Adolescent and Parent Perceptions of Legitimacy of Parental Authority: Developmental trajectories and relation to adjustment outcomes

Abstract

Perceptions of the legitimacy of parental authority (LoPA) reflect the degree to which parents and adolescents believe parental control over an adolescent’s behavior is appropriate. LoPA may impact the level of autonomy granted to adolescents and adolescents’ compliance with parental demands and rules. Although low perceived LoPA has consistently been linked to increased externalizing symptoms, the association with internalizing symptoms is less clear. Furthermore, few studies have considered whether and how age-related changes in LoPA are related to adjustment, and implications of LoPA trajectories for mental health symptoms. In a community sample of early adolescents (n = 387; 55% female; 83.1% non-Hispanic White; mean age at T1 = 11.61, SD = 0.55) and parents, latent growth curve modeling was utilized to characterize trajectories of LoPA perceptions and test associations with externalizing, anxious, and depressive symptoms. Parent and adolescent perceptions of LoPA generally declined with age although there were differences across domains. Adolescent perceptions of LoPA were prospectively linked to externalizing, anxious, and depressive symptoms and rate of change was associated with externalizing and depressive symptoms such that lower initial levels and steeper declines were associated with poorer adjustment, while parent perceptions of LoPA were not associated with adjustment outcomes.

Introduction

The development of autonomy is a key task of adolescence, and it has important implications for adjustment (Allen et al., 1994; McElhaney & Allen, 2001). A challenge for parents is to strike a balance between maintaining sufficient control to keep children safe, while not being excessively intrusive as to impede the development of autonomy. The myriad biological and social changes during adolescence requires parents to adapt their parenting to the changing needs of their child. Too much autonomy in early adolescence or too little in later adolescence can lead to adjustment problems (Lamborn et al., 1996; Smetana et al., 2004). In this paper we focus on parent and adolescent beliefs about the legitimacy of parental authority (LoPA) because these beliefs may impact parenting behavior and the degree to which adolescents are responsive to socialization efforts. Hence, LoPA is important for understanding autonomy development and adolescent adjustment. Our goal is to characterize developmental changes in parent and adolescent beliefs about LoPA and how these beliefs impact adjustment (depression, anxiety, and externalizing behavior).

Adolescent LoPA and Adjustment

Beliefs about the LoPA reflect the degree to which parental control over an area or domain of behavior is believed to be appropriate and acceptable. LoPA is commonly conceptualized through the framework of social domain theory, which posits that children develop social knowledge about multiple domains of issues (e.g., moral, social conventional) through interactions with their social environment, and that reasoning and judgments vary across these domains (Smetana, 2006; Turiel, 1983). Domains commonly include moral, conventional, prudential, personal, and multifaceted domains, and issues across these domains are primarily differentiated by their potential consequences to the self and others. The moral domain encompasses issues that affect others’ rights or welfare (e.g., lying to parents, hitting brothers and sisters). The conventional domain relates to issues that are based on social convention (e.g., talking respectfully to parents, cursing). The prudential domain includes issues that pertain to safety or harm to the self (e.g., drinking alcohol, driving recklessly). The personal domain includes issues that have consequences only to the individual (e.g., what clothes to wear, how much time to spend on the computer). Finally, the multifaceted domain includes issues that entail overlap between domains, most commonly the personal and conventional domains (e.g., hanging out with friends parents don’t approve of, when to clean bedroom) (Campione-Barr et al., 2020; Smetana, 2000). Levels of perceived LoPA vary across these domains, with the moral and conventional domains typically considered the most subject to legitimate parental authority, while personal issues are least likely to be considered subject to legitimate parental authority (Smetana et al., 2005).

There is evidence that adolescents’ perceptions of LoPA, both overall and within specific domains, are associated with their adjustment (Campione-Barr et al., 2020; Keijsers & Laird, 2014; Kuhn et al., 2014; Smetana & Asquith, 1994). The link between adolescent perceived LoPA and externalizing symptoms is relatively well-established. Adolescents who perceive their parents as having less overall legitimate authority are more likely to engage in delinquent or antisocial behavior (Cumsille et al., 2009; Laird et al., 2010; Trinkner et al., 2012) and those who perceive their parents as having less legitimate authority over prudential issues are more likely to use or plan to use alcohol and drugs (Handley & Chassin, 2013; Jackson, 2002). These associations are consistent with prior work suggesting that adolescents are more likely to obey parental rules when they believe parents have legitimate authority over the issue in question, even if they disagree with the rule (Darling et al., 2007, 2008). Low adolescent perceived LoPA could arise from poor quality relationship with parents, negative parenting practices (inconsistent or harsh discipline, permissiveness), or an authoritarian parenting style (low parental warmth coupled with high control), although this remains speculative, as no work has directly examined the factors that impact adolescent perceptions of LoPA. These parenting characteristics are associated with poor adjustment outcomes, including internalizing symptoms (Rose et al., 2018), substance use (Pinquart & Lauk, 2025), and emotion dysregulation (Goagoses et al., 2023), and it is possible that these associations may be partially mediated by adolescents’ perceptions of LoPA.

