Associations between Developmental Patterns of Negative Parenting and Emotion Regulation Development Across Adolescence

Abstract

Research has documented changes in parenting practices and in emotion regulation (ER) during adolescence. However, developmental trajectories of these constructs and how they may be linked are not clearly known. The present study examined longitudinal associations between developmental trajectories of negative parenting and developmental trajectories of ER (e.g., abilities and strategy use, including cognitive reappraisal and suppression). The sample included 167 adolescents (53% males) who were first recruited at age 13 or 14 years and assessed annually four times. Adolescents self-reported on the perceived degree of their parent’s negative parenting and ER. Growth mixture modeling revealed two distinct trajectories of negative parenting across adolescence: Class 1 contained the majority of adolescents (84%), with moderate initial levels of negative parenting that decreased across adolescence; Class 2 contained a smaller group of adolescents (16%), reporting moderate initial levels of negative parenting that increased across adolescence. Though growth curve modeling did not reveal significant growth in ER across time in the sample as a whole, results from a two-group model demonstrated that ER development significantly differs depending on adolescents’ experiences of negative parenting trajectories. Adolescents experiencing decreases in negative parenting showed significant increases in ER abilities and no significant changes in suppression. Adolescents experiencing increases in negative parenting exhibited significant decreases in ER abilities. Adolescent’s cognitive reappraisal was unaffected by negative parenting. The findings underscore the significant role of differential parenting environments in the development of ER abilities during adolescence.

Emotion regulation (ER) is an important developmental achievement, critical for health and well-being (Cole, et al., 2017). Developmental research has suggested that difficulties with ER emerge from an interaction between individual factors (e.g., temperamental and biological factors) and environmental factors (e.g., parenting factors; Goldsmith et al., 2008). Indeed, a number of studies have emphasized associations between child ER and parenting behavior, such that positive parenting (encompassed by warmth, support, and emotional safety) is facilitative of adolescents’ ER and its development, whereas negative parenting (reflected by criticism, harshness, and psychological control) may have a dampening effect (see Kiel & Kalomiris, 2015; Morris et al., 2017 for reviews). Yet, research examining associations between parenting risk factors (e.g., negative parenting) and ER development during adolescence—which may be a critical developmental period for ER development—is limited.

Prior theoretical and empirical work has emphasized that the number of biological, cognitive, and social changes that take place during adolescence make examination of not only ER development noteworthy (Steinberg, 2005), but also developmental changes within parenting factors (Ebbert et al., 2018). That is, numerous novel contexts and experiences during adolescence in addition to decreased opportunities for co-regulation with parents, require more advanced ER abilities that may be aided by increased neurobiological maturation, especially in prefrontal regions thought to underlie regulatory abilities (McRae et al., 2012; Steinberg, 2005). Likewise, the challenge of parenting adolescents coupled with the reorganization of the parent-child relationship during this period is likely to increase parental stress and may produce changes in parenting patterns (Steinberg & Silk, 2002). Yet, we do not have a clear understanding as to whether parenting behaviors change dynamically during adolescence, and if so, how such change patterns may be associated with ER development. To address this gap in the literature, the present study used prospective longitudinal data to examine patterns of within-person changes of negative parenting and their associations with developmental trajectories of ER.

Developmental Changes in Negative Parenting

In general, research has found that while attachment relations formed in early childhood tend to remain relatively stable across adolescence (Nickerson & Nagle, 2005), cognitive and physical changes stemming from puberty necessitate a reorganization of the parent-child relationship during adolescence. Previous research has suggested that adolescence is perceived by parents as a challenging stage of child-rearing, noting increased parental stress experienced during the transition into adolescence (Eisenberg et al., 2008). Indeed, the stress of the adjustment may weigh heavier on parents than on adolescents as the parent of an adolescent is faced with developmental changes that can make the task of parenting noticeably more challenging. This is in part due to the fact that many middle-aged adults tend to experience identity concerns as well as a low point in marriage and life satisfaction, which make it stressful to cope with their adolescents’ independence from and deidolization of them (Steinberg & Silk, 2002).

