we created mean composites of grandiose narcissism (superiority and exploitativeness) and vulnerable narcissism across ages 19 to 26. Grandiose narcissism at age 14 and 16 was used as a control variable for RQ4 to test whether observed associations between family experiences and emerging adult grandiose narcissism hold after controlling for prior levels of grandiose narcissism.

Family Environment Factors

All family factors were assessed at ages 10, 12, 14, and 16, except for parental conflict, which was assessed at ages 10, 12, and 14.

Parenting practices.

We assessed the three primary domains of parenting—warmth, hostility, and monitoring—separately for mothers and fathers, using a multi-method composite of child-reports, spouse-reports, and observational data (see Atherton et al., 2020, for details about how the parenting composites were created). The child and parents reported on parenting using the Behavioral Affective Rating Scale (Conger, 1989a), the Iowa Parenting Scale (Conger, 1989b), and the Parental Monitoring of Child Scale (Small & Kerns, 1993). Observational data come from coding of videotaped mom-child and dad-child interactions (Atherton et al., 2020).

Parent-child relationship quality.

The quality of the parent-child relationship was assessed using multi-informant data from mothers, fathers, and children. Youth rated the quality of their relationship with their mom and dad using a 3-item scale, and parents rated the quality of their relationship with their child using a 4-item scale (there were two overlapping items across the child and parent scales). All items were rated on a 4-point Likert scale, with the scale anchors varying across items. Example items include “How satisfied are you with your relationship with [NAME OF CHILD/MOM/DAD]?” and “How happy are you with the way things are between you and [NAME OF CHILD/MOM/DAD]?” Latent variables were created at each wave with child, mom, and dad reports as indicators of the latent variable, which allowed us to control for measurement error and for the unique biases of family members.

Quality of parental relationship.

The quality of parental relationship was assessed with the 5-item Quality of Marital/Romantic Relationship scale (Yeh et al., 2006). Reports were provided by mothers and fathers. Parents rated each item on a 4-point Likert scale, ranging from 1 (not at all true) to 4 (very true). Example items include “You have a good relationship,” “Your relationship with [PARTNER] is very stable,” and “Your relationship with [PARTNER] makes you happy.” The ratings by mothers and fathers were used as indicators of latent variables at each wave.

Relationship with closest sibling.

Relationship with one’s closest sibling was assessed using a 14-item scale developed for use in the California Families Project. Youth rated each item on a 4-point Likert scale, ranging from 1 (almost never or never) to 4 (almost always or always). Example items include “How often do you go to [SIB NAME] for advice or support?”, “[SIB NAME] accepts you no matter what you do,” and “How often do you and [SIB NAME] argue with each other?” (reverse scored). We used item parcels as indicators of latent variables at each wave.

Parental conflict.

Child’s perception of interparental conflict was assessed using a 45-item scale (Grych, Seid, & Fincham, 1992). Participants rated each item on a 4-point Likert scale, ranging from 1 (Not at all true) to 4 (Very true). Example items include “Your parents blame you when they have arguments or disagreements,” “When your parents argue or disagree, you worry that one of them will get hurt,” and “When your parents have an argument or disagreement, they usually work it out” (reversed). Factor analyses of the 45 items were conducted to identify empirically and psychologically coherent subscales, which we identified 3 relevant factors that items loaded onto: frequency/intensity, lack of resolution, and self-blame. For each subscale, we used item parcels as indicators of latent variables at each wave.

Parent-child conflict.

Parent-child conflict was assessed using a 10-item scale. Participants rated each item on a 4-point Likert scale, ranging from 1 (almost never or never) to 4 (almost always or always). Example items include “Over the past 3 months, how often have you and your mom [dad] disagreed with each other?”, “You and your mom [dad] gave each other dirty looks or rolled your eyes,” and “You and your mom [dad] yelled or raised your voices at each other.” Responses were collected from children, mothers, and fathers and were constructed to assess the parent-child relationship as a dyadic unit rather than as individual behaviors from each parent. The ratings by mothers, fathers, and children were used as indicators of latent variables at each wave.

Child defiance of parents.

Child defiance of parents was assessed using an 11-item scale (Ruiz & Gonzalez, 1988). Participants rated each item on a 4-point Likert scale, ranging from 1 (Almost never or never) to 4 (Almost always or always). Example items include “You argue with your mom until you get your way” and “You ignore your mom’s wishes and just do what you want.” Responses were collected from children, mothers, and fathers. The ratings by mothers, fathers, and children were used as indicators of latent variables at each wave.

Parenting goals.