Research on the associations between adolescents’ perceived LoPA and internalizing symptoms is more limited, but there is some evidence that higher adolescent perceived LoPA or lack of declines in LoPA perceptions over time may be linked to increased internalizing symptoms. In a latent class analysis, Wang and Faldowski (2014) found a trend such that adolescents characterized by greater endorsement of overall perceived LoPA exhibited high levels of depressive symptoms, although group differences were not statistically significant. In a longitudinal study, Campione-Barr et al. (2020) found that adolescents’ baseline perceptions of LoPA were not significantly associated with internalizing symptoms three years later; however, declines in LoPA for conventional and personal issues were associated with lower levels of depressive, but not anxious symptoms. These findings indicate the importance of considering change in LoPA and that LoPA may have a different association with depression vs. anxiety. The mechanisms underlying the link between adolescent perceived LoPA and internalizing symptoms are less clear than those for externalizing. One possibility is that sustained high levels of perceived LoPA, particularly at older ages, may derail autonomy development. Age-related increases in adolescent report of autonomy are generally associated with low levels of depression, high levels of self-worth, and adaptive emotional functioning (Qin et al., 2009; Smetana et al., 2004). Additionally, autonomous motivations for agreeing with parental values (i.e., motivated by understanding the benefit of the action, experienced as self-determined) have been associated with greater adolescent well-being than controlled motivations (i.e., motivated by feelings of coercion, experienced as self-discrepant) (Knafo & Assor, 2007). Lack of normative autonomy development, as manifested through high adolescent perceived LoPA or a lack of normative declines in LoPA, may thus be associated with increased internalizing symptoms.

Age-Related Changes in Adolescent LoPA

With age, adolescent perceived LoPA is typically expected to decline as they view more issues as within their own control, and prior work has found that the sharpest decline in overall LoPA is observed in early adolescence (Darling et al., 2008). This decline, however, is not uniform across all domains. Across adolescence, perceived LoPA appears to remain relatively stable for moral and conventional issues, even for older adolescents. Perceived LoPA for prudential and multifaceted issues appears to be more variable with some work indicating declines and others suggesting stability in these domains. The sharpest and most consistent decline in perceived LoPA across adolescence is typically seen in the personal domain (Campione-Barr et al., 2020; Cumsille et al., 2006; Smetana, 2000; Smetana et al., 2015).

This decrease in LoPA closely mirrors expected increases in autonomy and subsequent decreases in reliance on parental input over the course of adolescence. Despite the expected developmental changes in adolescent perceived LoPA, few studies have considered trajectories of adolescent perceived LoPA and their association with adjustment. There is some evidence that decreases in adolescent perceived LoPA are associated with lower levels of depressive symptoms, and that changes in LoPA are not associated with levels of externalizing problems (Campione-Barr et al., 2020). Smetana et al. (2004) examined adolescent perceptions of autonomy and found that high levels of adolescent perceived autonomy regarding personal issues in later adolescence were associated with low levels of depressive symptoms, but high levels of perceived autonomy in early adolescence were associated with high levels of depressive symptoms. Autonomy related to personal issues may reflect low levels of adolescent perceived LoPA. Taken together, this limited literature suggests that developmental patterns of LoPA may help advance our understanding of linkages to adjustment.

Adolescent vs. Parent Perceptions of LoPA

Limited work has considered whether parents’ perceptions of the legitimacy of their authority impact adolescents’ adjustment. Parental perceptions of LoPA could potentially impact adolescent adjustment by way of parenting behavior. Parents’ perceptions of the legitimacy of their own authority could impact family rules, parental monitoring, psychological control, and discipline practices. For example, high parent perceived LoPA could lead to high levels of monitoring which in turn protect against externalizing behavior and depressive symptoms (Racz & McMahon, 2011; Yap et al., 2014), but it could also lead to high levels of psychological control which could increase risk for internalizing and externalizing symptoms (Chyung et al., 2022; Yan et al., 2020). There has been limited research on the impact of parent perceived LoPA on parenting behaviors, although some work suggests that parents with an authoritarian or authoritative style maintain higher levels of perceived LoPA than those with a permissive style (Smetana, 1995). As adolescents age, it is important for parents to provide opportunities for autonomy by modifying their parenting behaviors in a developmentally appropriate manner.

The few studies that have examined adolescent and parent reports of LoPA suggest low cross-reporter associations with parents typically reporting higher perceived LoPA across domains, particularly in the personal domain (Campione-Barr et al., 2020; Smetana, 2000; Smetana & Asquith, 1994). Of the studies that have examined parent perceptions of LoPA and adolescent adjustment, the findings have been mixed. Some research suggests that high levels of parent reported LoPA are associated with increased adolescent conduct problems (Wang & Faldowski, 2014). However, in their three-year longitudinal study, Campion-Barr et al., (2020) found that high levels of baseline parent perceived LoPA and decreases over time in parent perceived LoPA regarding conventional, personal, and multifaceted issues, were associated with lower levels of adolescent problem behavior (i.e., depressive and externalizing, but not anxiety symptoms). In sum, the role of parent perceptions of the legitimacy of their own authority is unclear and warrants further investigation.

Summary and Hypotheses

While prior work has shown that LoPA is associated with adjustment outcomes, there are notable gaps in this research regarding the developmental course of LoPA and the potential influence of parent perceptions of LoPA. The current study addresses these gaps utilizing a longitudinal data set. First, growth trajectories for both parent and adolescent perceived LoPA were examined. Decreases in overall LoPA were hypothesized with individual differences in the rates of change. Differences in change across LoPA domains were also hypothesized. The steepest declines were expected to occur in the personal domain. Second, the study examined how adolescent and parent perceptions of LoPA relate to adolescent internalizing and externalizing symptoms. Given potential differences in the relation between LoPA and anxiety vs. depressive symptoms (Campione-Barr et al., 2020), the current study examined each domain of internalizing separately. For both parents and adolescents, it was hypothesized that a steeper than typical decline in LoPA would be associated with increased adolescent externalizing symptoms. Stable levels of LoPA might disrupt normative autonomy development and hence were expected to be associated with increased adolescent internalizing symptoms. No hypotheses were offered regarding the differences in associations with anxiety and depression given the limited literature. These proposed relationships were examined for overall LoPA as well as for domain-specific LoPA, given that developmental patterns vary across domains.