Though extant theoretical work has proposed probable shifts in parenting across adolescence, little is known about the developmental course of parenting behaviors that may emerge from the stress associated with the adolescent period. Some parents may cope with the challenges associated with the transition to adolescence—including parenting stress, adolescents’ autonomy assertions, and hostility and conflict—with increases in negative parenting (e.g., nagging, criticism, and yelling). Evidence demonstrates that adolescents perceive greater negative affect and lesser positive affect from their parents during adolescence compared to childhood (Eisenberg et al., 2008; Loeber et al., 2000). In turn, parents who perceive their children as more difficult are more likely to respond with hostile or negative parenting practices (Pettit et al., 2001; Weymouth & Buehler, 2016). Indeed, a meta-analysis demonstrated that negative parenting behaviors are often reactive types of parenting which are provoked by negative affective experiences such as frustration and irritation (Rueger, et al., 2011). Moreover, associations between parent and adolescent negative affect and between parent-adolescent hostility and intrusive parenting have been observed (Kim et al., 2001; Weymouth & Buehler, 2016).

Yet, previous research examining changes in negative parenting during adolescence has revealed mixed results. For example, one study reported a decline in rejection and inconsistent discipline across three years of early adolescence (Lengua, 2006), and another study reported stability in parental control, including proactive control, punitive control, and psychological control across four time points from early to late adolescence (Van Heel et al., 2019). Indeed, there has been evidence indicating that parent-child relationship quality decreased significantly at the onset of the teenage years but stabilized by late adolescence (Ebbert et al., 2018). However, another study reported that conflict in the parent-child relationship increased from early to middle adolescence and then decreased from middle to late adolescence (De Goede et al., 2009).

Taken together, available studies emphasize changing dynamics in the parent-child relationship, including increased negative parenting and conflict around early adolescence that seems to stabilize with development across adolescence. While many adolescents may pass through this period without significant difficulties in their relationships with their parents, some adolescents do not. That is, in some families, the unique demands placed on parents during adolescence may manifest as negative parenting practices (emphasized here as nagging, criticism, and yelling) with important implications for adolescent adjustment. Given the fact that these behaviors are targeted in many parenting interventions, yet are more likely to be observed in a typically developing sample than behaviors such as child abuse, clarification of differential developmental trends of negative parenting is important for identifying adolescents who are at risk of maladaptive development. Person-centered approaches, such as growth mixture modeling, allow for identification of latent subgroups within the sample (Muthén & Muthén, 2000), and are thus well suited to examine meaningful heterogeneity in parenting trajectories that may be obscured by variable-centered methods such as latent growth curve modeling. Yet, person-centered approaches have not been applied to examine unique classes of changes in negative parenting during adolescence.

Emotion Regulation

ER can be defined as a dynamic process encompassing both the internal (i.e., cognitive control, attention shifting, and physiological responses) and external (i.e., the socialization and coaching of ER via parents, teachers, peers and others) processes used to modify the experience and expression of an emotion in socially and contextually appropriate ways (Gross & Thompson, 2007; Morris et al., 2017; Thompson, 1994). The process model of ER by Gross and Thompson (2007) articulates the ways in which individuals can regulate emotions, in addition to the consequences of particular strategies. The model describes how individuals navigate life through an infinite amount of situation selections, or complex choices, with the ability to modify their responses to each one. However, regulation attempts require intentional thought and behavior to alter the emotion itself, the situation, or the meaning of the situation (Gross & Thompson, 2007). Gross (1998) proposes that strategies are classified according to their temporal position along the pathway of emotion generation. That is, strategies that are antecedent-focused (occur before an emotional response is fully activated) are often more adaptive than response-focused strategies (occur once an emotion has been fully activated) given their focus on responding to an emotion rather than modifying the emotion pathway (Gross & Thompson, 2007). Difficulties with ER may be reflected by trouble with the abilities necessary for the modification of emotions (e.g., awareness and understanding of emotions, acceptance of emotions, attention, and impulse control) or by employing ER strategies that are unsuccessful in the long-term, or successful in the short-term but with consequences for long-term well-being (Cole et al., 2017).