Parenting goals were assessed using an 11-item scale developed for use in the California Families Project. Mothers and fathers rated each item on a 4-point Likert scale, ranging from 1 (not at all important) to 4 (very important). Example items include “How important is it that [NAME OF CHILD] is popular?”, “How important is it that [NAME OF CHILD]” is courteous to other people?”, and “How important is it that [NAME OF CHILD] respects and pays attention to you?” The ratings by mothers and fathers were composited, separately for each parenting goal. We classified these goals as agentic or communal, and then constructed latent variable measures of agentic and communal parenting goals at each wave, with ratings by mothers and fathers used as indicators of latent variables at each wave.

Statistical Analyses

All analyses were conducted using R and a broad overview of the analytic plan is presented in Figure 1. We used structural equation modeling to compute latent narcissism and latent family factor variables at each wave. We used item parcels or different informants (child, mom, dad) as indicators of latent variables at each wave. To see which items constituted the parcels used for each measure, please see Supplemental Table S1. To create the narcissism and family factors composites, we used unit-weighted composites of the latent factor scores at each wave. For all analyses, we assessed adequate model fit by changes in chi-square and degrees of freedom, comparative fit index (CFI) less than or equal to .01, and Tucker-Lewis index (TLI) less than or equal to .01 (Chen, 2007; Cheung & Rensvold, 2002; Meade et al., 2008). We also noted values of the root-mean-square error of approximation (RMSEA), for which adequate fit of the model is indicated by values less than or equal to .06 (Hu & Bentler, 1998, 1999). For all research questions except RQ4, we tested models separately for all possible combinations of family factors and narcissism (superiority, exploitativeness, vulnerability).

To address RQ1, we created a latent variable composite of each of the family factors from ages 10 to 16 and a latent variable composite of each narcissism construct (superiority, exploitativeness, vulnerability) from ages 19 to 26 and examined correlations between the two latent variables using SEM. Factor loadings for each family factor and narcissism facet were separately estimated at each wave, and composites were unit-weighted across waves.

For RQ2, we asked whether family factors assessed at different ages are prospectively associated with narcissism. To do so, we used the three latent variable narcissism composites (superiority, exploitativeness, vulnerability) from ages 19 to 26 as our indicator of narcissism in emerging adulthood. Using structural equation modeling, we regressed the narcissism composites on the latent family factors at ages 10, 12, 14, and 16, separately for each measure of narcissism and each family factor. We compared a model that constrained the effects to be equal across ages to a model that freely estimated the effects. If the constrained model did not fit significantly worse than the freely estimated model, we concluded that the associations do not differ significantly across ages.

RQ3 examined whether associations between family factors and narcissism persist throughout emerging adulthood. To do so, we used the latent variable composites of family factors from ages 10 to 16. We regressed this composite onto narcissism at each available age (19, 21, 23, and 26), separately for each measure of narcissism and each family factor. We compared a model that constrained the effects to be equal across ages to one that freely estimated the effects. If the constrained model did not fit significantly worse than the freely estimated model, we concluded that the associations were the same across ages.

To address RQ4, we built upon the model used in RQ1 for grandiose narcissism to examine how the composite of family factors in late childhood to late adolescence (age 10 to 16) predicts the composite of narcissism in emerging adulthood (ages 19 to 26) when controlling for adolescent grandiose narcissism at age 14 and at age 16. These analyses were only done for grandiose narcissism (superiority & exploitativeness) because vulnerable narcissism was not assessed until age 19.

To address RQ5, we conducted multiple group SEM models for RQ1-4 by comparing a model in which all parameters were freely estimated across groups (boys and girls, youth born in the U.S. and Mexico) to one where all parameters were constrained to equality across groups. If the constrained model did not fit significantly worse than the freely estimated model, we concluded that the associations were the same across groups.

Transparency and Openness

Pre-registrations of the research questions, hypotheses, methods, and analysis plans, our analytic code, and supplemental materials are provided on the Open Science Framework: https://osf.io/vsp8g/overview?view_only=16574fba22c24051b137f089ba7019c0. CFP codebooks and research materials can be found on https://osf.io/ky7cw/.²

Results

Descriptive statistics for narcissism measures are shown in Table S2. Heat maps of concurrent associations between family factors and Waves 5 and 7 are presented in Figures S1 and S2, respectively. Figures of the structural equation model outputs of RQs 1-4 using grandiosity and warmth as examples are provided in Figures S3–S6, respectively.

Research Question 1: Are family factors in adolescence prospectively associated with narcissism in emerging adulthood?