Methods

Participants

The sample for this study was taken from a large longitudinal study that examined risk and protective factors for substance use initiation (Colder et al., 2013; Trucco et al., 2014). Recruitment for this study occurred between 2007 and 2009, and data were collected across 9 annual waves. The current study used data from the first three assessments (Waves 1, 2, and 3) when LoPA was assessed. Participants were recruited using random-digit-dialing of telephone numbers in Erie County, New York. Families were eligible to participate if they had a child between the ages of 10 and 12, were fluent in English, and had a caregiver willing to participate.

The final sample included 387 adolescent-parent dyads with the following racial and ethnic composition: 83.1% Non-Hispanic White, 9.1% African American, 2.1% Hispanic, 1.0% Asian, and 4.7% racially/ethnically mixed. 55% of adolescents were female and 45% were male. Median income was $70,000 (which is equivalent to approximately $105,000 in 2025, adjusting for inflation) with a range of $1,500 to $500,000. 76% of adolescents resided in two-parent households, 12.1% in divorced/separated households, 9.8% in single caregiver households, and 2.1% in other arrangements. 87.6% of caregivers who participated were female and 12.4% were male. 97.9% of caregivers who participated were biological or adoptive parents and 2.1% had another relationship to the adolescent (stepparent, grandparent, or other family member). Mean age for adolescents was 11.61 (SD 0.55), 12.63 (SD 0.58), and 13.61 (SD 0.57) at T1, T2, and T3 respectively. Attrition was low (7.5% total) across T1-T3, and dyads who did not complete T2 and/or T3 assessments did not significantly differ in demographics (sex, race/ethnicity, education, income, and marital status) from those who completed all three assessments.

Procedure

Adolescent-parent dyads completed annual interviews conducted at university research offices. Follow-ups took place approximately one year after the preceding assessment. All parents and adolescents provided informed consent and assent, respectively. For each session, parents and adolescents separately completed laboratory tasks and computer administered questionnaires assessing a variety of family, peer, and individual development factors. Questionnaires were read aloud to adolescents by research assistants to ensure comprehension, and answers were entered directly into a computer by the adolescent. Only questionnaire data were considered for this study. Participants were compensated $75, $85, and $100 for their participation at T1, T2, and T3 respectively. All procedures were reviewed and approved by the University at Buffalo IRB (Protocol #806056–1).

Measures

Legitimacy of Parental Authority (LoPA)

Parents’ and adolescents’ perceptions of LoPA were measured using 21-items taken from several existing LoPA measures (Darling et al., 2006; Smetana, 1988; Smetana & Bitz, 1996) to increase the number of items per domain (moral, conventional, prudential, personal, multifaceted) for broader representation and to enhance psychometric properties of domain scale scores. For each item, parents and adolescents reported their level of agreement with the statement “It’s ok for me/my parents to make rules about:” on a 5-point Likert scale (1 = Strongly Disagree, 5 = Strongly Agree). As LoPA items were compiled from multiple sources, Exploratory Factor Analyses (EFAs) were conducted using T1, T2, and T3 data to assess domains or subscales of LoPA and to guide creation of scale scores. Results of the EFAs are reported below.

Internalizing symptoms

Adolescents reported on their internalizing symptoms using DSM-5-Oriented Scales for Depressive Problems (11 items) and Anxiety Problems (7 items) for the Youth Self-Report (YSR) for ages 11–18 (Achenbach & Rescorla, 2001; Achenbach, 2013). Adolescents reported how well each item described them now or within the past 6 months on a scale of 0 (Not true) to 2 (Very true or often true). Items demonstrated strong internal consistency using ordinal alphas for both the Depressive Problems (α = .87, .92, and .89 at T1, T2, and T3 respectively) and Anxiety Problems (α = .80, .86, and .88) subscales.

Externalizing symptoms

Adolescents reported on their externalizing symptoms using the externalizing subscale (32 items) of the YSR for ages 11–18 (Achenbach & Rescorla, 2001). Adolescents reported how well each item described them now or within the past 6 months on a scale of 0 (Not true) to 2 (Very true or often true). Items demonstrated strong internal consistency using ordinal alphas; α = .94, .95, and .95.

Covariates

Adolescent sex (0 = male, 1 = female), parental education as measured by highest level of education completed by the participating caregiver (1 = Grade school, 7 = Graduate/Professional school), and receipt of public assistance (no/yes) were included as covariates.

Data Analytic Plan

Factor Analysis of LoPA

Exploratory Factor Analysis (EFA) was used to determine the number of factors that best represented the domains of LoPA. First, a Principal Components Analysis (PCA) was conducted to examine eigenvalues including parallel analysis to determine the number of factors to extract. Principal Factor Analysis (PFA) was then used to determine the most interpretable solution. Separate EFAs were conducted for T1, T2, and T3, and for adolescent and parent report. This allowed examination of the consistency of the factor structure across time and reporter. Summary of the results are presented below and details of the EFA can be found in supplemental materials.