However, competency in ER may not necessarily equate to use of adaptive ER strategies. First, capacity for adaptive ER may not align with socialized adaptive strategies, and second, many strategies are considered to be both adaptive and maladaptive depending on the context. According to the process model of ER by Gross and Thompson (2007), ER strategies can be differentiated by when they have the greatest impact on the emotion process. For example, cognitive reappraisal, whereby one may alter how to think about a situation or their capacity to manage the demands it poses, is typically viewed as an adaptive ER strategy (Gross & John, 2003). Cognitive reappraisal involves consciously reinterpreting a stressful or negative situation with a positive interpretation, and those that actively engage in cognitive reappraisal tend to experience more positive affect, less anxious and depressive symptoms, and better overall adjustment (Gross & John, 2003). In contrast, suppression, which involves inhibition of emotional expression and of the subjective experience of a distressing emotion, is often considered a maladaptive ER strategy. Though it can be beneficial in the short term, over time it can increase physiological arousal associated with distressing emotions (Gross & Thompson, 2007; John & Gross, 2004). Those who frequently use suppression tend to demonstrate less positive affect and have trouble overcoming negative affect (Gross & John, 2003).

Patterns of Emotion Regulation Development

Although research examining developmental trajectories of ER abilities is limited, extant work suggests that ER continues to develop from early life through adulthood, alongside cognitive development. Indeed, the developmental period of adolescence reflects a profound shift in cognitive abilities, including “a more fully conscious, self-directed and self-regulating mind” (Steinberg, 2005). Neuroscience research also suggests steady maturation in the prefrontal cortex, underling cognitive advancements in adolescence (Kalisch, 2009; Ochsner & Gross, 2008; Paus, 2005). Such increasingly complex and adaptive cognitive capabilities across time have significant implications for adolescent emotional development, with emotional and cognitive processes consistently interreacting to inform ER (Ochsner & Gross, 2008; Steinberg, 2005). Accordingly, ER typically exhibits positive growth from childhood to adulthood (McRae et al., 2012). For example, a longitudinal study investigating within-person changes in parent- and adolescent-reported ER abilities demonstrated increases from early adolescence into young adulthood (Hardy et al., 2019). Furthermore, using latent change score modeling, another longitudinal study reported positive constant change in ER abilities across adolescence with a slight deceleration in ER change from year-to-year (Herd et al., 2020). As such, prior studies seem to consistently suggest increasing trends in ER abilities across adolescence.

Given that higher-order cognitive abilities may underlie ER strategies, such as reappraisal (Schweizer et al., 2020), it follows then that cognitive reappraisal (supported by high-order cognitive abilities) may exhibit similar increases across adolescence. However, research commenting on developmental patterns of suppression is mixed. While abilities underlying the use of more adaptive strategies may increase, the motivation to use suppression may increase alongside heightened awareness of consequences of certain emotional displays (Gross & Cassidy, 2019). Indeed, adolescents use cognitive ER strategies, including cognitive reappraisal, less frequently than adults (Garnefski et al., 2002) suggesting that use of cognitive reappraisal may increase with development. One study by McRae et al. (2012) compared how behavioral and neural ER may differ among older children, adolescents, and young adults. Results provide support for a linear increase in cognitive reappraisal ability across developmental periods, both behaviorally and in activation of the left ventrolateral prefrontal cortex, which has been previously associated with cognitive reappraisal in adults. However, Gullone et al. (2010) demonstrated that younger adolescents relied on suppression more so than older adolescents, whereas no significant difference existed for cognitive reappraisal. Furthermore, findings regarding within-person changes revealed that cognitive reappraisal use showed no significant change across time, whereas suppression use decreased over two years among adolescent girls, but not adolescent boys. Given the limited and inconsistent evidence, it is not clear what normative developmental patterns of cognitive ER strategies may be.