We examined prospective associations between family factors averaged across ages 10-16 and narcissism averaged across ages 19-26. We found numerous significant associations (see Table 2). For parenting behaviors, warmth positively predicted grandiosity (r = .14) and superiority (r = .24), hostility positively predicted exploitativeness (r = .19) and vulnerability (r = .17), and monitoring positively predicted superiority (r = .15). These results support our hypotheses that more toxic aspects of parenting (i.e., hostility) are associated with higher levels of narcissism, but only for maladaptive forms of narcissism such as exploitativeness and vulnerability. In contrast, we did not find support for our hypotheses that positive aspects of parenting (i.e., warmth) would be associated with lower levels of narcissism, for any form of narcissism.

Table 2.

RQ1: Associations between Composites of Family Factors (Age 10 to 16) and Composites of Narcissism (Age 19 to 26)

	Grandiosity		Superiority		Exploitativeness		Vulnerability
Predictor	r [CI]	p	r [CI]	p	r [CI]	p	r [CI]	p
Parenting Behaviors
Warmth	0.14 [0.04, 0.23]	0.004	0.24 [0.14, 0.33]	< .001	0.03 [−0.07, 0.13]	0.555	−0.08 [−0.17, 0.01]	0.084
Hostility	0.09 [−0.01, 0.18]	0.076	−0.04 [−0.14, 0.06]	0.405	0.19 [0.09, 0.28]	< .001	0.17 [0.08, 0.26]	< .001
Monitoring	0.09 [−0.01, 0.18]	0.076	0.15 [0.06, 0.25]	0.002	0.02 [−0.08, 0.12]	0.721	−0.06 [−0.15, 0.03]	0.219
Relationship Quality
Parent-Child Relationship Quality	0.00 [−0.10, 0.11]	0.945	0.11 [0.00, 0.21]	0.040	−0.09 [−0.19, 0.02]	0.097	−0.01 [−0.11, 0.09]	0.842
Parental Relationship Quality	0.07 [−0.03, 0.17]	0.196	0.11 [0.01, 0.21]	0.035	0.03 [−0.07, 0.13]	0.566	−0.02 [−0.11, 0.08]	0.757
Sibling Relationship Quality	0.13 [0.02, 0.24]	0.018	0.19 [0.08, 0.30]	0.001	0.06 [−0.05, 0.17]	0.287	−0.02 [−0.13, 0.09]	0.711
Relationship Conflict
Parent-Child Conflict	0.13 [0.03, 0.23]	0.009	0.03 [−0.07, 0.13]	0.548	0.20 [0.10, 0.30]	< .001	0.12 [0.03, 0.22]	0.011
Frequency of Parental Conflict	−0.01 [−0.12, 0.10]	0.799	−0.12 [−0.23, −0.01]	0.039	0.07 [−0.04, 0.18]	0.196	0.11 [0.00, 0.22]	0.045
Lack of Resolution of Parental Conflict	−0.09 [−0.20, 0.01]	0.090	−0.11 [−0.23, −0.00]	0.048	−0.06 [−0.17, 0.05]	0.263	0.04 [−0.07, 0.15]	0.484
Self-Blame in Parental Conflict	0.05 [−0.08, 0.18]	0.448	−0.10 [−0.23, 0.03]	0.131	0.15 [0.03, 0.28]	0.018	0.25 [0.13, 0.37]	< .001
Child Defiance of Parents	0.28 [0.18, 0.39]	< .001	0.19 [0.08, 0.30]	0.001	0.32 [0.21, 0.42]	< .001	0.07 [−0.04, 0.18]	0.206
Parenting Goals
Agentic Parenting Goals	0.13 [0.03, 0.23]	0.014	0.12 [0.01, 0.22]	0.030	0.12 [0.01, 0.22]	0.031	−0.01 [−0.11, 0.09]	0.882
Communal Parenting Goals	−0.00 [−0.10, 0.10]	0.965	−0.00 [−0.10, 0.10]	0.986	−0.00 [−0.11, 0.10]	0.934	−0.03 [−0.13, 0.07]	0.552

Note: Values in bold are significant at a p < .05 level. Values in brackets are the 95% confidence intervals.

We found further evidence supporting associations between positive aspects of the family environment and more adaptive forms of narcissism. Parent-child relationship quality (r = .11) and parental relationship quality (r = .11) were positively associated with superiority, and sibling relationship quality was associated with grandiosity (r = .13) and its facet superiority (r = .19).