Growth Modeling

Growth models were estimated in Mplus 8.4 (Muthén & Muthén, 1998–2017) using Robust Full-Information Maximum Likelihood estimation (MLR) to accommodate missing data and non-normality. Some skew was observed in adolescent and parent perceived prudential/delinquent LoPA and for the adjustment outcome variables. Missing data for any given variable ranged from 0.3% to 4.7%. Chi-square and analysis of variance tests comparing those with and without missing data suggested no significant differences for demographics (age, sex, race, parental education, public assistance), predictors, or outcome variables. The low attrition rate and lack of differences along with using MLR which uses full-information estimation suggests that missing data did not have a substantial impact on the findings of the current study. First, unconditional growth in overall LoPA for both adolescents and parents was tested, then separate unconditional growth models were estimated for each identified LoPA domain for both adolescents and parents.

Nested chi-square tests were used to compare an intercept only model to a linear growth model to test whether there was significant change in LoPA perceptions over time. The intercept represented T1 LoPA levels and slope loadings were set to 0, 1, and 2. A combination of nested chi square tests as well as Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), and Root Mean Square Error Approximation (RMSEA) were used to establish the best fitting growth models. Acceptable fit ranges for each index are as follows: CFI and TLI: < .90 is poor, .90–.94 is acceptable, > .95 is good; RMSEA: > .08 is poor, .05–.07 is acceptable, < .05 is good (Hu & Bentler, 1999; Marsh et al., 2004).

After determining the best fitting unconditional growth model, the growth factors were included in a structural equation model with covariates (sex, SES indices), and outcomes specified as observed variables (anxiety, depressive, and externalizing symptoms) controlling for baseline levels of each outcome. Figure 1 depicts the conditional growth model tested. Separate models were estimated for overall LoPA and for each identified LoPA factor for both adolescents and parents to reduce model complexity. Model chi square, CFI, TLI, and RMSEA were used to assess model fit.

Figure 1. Proposed Path Model.

Notes. T1 = time one, T3 = time three, EXT = externalizing, INT = internalizing, LoPA = legitimacy of parental authority. Although not depicted in the figure, sex and SES (as measured by parental education and receipt of public assistance) were included as covariates.

Results

Factor Analysis of LoPA

Adolescent Report of LoPA

PCA analysis of eigenvalues and parallel analysis suggested a 2 or 3 factor model, but there was considerable cross-loading. After running the PFA and trimming cross-loading items, the final model included 3 factors at each wave: personal issues (8 items), moral/conventional issues (5 items), and prudential/delinquent issues (3 items). A more detailed summary of the factor analysis and factor loadings can be found in supplemental materials. Factors demonstrated acceptable internal consistency at all waves: personal issues αs = .82, .85, .89; moral/conventional issues αs = .78, .79, .80; prudential/delinquent issues αs = .85, .85, .89. Items were combined to form scale scores by taking the average of items within each factor at each wave.

Parent Report of LoPA

Analysis of eigenvalues from PCA and parallel analysis of the parent report LoPA items indicated a three-factor solution was most appropriate, but again there were some cross-loading items. After eliminating cross-loading times, the final PFA factor solution included three factors similar to those found in the solution for adolescent reported item: personal issues (7 items), moral/conventional issues (5 items), and prudential/delinquent issues (2 items) (see supplement for more detail of EFA and factor loadings). Factors demonstrated acceptable internal consistency at all waves: personal issues αs = .79, .81, .82; moral/conventional issues αs = .78, .76, .81; prudential/delinquent issues αs = .94, .94, .94. Items were averaged within factor to form scale scores at each wave.

In sum, final factor solutions included 16 items for adolescent report and 14 items for parent report. Of the items included in both the parent and adolescent solutions, only one item (cleaning room) loaded on a different factor for parents (personal) vs. adolescents (moral/conventional).

Unconditional Growth Models

The linear model provided an adequate fit to the data for overall LoPA, and means for the growth factors for both parents and adolescents were statistically significant (see Table 1). There was significant variance in the intercept in both parent and adolescent models, but the slope variance was statistically significant in the adolescent model only. The findings suggest a decline in parent and adolescent reported overall LoPA with significant individual variability in overall LoPA at T1 and significant individual differences in rate of change in LoPA for adolescent report only.

Table 1.

Results for Unconditional Growth Models

	Parent		Adolescent
Overall LoPA	Mean	Variance	Mean	Variance
Intercept	4.37***	0.11***	3.87***	0.20***
Slope	−0.02*	0.01	−0.11***	0.03**
Covariance	−0.003		−0.004
Chi Square	χ2(1) = 0.40, p = .53		χ2(1) = 0.63, p = .43
CFI; RMSEA	1.00; 0.00		1.00; 0.00
Personal LoPA	Mean	Variance	Mean	Variance
Intercept	4.02***	0.25***	3.43***	0.35***
Slope	−0.04**	0.02	−0.18***	0.07**
Covariance	−0.014		−0.003
Chi Square	χ2(1) = 0.05, p = .83		χ2(1) = 1.13, p = .29
CFI; RMSEA	1.00; 0.00		1.00; 0.02
Moral/Conventional LoPA	Mean	Variance	Mean	Variance
Intercept	4.65***	0.06***	4.07***	0.25***
Slope	0.01	−0.02	−0.04	0.04*
Covariance	0.013		−0.029
Chi Square	χ2(1) = 2.55, p = .11		χ2(1) = 0.02, p = .90
CFI; RMSEA	0.99; 0.06		1.00; 0.00
Prudential/Delinquent LoPA	Mean	Variance	Mean	Variance
Intercept	4.88***	0.05**	4.72***	0.10**
Slope	−0.01	0.01	−0.07***	0.001
Covariance	−0.007		0.011
Chi Square	χ2(1) = 0.01, p = .93		χ2(1) = 0.23, p = .63
CFI; RMSEA	1.00; 0.00		1.00; 0.00