Associations between Negative Parenting and Emotion Regulation

Previous work has shown that the development of ER abilities and strategies is a transactional process, affected by both child and parent factors. The quality of parent-child interactions forms the basis of ER development (Cole et al., 2017). In adolescence, adaptive ER development is best facilitated by parents who encourage autonomous ER while providing a supportive environment in which adolescents’ acceptance of and coping with negative emotions is encouraged and supported (see Morris et al., 2007 for a review). In contrast, negative parenting practices may impede ER development by increasing emotional arousal and diminishing opportunities to practice adaptive, effective ER. For example, cross-sectional studies demonstrated that high levels of maternal psychological control predicted emotional dysregulation among adolescents (Cui et al., 2014; Luebbe et al., 2014). Similarly, suboptimal parenting behaviors, such as inconsistent discipline and corporal punishment, were related to adolescents’ suppression (Balan et al., 2017). Further, longitudinal research on self-regulation (a broader construct related to ER) has shown that high levels of parents’ psychological control predicted declines in adolescent self-regulation (Rogers et al., 2019) and that high levels of maternal discipline were associated with slower increases in adolescent self-regulation (Moilanen & Rambo-Hernandez, 2017). Finally, high levels of parental alienation predicted more use of suppression among adolescents (Gresham & Gullone, 2012). Taken together, these studies suggest that negative parenting behaviors may be harmful to adaptive ER development during adolescence. Not only do negative parenting practices model dysregulated behaviors, but they may also force adolescents to suppress their emotions in order to avoid provoking further parental hostility. Furthermore, parents high in negativity are less likely to foster an atmosphere in which support and guidance is given to adolescents to learn and practice adaptive ER strategies. However, this body of work often relies on assessing levels of negative parenting and their association with levels of ER. The novelty of the present study is that by using a person-centered approach we may identify potentially multiple distinct developmental trajectories of negative parenting, and can then examine whether and how differences in the parenting context during adolescence may differentially shape adolescents’ trajectories of ER development.

Present Study

The aims of the present study were three-fold: First, we tested the possibility of heterogeneous patterns of changes in negative parenting across adolescence, using growth mixture modeling. Given limited and mixed findings on patterns of change in parenting constructs during adolescence, we did not have specific hypotheses about the number of classes that may emerge nor their trajectories. Second, we used growth curve modeling to clarify trajectories of ER development across adolescence, including ER abilities and two putative ER strategies: cognitive reappraisal and suppression. In general, we expected that ER would become more adaptive across adolescence. Thus, we hypothesized that ER abilities would demonstrate positive growth trajectories, even though previous research examining developmental trajectories of ER is limited and the results are inconsistent. Furthermore, given that prefrontal maturation during adolescence underlies many higher order cognitive abilities, we similarly hypothesized that cognitive reappraisal use would increase as advanced cognitive abilities may make it less effortful. Likewise, we hypothesized that suppression use would decrease as abilities underlying the use of more adaptive ER strategies may increase. Alternatively, the motivation to use suppression as a result of heightened awareness of consequences of certain emotional displays may lead to increases (Gross & Cassidy, 2019). Finally, we sought to link the identified patterns of within-person changes in negative parenting across adolescence to adolescent developmental trajectories of ER. Considering negative parenting as a contextual risk factor for ER development, we hypothesized that differential experiences of negative parenting would be related to developmental trajectories of ER, such that adolescents exposed to higher levels of negative parenting across time (whether stable or increasing) would exhibit compromised ER development.