Interestingly, all three measures of conflicts between family members were associated with both adaptive and maladaptive forms of narcissism. First, parent-child conflict was positively associated with grandiosity (r = .13), exploitativeness (r = .20), and vulnerability (r = .12). Second, child defiance of parents was positively associated with grandiosity (r = .28), superiority (r = .19), and exploitativeness (r = .32). Third, frequency/intensity of parental conflict was positively associated with vulnerability (r = .11) but negatively associated with superiority (r = −.12); self-blame was positively associated with exploitativeness (r = .15) and vulnerability (r = .25); and, lack of resolution was negatively associated with superiority (r = −.11).

Agentic and communal parenting goals in adolescence also predicted adult narcissism. Consistent with our predictions, agentic parenting goals were positively associated with grandiosity (r = .13), superiority (r = .12), and exploitativeness (r = .12). Contrary to our predictions, communal parenting goals were not associated with any form of narcissism.³

Research Question 2: Is early adolescence a sensitive period for the development of adult narcissism, or are the associations stronger for more proximal family experiences?

We tested whether more distal (e.g., age 10) or more proximal (e.g., age 16) family experiences would have a stronger impact on the development of adult narcissism by regressing the composite of narcissism onto individual waves of family factors and comparing a freely estimated to a constrained model. Model comparisons between constrained and freely estimated effects are shown in Table S4. The constrained model fit significantly worse than the freely estimated model in only two out of 52 cases (3.8%; see Table S5 for main effects), indicating that there were few differential associations across time (the two exceptions being parental hostility and vulnerable narcissism, and lack of resolution of parental conflict and vulnerable narcissism). Notwithstanding these two exceptions, the overall pattern of findings does not support the idea that the timing of family experiences differentially impacts the development of narcissism.

Research Question 3: How enduring are the prospective associations between adolescent family experiences and emerging adult narcissism?

We examined the persistence of the effects of family experiences on narcissism by regressing individual waves of narcissism onto the composite of each family factor and comparing a freely estimated model to a constrained model (see Table S6). The constrained model fit significantly worse in only three out of 52 cases (5.8%; see Table S7 for main effects). None of the three significant cases (self-blame and exploitativeness, child defiance and grandiosity, child defiance and exploitativeness) showed a systematic age difference (e.g., progressively stronger or weaker effects from age 19 to 26). Overall, these findings do not support our hypothesis that family experiences would be more strongly associated with narcissism earlier (versus later) in emerging adulthood.

Research Question 4: Do any observed associations between adolescent family experiences and adult grandiose narcissism hold after controlling for grandiose narcissism in adolescence?

Table 3 shows the prospective associations between adolescent family experiences and adult grandiose narcissism, controlling for grandiose narcissism at age 14 or 16⁴. (Vulnerable narcissism was not assessed until age 19, and therefore prior levels cannot be controlled.) Most significant associations no longer held after controlling for prior levels of narcissism. However, the prospective effects still held between parental warmth and exploitativeness, parental monitoring and exploitativeness, parental relationship quality and exploitativeness, lack of resolution and exploitativeness, self-blame and exploitativeness, and agentic parenting goals and grandiosity (superiority and exploitativeness). Nearly all of the aforementioned associations were significant when controlling for narcissism at age 14, but not at age 16. Together, the findings suggest that family experiences predict adult narcissism via their concurrent effects on adolescent narcissism.

Table 3.

RQ4: Prospective Associations of Family Experiences at Ages 10 to 16 and Narcissism at Ages 19 to 26, Controlling and not Controlling for Narcissism at Ages 14 and 16

	Grandiosity	Superiority	Exploitativeness
Predictor	r (Age 14, 16)	r (Age 14, 16)	r (Age 14, 16)
Parenting Behaviors
Warmth	0.14** (0.06, 0.06)	0.24*** (0.02, 0.04)	0.03 (0.13**, 0.03)
Hostility	0.09 (0.02, 0.05)	−0.04 (0.03, −0.00)	0.19* (−0.06, 0.07)
Monitoring	0.09 (0.03, 0.01)	0.15** (−0.02, −0.00)	0.02 (0.12*, 0.04)
Relationship Quality
Parent-Child Relationship Quality	0.00 (0.00, −0.05)	0.11* (−0.05, 0.00)	−0.09 (0.10, −0.06)
Parental Relationship Quality	0.07 (0.09, 0.03)	0.11* (0.06, 0.06)	0.03 (0.16*, 0.03)
Sibling Relationship Quality	0.13* (0.06, 0.01)	0.19** (0.08, 0.08)	0.06 (0.04, −0.05)
Relationship Conflict
Parent-Child Conflict	0.13** (0.02, 0.07)	0.03 (0.05, 0.02)	0.20*** (−0.05, 0.08)
Frequency/Intensity of Parental Conflict	−0.01 (−0.12, −0.01)	−0.12* (−0.04, −0.02)	0.07 (−0.05, −0.06)
Lack of Resolution of Parental Conflict	−0.09 (−0.05, 0.00)	−0.11* (0.07, 0.01)	−0.06 (−0.17**, −0.02)
Self-Blame in Parental Conflict	0.05 (−0.05, 0.06)	−0.10 (−0.00, −0.08)	0.15* (−0.17*, 0.15*)
Child Defiance of Parents	0.28*** (0.04, 0.02)	0.19** (0.06, 0.02)	0.32*** (−0.12, 0.01)
Parenting Goals
Agentic Parenting Goals	0.13* (0.23***, 0.05)	0.12* (0.17**, 0.08)	0.12* (0.23***, 0.03)
Communal Parenting Goals	−0.00 (0.07, −0.04)	−0.00 (0.02, 0.01)	−0.00 (0.10, −0.06)