Notes. T1 = time one, T3 = time three, LoPA = legitimacy of parental authority. Covariance reported for intercept and slope. Good fit indicated by non-significant Chi Square value, CFI > 0.95, TLI > 0.95, RMSEA < 0.05.

^*= p < .05,

^**= p < .01,

^***= p < .001. N = 387.

Models for each of the specific LoPA domains (personal, moral/conventional, prudential/delinquent) suggested adequate fit for the linear growth model and showed results similar to overall LoPA (see Table 1). Intercept means and variances were statistically significant for parents and adolescents in all models. For the personal domain, the slope mean was significant for both parent and adolescent, but only adolescent slope variance was statistically significant. For the moral/conventional domain, the slope mean was not significant for either the parent or adolescent model, but there was significant slope variance for the adolescent model only. Finally, for the prudential/delinquent domain, the slope mean was significant only for the adolescent model, and there was no significant slope variance for either model. These findings suggest a decline in parent and adolescent reports of LoPA for personal issues, no change over time for parent or adolescent reports for the moral/conventional domain, and a decline for only adolescent report of prudential/delinquent domain. Additionally, there was significant individual variability in T1 LoPA for all domains for both parents and adolescents, but significant individual differences in rate of change was observed only for adolescent reports of the personal and moral/conventional domains.

Across all models, the covariance between the intercept and slope was not statistically significant. In all models, adolescents reported lower mean intercepts and steeper mean declines in LoPA than parents. Consistent with hypotheses and prior work, personal LoPA demonstrated the lowest intercept and steepest decline for both parents and adolescents.

Conditional Growth Models

Unstandardized coefficients for parent and adolescent structural equation models for overall, personal, moral/conventional, and prudential/delinquent LoPA are presented in Table 2. The slope variance was set to zero where unconditional models showed no slope variance (e.g., parent personal, moral/conventional, and prudential/delinquent models, adolescent prudential/delinquent model), and for these models the slope was not included as a predictor of the symptom outcomes. All models demonstrated good or acceptable fit across at least three of four fit indices (Table 3).

Table 2.

Results for Conditional Growth Models

	Endogenous Variables
	Parent				Adolescent
Overall LoPA
Predictors	Intercept	T3 ext	T3 anx	T3 dep	Intercept	Slope	T3 ext	T3 anx	T3 dep
Child Sex	0.05	0.01	0.06	0.04*	0.10	0.02	0.04*	0.08**	0.06**
Parent Education	0.02	−0.03***	−0.04**	−0.02**	−0.01	0.02	−0.02**	−0.03**	−0.01
Public Assistance	0.06	0.07*	−0.02	0.00	−0.03	−0.08	0.03	−0.04	−0.03
T1 ext	0.09	0.59***	-	-	−0.93***	−0.06	0.49***	-	-
T1 anx	−0.12	-	0.50***	-	0.07	−0.04	-	0.45***	-
T1 dep	0.04	-	-	0.47***	−0.20	0.06	-	-	0.45***
Intercept	-	0.01	0.04	0.04	-	-	−0.14***	−0.12**	−0.08**
Slope	-	-	-	-	-	-	−0.41***	−0.13	−0.34**
R 2	0.02	0.32***	0.22***	0.26***	0.18***	0.04	0.54***	0.27***	0.39***
Personal LoPA
Predictors	Intercept	T3 ext	T3 anx	T3 dep	Intercept	Slope	T3 ext	T3 anx	T3 dep
Child Sex	0.08	0.01	0.06	0.04*	0.12	0.02	0.03	0.08**	0.06**
Parent Education	0.02	−0.03***	−0.04**	−0.02*	−0.03	0.02	−0.03***	−0.04**	−0.02*
Public Assistance	0.14	0.07*	−0.02	0.00	0.04	−0.21*	0.02	−0.02	−0.03
T1 ext	0.15	0.59***	-	-	−1.07**	−0.08	0.50***	-	-
T1 anx	−0.13	-	0.46***	-	0.03	−0.00	-	0.43***	-
T1 dep	0.03	-	-	0.47***	−0.27	0.09	-	-	0.47***
Intercept	-	−0.00	0.02	0.03	-	-	−0.11***	−0.11***	−0.05*
Slope	-	-	-	-	-	-	−0.23***	0.00	−0.17**
R 2	0.02	0.32***	0.22***	0.26***	0.16**	0.06	0.50***	0.28***	0.34***
Moral/Conventional LoPA
Predictors	Intercept	T3 ext	T3 anx	T3 dep	Intercept	Slope	T3 ext	T3 anx	T3 dep
Child Sex	0.01	0.01	0.06*	0.04*	0.12	0.01	0.03	0.08**	0.06**
Parent Education	0.03*	−0.03***	−0.04**	−0.02**	0.00	0.02	−0.02**	−0.03**	−0.01
Public Assistance	0.02	0.07	−0.02	0.00	−0.04	0.01	0.01	−0.02	0.00
T1 ext	0.08	0.59***	-	-	−0.86**	−0.02	0.51***	-	-
T1 anx	−0.15*	-	0.46***	-	0.03	−0.05	-	0.45***	-
T1 dep	0.08	-	-	0.47***	−0.17	−0.03	-	-	0.43***
Intercept	-	0.04	0.09	0.06	-	-	−0.11***	−0.09*	−0.10***
Slope	-	-	-	-	-	-	−0.22*	−0.22	−0.31*
R 2	0.04	0.32***	0.23***	0.26***	0.14**	0.02	0.47***	0.27***	0.39***
Prudential/Delinquent LoPA
Predictors	Intercept	T3 ext	T3 anx	T3 dep	Intercept		T3 ext	T3 anx	T3 dep
Child Sex	0.02	0.01	0.06*	0.04*	0.05		0.02	0.06*	0.05**
Parent Education	0.01	−0.03***	−0.04**	−0.02*	0.04*		−0.02**	−0.03**	−0.01
Public Assistance	−0.07	0.07*	−0.02	0.01	−0.08		0.06	−0.02	−0.01
T1 ext	−0.08	0.59***	-	-	−0.81***		0.55***	-	-
T1 anx	−0.03	-	0.46***	-	0.12		-	0.45***	-
T1 dep	0.03	-	-	0.47***	0.06		-	-	0.44***
Intercept	-	0.04	−0.00	0.03	-		−0.10*	−0.04	−0.12**
R 2	0.02	0.32***	0.22***	0.26***	0.19**		0.36***	0.22***	0.29***