Method

Participants

The community sample included 167 adolescents (53% males) and their parents (82% biological mothers) from a southeastern state of the United States of America, who participated in annual assessments across four years. The sample size was determined using an a priori power analysis that was conducted as part of the grant submission that funded the larger project that this study was a part of. Participant enrollment was completed once the target sample size was reached. The sample size provided more than adequate power to detect a modest hypothesized effect. Adolescents were 13 to 14 years of age at Time 1, 14 to 15 years of age at Time 2, 15 to 16 years of age at Time 3, and 16 to17 years of age at Time 4 (M = 14.07, SD = 0.54 for Time 1; M = 15.05, SD = 0.54 for Time 2; M = 16.07, SD = 0.56 for Time 3; and M = 17.01, SD = 0.55 for Time 4). Eighty-two percent of adolescents identified as Caucasian, 12% African-American, and 6% other. Eighty-eight percent of parents identified as Caucasian, 10% African American, and 2% other. Median family income was $35,000 – $49,999 per year at all times. At Time 1, 157 adolescents participated. At Time 2, 10 adolescents were added (to offset annual attrition) for a final sample of 167 (150 at Time 2, 147 at Time 3, and 150 at Time 4). However, 24 adolescents did not participate at all possible time points for reasons including: ineligibility for tasks (n = 2), declined participation (n = 17), and lost contact (n = 5) during the follow-up assessments. We performed attrition analyses using general linear model (GLM) univariate procedure to determine whether there were systematic predictors of missing data. Results indicated that rate of participation (indexed by proportion of years participated to years invited to participate) was not significantly predicted by age, income, sex, or race (p < .58) or study variables at Time 1 (p > .19).

Procedures

Data included in the present study were collected as part of a larger project. Adolescent participants and their parents were recruited via email announcements, flyers, and snowball sampling (word-of-mouth). Data collection was administered at university offices where participants completed self-report questionnaires, behavioral and neuroimaging tasks, and were interviewed by trained research assistants. The study duration was on average five hours long and participants were compensated monetarily for their time. All procedures were approved by the institutional review board of Virginia Tech and written informed consent or assent was received from all participants.

Measures

Negative parenting.

Adolescents reported on aspects of their relationship with their parent (primary caregiver) using the Parent-Child Relationship Scale (Hetherington & Clingempeel, 1992) at Times 1–4. Participants responded to eight items on a Likert-type scale with responses ranging from “1 = Extremely” to “5 = Not at all”. We created a negative parenting score based on three items from this measure. The items include, “How much does this person yell at you after he or she has had a bad day?”, “How much does this person nag you about what you are doing wrong?”, and “How much does this person criticize you?”. Items were reverse scored such that higher scores represent greater negativity from the parent to the adolescent. The negativity score demonstrates acceptable reliability in our sample (α= .72 – .80 across the four time points).

Emotion regulation abilities.

The Emotion Regulation Checklist (ERC; Shields & Cicchetti, 1997) was measured at Times 1–4, to capture adaptive emotion regulation abilities, including socially appropriate emotional displays, empathy, and emotional self-awareness. Adolescents self-reported on a four-point Likert scale (from “1 = Rarely/Never” to “4 = Almost Always”) about how they respond to different situations. Mean scores were calculated across 8 items that reflected the emotion regulation subscale, such that higher scores were indicative of better emotion regulation. Sample items include “I can say when I am feeling sad, angry or mad, fearful or afraid” and “I show concern and understanding when others are upset or distressed”. While the subscale demonstrated relatively low reliability (α = 0.49–0.63) in the current sample, this is similar to figures in past research that also used adolescents (Zaremba & Keiley, 2011).

Emotion regulation strategies.