Note: Values outside of the parentheses represent the prospective correlations (from Table 2) without controlling for prior narcissism. Values inside the parentheses represent the partial correlations when controlling for narcissism at age 14 (left side) or age 16 (right side). Values in bold are significant at a p < .05 level.

^*indicates significance levels of p-value < .05,

^**indicates significance levels of p-value < .01, and

^***indicates significance values of < .001. Please see Table S18 for full p-values.

We also ran non-preregistered exploratory analyses to examine whether age 16 narcissism mediated the effects of composites of family factors on composites of narcissism. We found significant indirect effects between grandiosity and warmth, monitoring, sibling relationship quality, parent-child conflict, lack of resolution of parental conflict, and child defiance of parents; superiority and warmth, monitoring, parent-child and sibling relationship quality, lack of resolution of parental conflict, and child defiance of parents; and vulnerability and hostility, sibling relationship quality, parent-child conflict, frequency of parental conflict, and child defiance of parents. In other words, these family experiences are associated with adolescent narcissism at age 16, which, in turn, is associated with narcissism in emerging adulthood. Full results are presented in Table S8.

Research Question 5: Do the findings for RQ1 and RQ4 differ for boys versus girls (sex) or for youth born in the U.S. versus Mexico (nativity)?

Descriptive statistics for narcissism by sex and nativity are presented in Table S9.

We first tested whether the findings from RQ1 to RQ4 generalize across sex using multi-group SEM analyses that compared constrained to freely estimated models. In the vast majority of cases (219 out of 232, or 94%), the findings did not differ significantly between boys and girls (see Tables S10–11 for freely estimated parameters, and see Tables S12–13 for model fit comparisons). For the 13 cases where constraining the effects to be equal across sex fit significantly worse than freely estimating the effects, the associations for boys tended to be stronger than those for girls. However, we are hesitant to interpret these effects because the number of significant effects is similar to what would be expected by chance.

In most cases (202 out of 233, or 87%), the findings did not differ significantly by nativity. Table 4 shows freely estimated effects by nativity for the significant cases that generalized across RQ1 and RQ4 (see Tables S14–15 for RQ2-3 findings, and Tables S16–17 for model fit comparisons). Intriguingly, many of the positive significant associations between narcissism and positive aspects of the family environment, such as warmth, monitoring, and parent-child and sibling relationship quality, held only for U.S. born youth. In contrast, Mexican-born youth had either null or negative associations between narcissism and positive aspects of the family environment, and had positive significant associations between narcissism and negative aspects of the family environment like hostility.

Table 4.

RQ5: Freely Estimated Associations between Family Experiences and Narcissism, Separately for U.S. Born Youth and Mexican-Born Youth