Notes. T1 = time one, T3 = time three, LoPA = legitimacy of parental authority, ext = externalizing, anx = anxiety, dep = depression. Slope is not included as a predictor or endogenous variable for models where there was no significant slope variance.

^*= p < .05,

^**= p < .01,

^***= p < .001. N = 387.

Table 3.

Model Fit Indices for Conditional Growth Models.

	Chi Square	CFI	TLI	RMSEA
Parent Overall	χ2(27) = 42.44, p = .03	0.99	0.97	0.04
Adolescent Overall	χ2(16) = 14.52, p = .56	1.00	1.00	0.00
Parent Personal	χ2(27) = 50.22, p < .01	0.98	0.96	0.05
Adolescent Personal	χ2(16) = 17.32, p = .37	1.00	1.00	0.02
Parent Moral/Conventional	χ2(27) = 45.80, p = .01	0.98	0.95	0.04
Adolescent Moral/Conventional	χ2(16) = 14.37, p = .57	1.00	1.00	0.00
Parent Prudential/Delinquent	χ2(27) = 27.10, p = .46	1.00	1.00	0.00
Adolescent Prudential/Delinquent	χ2(27) = 49.89, p < .01	0.97	0.94	0.05

Note. Good fit indicated by non-significant (p > .05) Chi Square value, CFI > 0.95, TLI > 0.95, RMSEA < 0.05.

For overall LoPA, the parent intercept was not associated with T3 externalizing, anxious, or depressive symptoms. For the adolescent model, the intercept was significantly negatively associated with T3 externalizing, anxious, and depressive symptoms, and the slope was negatively associated with T3 externalizing and depressive (but not anxious) symptoms. Low initial levels of adolescent reported overall LoPA were associated with high levels of problem behavior three years later, and steeper declines in overall LoPA than average were associated high levels of externalizing and depressive symptoms.

For both personal and moral/conventional LoPA, the intercept for parents was not associated with T3 externalizing, anxious, or depressive symptoms. For the adolescent model, the intercept was significantly negatively associated with T3 externalizing, anxious, and depressive symptoms, and the slope was negatively associated with T3 externalizing and depressive (but not anxious) symptoms. Low initial levels of adolescent reported personal and moral/conventional LoPA were associated with high levels of problem behavior three years later, and steeper declines than average were associated high levels of externalizing and depressive symptoms.

For prudential/delinquent LoPA, the intercept for parents was not associated with T3 externalizing, anxious, or depressive symptoms. For the adolescent model, the intercept was significantly negatively associated with T3 externalizing and depressive (but not anxious) symptoms. Low initial levels of adolescent reported prudential/delinquent LoPA were associated with high levels of externalizing and depressive symptoms three years later.

We expected links between LoPA in prudential/delinquent domain and externalizing symptoms, but the relation between prudential/delinquent issues and depressive symptoms was surprising. Externalizing and depressive symptoms were highly correlated (r = .64) suggesting that associations with depressive symptoms may be confounded with externalizing symptoms. Accordingly, additional post hoc analyses were conducted predicting T3 depressive symptoms while controlling for T3 externalizing symptoms and vice versa. Controlling for T3 externalizing resulted in a non-significant association between adolescent perceived LoPA prudential/delinquent intercept and depressive symptoms (β = .01, p = .85), but the association with externalizing symptoms remained significant (β = −.09, p = .04) when controlling for T3 depressive symptoms. This indicates that the effect of prudential/delinquent LoPA on depressive symptoms was attributable to externalizing symptoms.

Sensitivity Analysis

To determine whether significant results for adolescent LoPA only were due to shared reporters across the independent variable and outcomes, conditional growth models were run predicting parent report of adolescent adjustment outcomes (using the Child Behavior Checklist). Correlations across reporters for adjustment outcomes were .40, .22, and .18 for T3 externalizing, anxious, and depressive symptoms, respectively. Results for these models are presented in Table S1 and all models demonstrated acceptable fit (Table S2). Neither parent nor adolescent report LoPA were associated with parent report of T3 externalizing, anxious, or depressive symptoms.