Adolescents’ ER strategies were measured at Times 1–4 with self-report on the Emotion Regulation Questionnaire (ERQ; Gross & John, 2003). This questionnaire includes 10 items that describe emotion regulation strategies with responses ranging from “1 = Strongly Disagree” to “7 = Strongly Agree”. There are two subscales, suppression and reappraisal. Sample items include, “When I am feeling positive emotions, I am careful not to express them” and “When I want to feel less negative emotion (such as sadness or anger), I change what I am thinking about”. The ERQ demonstrates acceptable reliability in our sample (α= .75 – .86 across the four time points for reappraisal; α = .67 – .72 across the four time points for suppression).

Data Analytic Plan

Descriptive statistics and correlations for all study variables are presented in Tables 1 and 2. For all variables, descriptive statistics were used to assess for normal distributions and outliers. Skewness and kurtosis were also examined, and levels less than 3 and 10, respectively, were considered acceptable (Kline, 2011). All variables demonstrated acceptable levels of skewness and kurtosis. Outliers (N = 4) were identified as values deviating more than 3.29 SD from the mean (Tabachnick & Fidell, 2001) and were Winsorized to retain statistical power and attenuate bias resulting from elimination (Ghosh & Vogt, 2012). Multivariate general linear modeling analyses indicated that demographic covariates, including age, sex, race, and family income were not significant predictors of study variables (ps > .07) and were therefore not included as covariates in our analyses. The only exception was the effect of adolescent sex on ER abilities at Time 1, but given the small effect size (η ² = .06) and the fact that it only predicted ER abilities at one of the four time points, we decided against including sex as a covariate.

Table 1.

Descriptive Statistics of Study Variables

	M	SD	Min	Max
Whole Sample
1. Negative Parenting T1	2.07	0.90	1.00	5.00
2. Negative Parenting T2	2.20	0.84	1.00	5.00
3. Negative Parenting T3	2.27	0.91	1.00	5.00
4. Negative Parenting T4	2.18	0.88	1.00	5.00
5. ER Abilities T1	3.13	0.37	2.25	3.88
6. ER Abilities T2	3.10	0.41	2.00	4.00
7. ER Abilities T3	3.14	0.41	2.13	3.88
8. ER Abilities T4	3.15	0.41	2.13	4.00
9. Reappraisal T1	4.93	0.94	2.33	7.00
10. Reappraisal T2	4.87	0.88	2.50	7.00
11. Reappraisal T3	4.92	1.07	1.67	7.00
12. Reappraisal T4	4.96	0.94	2.17	7.00
13. Suppression T1	3.78	1.17	1.25	7.00
14. Suppression T2	3.85	1.14	1.25	6.75
15. Suppression T3	3.78	1.16	1.25	6.25
16. Suppression T4	3.85	1.19	1.50	7.00
ER in the Decreasing Group
1. ER Abilities T1	3.14	0.35	2.38	3.88
2. ER Abilities T2	3.11	0.40	2.00	3.88
3. ER Abilities T3	3.17	0.40	2.13	3.88
4. ER Abilities T4	3.21	0.36	2.25	4.00
5. Reappraisal T1	4.91	0.94	2.33	7.00
6. Reappraisal T2	4.87	0.89	2.50	7.00
7. Reappraisal T3	4.97	1.05	1.67	7.00
8. Reappraisal T4	4.99	0.96	2.17	7.00
9. Suppression T1	3.78	1.20	1.25	7.00
10. Suppression T2	3.88	1.14	1.25	6.75
11. Suppression T3	3.74	1.12	1.25	6.25
12. Suppression T4	3.77	1.14	1.50	7.00
ER in the Increasing Group
1. ER Abilities T1	3.11	0.44	2.25	3.75
2. ER Abilities T2	3.06	0.45	2.00	4.00
3. ER Abilities T3	3.04	0.45	2.25	3.88
4. ER Abilities T4	2.90	0.51	2.13	3.75
5. Reappraisal T1	5.02	0.98	3.00	7.00
6. Reappraisal T2	4.91	0.89	3.33	6.50
7. Reappraisal T3	4.74	1.15	1.67	6.83
8. Reappraisal T4	4.83	0.83	2.50	6.17
9. Suppression T1	3.78	1.05	1.75	5.75
10. Suppression T2	3.71	1.13	2.00	6.50
11. Suppression T3	3.93	1.32	1.50	6.25
12. Suppression T4	4.17	1.36	1.75	6.50

Note: T1 = Time 1; T2 = Time 2; T3 = Time 3; T4 = Time 4.