		U.S. born youth		Mexican-born youth
Association	RQ	r [CI]	p	r [CI]	p
Warmth
Warmth → Superiority	RQ1	0.21 [0.13, 0.30]	< .001	0.06 [−0.05, 0.17]	0.299
Warmth → Vulnerability	RQ1	−0.01 [−0.06, 0.04]	0.686	−0.14 [−0.26, 0.03]	0.013
Warmth → Grandiosity (controlling for age 14 Grandiosity)	RQ4	0.08 [0.01, 0.16]	0.028	−0.07 [−0.19, 0.05]	0.268
Warmth → Exploitativeness (controlling for age 14 Exploitativeness)	RQ4	0.17 [0.07, 0.26]	0.001	−0.04 [−0.15, 0.06]	0.418
Hostility
Hostility → Grandiosity	RQ1	0.02 [−0.08, 0.11]	0.722	0.19 [0.05, 0.32]	0.007
Hostility → Superiority	RQ1	−0.10 [−0.20, 0.00]	0.056	0.08 [−0.04, 0.20]	0.174
Hostility → Grandiosity (controlling for age 14 Grandiosity)	RQ4	−0.05 [−0.13, 0.04]	0.283	0.15 [0.02, 0.27]	0.023
Hostility → Superiority (controlling for age 14 Superiority)	RQ4	−0.03 [−0.12, 0.06]	0.530	0.12 [0.01, 0.24]	0.039
Hostility → Exploitativeness (controlling for age 14 Exploitativeness)	RQ4	−0.12 [−0.24, −0.01]	0.036	0.10 [−0.02, 0.22]	0.094
Hostility → Exploitativeness (controlling for age 16 Exploitativeness)	RQ4	0.01 [−0.07, 0.10]	0.743	0.16 [0.05, 0.27]	0.003
Monitoring
Monitoring → Vulnerability	RQ1	0.01 [−0.04, 0.06]	0.771	−0.13 [−0.23, −0.03]	0.011
Monitoring → Exploitativeness (controlling for age 14 Exploitativeness)	RQ4	0.13 [0.04, 0.23]	0.004	−0.02 [−0.11, 0.08]	0.740
Parent-Child Relationship Quality
Parent-Child Relationship Quality → Grandiosity	RQ1	0.20 [−0.08, 0.48]	0.162	−0.40 [−0.80, −0.01]	0.045
Parent-Child Relationship Quality → Superiority	RQ1	0.47 [0.18, 0.77]	0.002	−0.18 [−0.53, 0.17]	0.314
Parent-Child Relationship Quality → Exploitativeness	RQ1	−0.04 [−0.36, 0.28]	0.803	−0.49 [−0.83, −0.15]	0.005
Parent-Child Relationship Quality → Grandiosity (controlling for age 14 Grandiosity)	RQ4	0.23 [−0.03, 0.48]	0.084	−0.37 [−0.81. 0.07]	0.097
Parent-Child Relationship Quality → Superiority (controlling for age 16 Superiority)	RQ4	0.15 [−0.10, 0.41]	0.248	−0.21 [−0.48, 0.05]	0.111
Parent-Child Relationship Quality → Exploitativeness (controlling for age 14 Exploitativeness)	RQ4	0.52 [0.17, 0.87]	0.003	−0.31 [−0.61, −0.02]	0.036
Sibling Relationship Quality
Sibling Relationship Quality → Exploitativeness (controlling for age 14 Exploitativeness)	RQ4	0.21 [0.00, 0.42]	0.049	−0.29 [−0.62, 0.05]	0.090
Resolution of Parental Conflict
Resolution of Parental Conflict → Vulnerability	RQ1	−0.06 [−0.30, 0.17]	0.599	0.65 [0.03, 1.28]	0.040

Note: Values in bold are significant at a p < .05 level. Only models that showed significant sex differences in the effects are included in the table. Values in brackets are the 95% confidence intervals.

Discussion

Our study aimed to identify the family factors in late childhood and early adolescence that are associated with narcissism in emerging adulthood using multimethod data from a longitudinal study of Mexican-origin youth. We found evidence linking earlier parenting behaviors, relationship quality, relationship conflict, and parenting goals to emerging adult narcissism. These findings held across long periods of time (e.g., 10 years), but not after controlling for prior levels of narcissism.

Are family factors in adolescence associated with narcissism in emerging adulthood, and is there a sensitive period of development?

In alignment with our expectations that toxic family environments would cultivate narcissistic development, parental hostility and elevated conflict between parents and children were significantly associated with higher levels of exploitativeness and vulnerability, but not superiority. Hostile parenting may contribute to tense parent-child dynamics that lead children to repeat dysfunctional patterns in their adult lives as a manifestation of their exploitative tendencies. The self-blame subscale of parental conflict predicted subsequent vulnerability and exploitativeness, whereas the frequency and intensity of parental clashes predicted vulnerability. In other words, children who felt more at fault for frequent volatile disputes had more fragile egos in young adulthood, suggesting that these youth tended to internalize their role in parental arguments in ways that may have led them to feel a constant (and unsatisfied) need to validate their insecure selves.