To address potential impacts of differences in LoPA items used to form domain scores across reporters, additional sensitivity analyses were conducted using the same items for parent and adolescent models. Results were largely the same, with the exception of a significant positive association between parent overall and moral/conventional LoPA and T3 depressive symptoms, suggesting different pools of items did not have a large impact on the findings. Further details of these analyses are presented in the supplemental materials.

Discussion

Age-Related Changes in LoPA

Consistent with prior research, growth models indicated age-related changes in LoPA perceptions for both parents and adolescents. These changes also varied across domains, with significant decreases in perceived LoPA in the personal domain, significant decreases for the prudential/delinquent domain but only for adolescent reported items, and no significant change in the moral/conventional domain. This builds on work examining adolescent perceptions that has found stability in moral/conventional issues and decreases in personal issues and helps to clarify the mixed literature on changes over time in prudential issues, as prior work looking at developmental changes in prudential LoPA perceptions has found decreases, no change, and even increases with age (Campione-Barr et al., 2020; Darling et al., 2008; Smetana, 2000; Smetana et al., 2015).

While these trajectories are somewhat well established for adolescents, research on changes over time for parent perceived LoPA is much more limited. Campione-Barr et al. (2020) assessed both parent and adolescent perceptions over time but focused primarily on changes in discrepancies between parent and adolescent. They found a significant increase in LoPA discrepancy, which appeared to be driven by decreases in adolescent perceptions and relative stability in parent perceptions. This is consistent with our findings that parent perceptions of overall LoPA tended to show less steep decline than adolescent perceptions. Additionally, while there was individual variation in the rate of change for adolescents in perceptions of overall, personal, and moral/conventional LoPA, there were no individual differences in the rate of change for parents. This suggests that in this sample LoPA change based on parent report is more uniform than LoPA change based on adolescent report.

LoPA and Adjustment

Consistent with prior research, adolescents’ perceptions of LoPA were associated with externalizing, depressive, and anxiety symptoms. Contrary to hypotheses that there would be differential impacts of LoPA on externalizing vs. internalizing symptoms, whenever there was a significant relationship between adolescent perceived LoPA and adjustment, lower T1 levels of adolescent perceived LoPA and steeper decreases over time were associated with poorer adjustment outcomes.

This finding differs somewhat from the findings of Campione-Barr et al. (2020), who found no association between adolescent LoPA intercepts and adjustment, and that decreases in adolescent perceived LoPA over time were associated with fewer depressive symptoms. The sample in that study, however, had more age variability and followed up with adolescents over a four-year period. It is possible that the relationship between adolescent perceived LoPA and depressive symptoms changes over time, with decreases in earlier adolescence (as in the current sample) being associated with increased depressive symptoms but decreases later in adolescence (as in the Campione-Barr et al. (2020) sample) being associated with decreased depressive symptoms. Declines in adolescent perceived LoPA may be more adaptive during later adolescence as higher levels of autonomy become more normative and developmentally appropriate.

The relationship between adolescent LoPA and outcomes differed between anxious vs. depressive symptoms, with adolescent perceived LoPA more consistently related to depressive symptoms. This is somewhat consistent with the findings of Campione-Barr et al. (2020) and provides support for future research to examine domains of internalizing separately. The more consistent relationship between adolescent perceived LoPA and depression could potentially be due to adolescents with low LoPA perceptions experiencing depressive symptoms as a result of feeling as though their autonomy is unfairly constrained. This is consistent with learned helplessness theory that posits a lack of perceived control is linked to depressive symptoms, which has been supported by findings in adolescents (Brown & Siegel, 1988; Ye et al., 2024). Additionally, adolescents with steeper declines in LoPA likely experience increasing discrepancies between their LoPA perceptions and those of their parents, potentially leading to increased conflict and tension, which may result in depressive symptoms.

In contrast to adolescent perceptions, parent perceptions of LoPA were not related to any adolescent adjustment outcomes. This is contrary to Campione-Barr and colleagues’ (2020) findings that both intercepts and slopes of parent LoPA perceptions were associated with depressive, anxious, and externalizing symptoms. These associations varied across birth order with parent perceived LoPA in some domains at certain timepoints associated with adjustment only for first- or second-born adolescents, indicating that this may be an important moderator to examine further in future research. Prior work has found that parents perceive themselves as granting more decision-making autonomy to second-born adolescents than first-born adolescents in some domains, potentially as a result of learning from experience and adjusting parenting behaviors with second-born children (Campione-Barr et al., 2015; Wray-Lake et al., 2010). It will be important for future work to replicate these complex interactions.

Sensitivity analyses utilizing a common pool of LoPA items across parents and adolescents resulted in a significant positive association between overall and moral/conventional parent perceived LoPA and adolescent reported depressive symptoms. This could potentially result from parents implementing more restrictive parenting as a result of their high levels of perceived LoPA, which could in turn negatively impact adolescent adjustment. To further understand this pattern of findings in future research, it may also be useful to consider if and how parent LoPA perceptions influence parenting behavior. These perceptions may only impact adolescent adjustment to the extent that they tangibly alter parenting behavior, such as psychological control, unilateral decision making, and harsh parenting. The general lack of associations between parent perceived LoPA and adolescent adjustment is consistent with a broader literature that suggests that adolescents’ perceptions of parenting behavior may be stronger predictors of adjustment than parent perceptions (e.g., Abar et al., 2015; Cottrell et al., 2003; Maurizi et al., 2012).

Broadly speaking, these findings indicate that lower levels of adolescent perceived LoPA are associated with poorer adjustment outcomes during early adolescence. Perceived LoPA could be an important target for intervention to address concerns related to both externalizing and internalizing symptoms. Given these results, interventions aimed at increasing adolescent perceived LoPA may be beneficial to adolescent well-being. Prior work has found that adolescent perceived LoPA is associated with parental monitoring and support (Darling et al., 2008), which may indicate that targeting these specific parenting behaviors could lead to changes in adolescent perceived LoPA.

Factor Analysis of LoPA

This study is one of few that has examined the dimensionality of LoPA items using factor analysis. For both parents and adolescents, a three-factor solution was determined to be the most interpretable, with factors that correspond to the personal, moral/conventional, and prudential domains identified in prior literature. Although moral and conventional issues have primarily been measured separately due to their theoretical distinction, previous work has shown that both types of issues tend to be rated very high by both parents and adolescents on LoPA questionnaires (Smetana, 2000; Smetana & Asquith, 1994). Additionally, moral and conventional issues are conceptually similar in that they are both socially regulated (vs. affecting the individual only; Smetana, 2006). This has led some researchers to combine these domains into one scale, which was supported by a CFA (Smetana et al., 2015) and is consistent with our finding that moral and conventional items loaded onto the same factor.

Results of our factor analysis suggested that the LoPA domains were represented by a slightly different sets of items across reporter. Nonetheless, there was sufficient overlap in parent and adolescent items to warrant similar interpretation of the factors based on prior theoretical categorizations. That the factor solutions differed across report may indicate that different issues are more representative of those domains for parents vs. adolescents. This is not surprising, given the low correlations across parent and adolescent LoPA perceptions (ranging from .07 to .14 within wave), which indicate that these are distinct perspectives and may therefore be better captured by different items for each reporter. Low parent-adolescent correlations have reported in other studies as well (Smetana, 2000). As noted, sensitivity analyses examining results using common LoPA items for both reporters produced largely the same results, with the exception of a significant positive association between parent overall and prudential/delinquent LoPA intercepts and T3 adolescent reported depressive symptoms.

One limitation of the methodology used for this EFA is that retaining items that loaded onto only one factor may have removed items that would be considered multifaceted. Yet, even if we tolerated some cross-loading, we did not see evidence for a multifaceted dimension. One study that utilized factor analysis on similar items found some evidence for minor cross-loading of items considered multifaceted or “overlapping”; however there wasn’t strong evidence for factorial complexity, as these items formed their own distinct multifaceted or “overlapping” domain factor separate from prudential/conventional and personal domain factors (Hasebe et al., 2004). Future research may benefit from considering how better to include items that are conceptualized as multifaceted.

Limitations and Future Directions

The results of this study help to further illuminate the developmental changes in adolescent and parent perceptions of LoPA and the association between LoPA and adolescent adjustment outcomes. There are, however, still many unanswered questions that would be beneficial to examine in future research. The age range of this sample was limited to early and middle adolescence, with adjustment outcomes measured at approximately ages 13–14. As such, effects of LoPA perceptions on later adjustment outcomes could not be examined. Future research would benefit from examining outcomes over a longer follow-up period, as changes in adolescent autonomy development and the role of LoPA perceptions may differ in older adolescence. While there were significant decreases in LoPA for adolescents over the three waves, adolescents still tended to agree that their parents did have at least some legitimate authority to make rules about various issues, suggesting the importance of continued parental control.

The current sample included adolescents from the United States that were predominately White, and generally highly educated and affluent, limiting the generalizability of these findings across different racial/ethnic and cultural backgrounds and socioeconomic statuses. Prior work examining LoPA perceptions has been conducted in different populations, including African American adolescents in the US (Smetana, 2000) as well as adolescents in South America (Cumsille et al., 2006, 2009; Darling et al., 2008), Asia (Hasebe et al., 2004), and the Middle East (Smetana et al., 2015). Based on this work, developmental trajectories of adolescent LoPA perceptions appear to be relatively similar across cultures, although there may be subtle differences in perceptions about specific issues (e.g., dating; Hasebe et al., 2004). Although LoPA perceptions appear similar across cultures, the associations between LoPA perceptions and adjustment outcomes may differ across cultural contexts and future work would benefit from examining these associations in more diverse samples.

Additionally, while parent perceptions were unrelated to adolescent adjustment outcomes in this sample, it is possible that the relation between adolescent and parent perceptions may be meaningful above and beyond individual reports. The trajectories of change over time found in this study indicate that discrepancies between adolescent and parent perceptions increase as adolescents age, with parent perceptions declining less rapidly, if at all. Examining discrepancies may yield additional insight into whether and how parent LoPA perceptions relate to adolescent adjustment. There are significant limitations associated with using discrepancy scores, however, and future work should attempt to account for these limitations, potentially utilizing polynomial regression analysis (Laird & De Los Reyes, 2013) or latent difference scores (de Haan et al., 2018).

Data Availability:

The data that support the findings of this study are openly available in ICPSR at https://www.icpsr.umich.edu/web/ICPSR/studies/37620, reference number ICPSR 37620.