Counter to our expectations, we found that various positive aspects of the family environment from ages 10 to 16 were precursors of the superiority facet of grandiosity from ages 19 to 26. For example, parental warmth and monitoring, parent-child relationship quality, parental relationship quality, and sibling relationship quality all predicted higher levels of superiority. There are several explanations for why positive aspects of the family environment were associated with more adaptive characteristics of narcissism (specifically superiority). First, warm and supportive parenting, as well as healthy and fulfilling relationships with family members, may engender positive beliefs about oneself and one’s capabilities, as expressed in items such as “You are a really special person.” Second, positive relationship quality between parents and between the parent and the child may be reciprocally related to parenting behaviors. For example, warm parenting may foster trusting and supportive family relationships, and positive relationships can provide the necessary space to cultivate a nurturing parenting style (Horton et al., 2006).

One explanation for the somewhat divergent findings for exploitativeness/vulnerability versus superiority is that there are both adaptive and maladaptive forms of narcissism (Cai & Luo, 2018). This pattern can be understood in the context of longstanding theories of narcissistic development, and indeed, our support for both psychoanalytic and social learning perspectives highlights that each may differentially apply to narcissistic domains. Psychoanalytic theory, which argues that cold parenting (and perhaps by extension, negative family characteristics) induces narcissism, is most applicable to exploitativeness and vulnerability (Kernberg, 1975; Kohut, 1977). Conversely, the social learning theory, which posits that indulgent parenting predicts narcissism, applies more to superiority (Millon, 1981; Wetzel & Robins, 2016). Together, our findings suggest that these theories should be jointly considered as narcissism is best understood as a system composed of both the grandiose and vulnerable self.

A few family factors contributed to both “adaptive” and “maladaptive” aspects of narcissism. For example, agentic parenting goals were positively linked to superiority and exploitativeness. Although this finding is in line with prior research indicating that narcissism is associated with self-interested agency (Findley & Ojanen, 2013), our study uniquely connects agentic parenting goals with narcissism development. Children are highly susceptible to the thoughts and behaviors of their parents and could internalize the goals that parents (knowingly or unknowingly) place on them. Moreover, goals that parents convey to their children are theorized to influence parenting behaviors (Mageau et al., 2016); for example, goals that are related to children’s performance (i.e., wanting children to perform better than peers) are associated with controlling parenting as opposed to goals that are oriented toward mastery (i.e., wanting children to learn new skills and grow; Gonida & Cortina, 2014; Mageau et al., 2016). Parenting goals could therefore be one component involved in explaining the psychodynamic pathway from parental hostility to narcissism. Perhaps it is through having more performance-related goals that parents are fostering environments in which children learn that achievement contexts and external validation are critical for self-worth. It is worth noting that we did not directly examine performance versus mastery goals, but rather agentic versus communal goals; nonetheless, there may be ties between agentic goals and performance goals that remain to be explored. Contrary to our expectations, communal parenting goals were not negatively associated with any narcissism facets. This suggests that the absence of communal goals is not as influential as the presence of agentic goals, meaning that valuing achievement and agency play a more central role in engendering narcissistic traits than communal values play in preventing them.

Defiance of parents also predicted both superiority and exploitativeness, indicating that these facets of narcissism are both rooted in childhood oppositional behavior. Disobedient tendencies may be an early sign that a child has inflated self-beliefs and a lack of deference toward authority figures. This finding supports prior research that temperamental characteristics like impulse dysregulation (i.e., pushing and stretching limits) as early as preschool ages are predictive of young adult narcissism (Carlson & Gjerde, 2009). Previous work using the same sample, albeit across a much shorter timespan, found that exploitativeness is predictive of the defiance subscale of oppositional defiance disorder (Wetzel et al., 2021); our research expands upon this by suggesting that defiance also in turn predicts exploitativeness, reflecting a dysfunctional developmental cycle where the two negative characteristics reinforce each other over time. Behaviorally, exploitativeness is indexed primarily by manipulative action tendencies (i.e., “You are good at getting people to do things your way”; Wetzel et al., 2021), so being able to bend the rules with primary caregivers in childhood (i.e., “You ignore your mom’s wishes and just do what you want”) may give individuals the impetus to do the same toward others in adulthood. Cognitively, successfully rebelling against authority figures may fuel delusions of grandeur and reinforce the belief that one is superior to others. Where this defiance derives from is not the subject of this study. There is reason to suspect that genetic influences play a large part (Hudziak et al., 2005). Environmental influences such as overly strict parenting could also contribute to a hostile environment that then brings about defiance in offspring. We do not explicitly examine the interactions between various family factors in producing narcissism in this study—for example, it could be via parental hostility that defiance is linked to exploitativeness—leaving open future avenues for investigation.

Furthermore, we aimed to understand not only which family factors prospectively predict narcissism, but also when these family factors have the largest impact, and how lasting these impacts are. Contrary to our hypotheses, we did not find support for the idea of a critical window or that more proximally assessed family factors would have the strongest effects. It is possible that (1) the cumulative impact of family processes over time contributes to narcissism, (2) family factors from ages 10 to 16 are equally impactful on narcissism, (3) the “critical window” for which family factors are most influential occur outside of our age of assessment, or (4) there are substantial individual differences in when youth are most affected by their childhood experiences, making it difficult to find a normative pattern. Future research should explore these possibilities.

Do any observed associations between adolescent family experiences and adult grandiose narcissism hold after controlling for grandiose narcissism in adolescence?

Most prospective associations between adolescent family experiences and emerging adult grandiose narcissism no longer held after controlling for prior levels of narcissism. These findings imply that family factors exert influence on narcissism at earlier ages than when we first assessed narcissism. Several associations held after controlling for age 14 narcissism, but with one exception (self-blame for parental conflict→vulnerability), none were significant after controlling for narcissism at age 16. These effects suggest that family factors may have a unique impact on narcissistic development that is not fully accounted for by grandiosity at age 14, and that by 16, narcissism has already developed from the cumulative prior effects of family factors. In general, the findings suggest a developmental process in which family factors shape the development of narcissism during adolescence, and another in which adolescent narcissism remains stable into adulthood. In turn, these two processes together drive the prospective effects of the family environment on adult narcissism. Thus, the findings highlight the importance of collecting longterm longitudinal data to better disentangle concurrent from prospective associations.

Do the findings differ for boys versus girls or for youth born in the U.S. versus Mexico?

In nearly all cases, the associations between family factors and narcissism did not differ significantly between boys and girls, but the associations differed more frequently for youth born in the U.S. versus Mexico. Certain positive aspects of the family environment, such as parental warmth, monitoring, and parent-child relationship quality, were positively associated with narcissism for U.S.-born youth but were not associated or negatively associated with narcissism for Mexican-born youth. In contrast, certain negative aspects of the family environment, such as hostility, held only for Mexican-born youth. Together, these findings suggest that the cultivation of narcissistic tendencies in U.S.-born youth is particularly sensitive to positive influences from the family environment, whereas Mexican-born youth are particularly sensitive to negative influences from the family environment. The extent to which individualistic versus collectivistic values contribute to these differences warrants further investigation. Based on cultural normativeness theory (Lansford et al., 2018), we can speculate that parental hostility may be especially adverse for Mexican-origin youth because they grew up in a culture that places great value on warm and cohesive familial bonds. It is worth noting, however, that the findings were similar for boys and girls and for youth born in the U.S. and Mexico in the vast majority of cases, which supports the generalizability of the findings.

Limitations and Conclusion

The present study has a number of strengths that contribute to a richer understanding of narcissistic development, such as the use of multimethod data to combat shared method variance, the 17-year longitudinal assessment of a variety of unexplored family factors as predictors of narcissism, and the study of an ethnic minority sample that is underrepresented in the field. Nonetheless, it is important to acknowledge the limitations that future research could further investigate. First, vulnerable narcissism was first measured at age 19, whereas other facets were first measured at age 14, spanning different developmental windows. Second, our earliest assessment of family factors occurred at age 10 which covers later but not earlier childhood. As a result, we could not evaluate the extent to which earlier family environments—often thought to play a more formative role in development—impacts later narcissism. Third, we did not have a measure of parental overvaluation, an important theoretical piece to narcissistic growth under the framework of the social learning theory. Including parental overindulgence would provide a more comprehensive picture to understanding the development of narcissism in children.

In summary, our findings suggest that narcissism does not emerge from a single type of family experience, but rather through various avenues, from parenting behaviors to family relationships. Specifically, positive aspects of the family environment such as warm parenting and positive family relationships predicted more adaptive forms of narcissism (superiority), whereas negative aspects of the family environment, such as parental hostility and conflict, predicted more maladaptive forms of narcissism (exploitativeness, vulnerability), thereby reconciling previous conflicting theories regarding the role of parenting on narcissistic development. Certain factors, such as child defiance and agentic parenting goals, predicted both adaptive and maladaptive narcissism. There was little evidence of a critical window during which family influences exert the strongest effects, suggesting that family experiences may have cumulative impacts across childhood and adolescence that shape personality development over time. Our study is among the first to paint a more holistic picture of the prospective effects of the family environment on the development of emerging adult narcissism using long-term longitudinal data, and highlights the importance of both positive and negative family factors in contributing to a more nuanced understanding of the etiology of narcissistic traits.