Table 2.

Bivariate Correlations of Study Variables

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
1. Negative Parenting T1	-
2. Negative Parenting T2	.64**	-
3. Negative Parenting T3	.53**	.63**	-
4. Negative Parenting T4	.57**	.50**	.69**	-
5. ER abilities T1	−.17*	−.15	−.14	−.11	-
6. ER abilities T2	−.08	−.13	−.19*	−.02	.50**	-
7. ER abilities T3	−.11	−.17*	−.24**	−.07	.51**	.52**	-
8. ER abilities T4	−.08	−.09	−.27**	−.23**	.38**	.52**	.57**	-
9. Reappraisal T1	−.16*	−.01	−.03	−.10	.29**	.19*	.13	.19*	-
10. Reappraisal T2	−.18*	−.07	−.13	−.08	.34**	.34**	.27**	.21*	.44**	-
11. Reappraisal T3	−.20*	−.07	−.20*	−.12	.20*	.32**	.33**	.25**	.30**	.53**	-
12. Reappraisal T4	−.18*	−.11	−.18*	−.17	.15	.25**	.18*	.41**	.33**	.28**	.42**	-
13. Suppression T1	−.00	.06	.04	−.01	−.45**	−.33**	−.28**	−.20*	.04	−.07	−.09	−.02	-
14. Suppression T2	−.02	.06	.01	−.12	−.29**	−.39**	−.28**	−.26**	−.05	−.06	.02	−.05	.58**	-
15. Suppression T3	.03	.06	.10	.02	−.31**	−.31**	−.50**	−.41**	−.05	−.14	−.01	−.07	.53**	.62**	-
16. Suppression T4	.02	.10	.11	.06	−.26**	−.40**	−.43**	−.56**	.00	−.05	−.14	−.10	.44**	.49**	.61**

Notes:

^*p < .05,

^**p < .01.

T1 = Time 1; T2 = Time 2; T3 = Time 3; T4 = Time 4.

The hypothesized models were tested using Structural Equation Modeling (SEM) in Mplus statistical software version 8 (Muthén & Muthén, 1998–2018). Model fit was assessed by χ² value, degrees of freedom, corresponding p-value, Root Mean Square Error of Approximation (RMSEA), and Confirmatory Fit Index (CFI). RMSEA values less than .08 and CFI values greater than .90 were considered an acceptable fit (Browne & Cudeck, 1993). Full information maximum likelihood (FIML) estimation procedure (Arbuckle, 1996) was used to address missing data given its superiority to those obtained with listwise deletion or other ad hoc methods (Schafer & Graham, 2002).

For ER abilities and ER strategies, univariate unconditional growth curve models were specified to assess change over time. Linear growth and nonlinear growth models were tested to fit the observed data patterns across the four time points. The first latent factor was the intercept, with all factor loadings fixed to one. The second latent factor was the slope, indicating rate of change over time. In the linear growth model, a linear pattern of change was assumed and factor loadings for the slope factor were fixed to 0, 1, 2, and 3. The latent basis growth model was used to test a non-liner growth pattern because it allowed the data to estimate the shape of growth by fixing the first and last time points (to 0 and 1, respectively) and freely estimating the second and third. In the latent basis growth model, the estimated factor loadings represented the proportion of total change that has occurred since the first measurement occasion. The χ² difference test was used to compare these nested models and the most parsimonious model with acceptable fits was chosen as the best-fitting model (see Table 3).

Table 3.

Model Fit for Univariate Growth Models

Model Label	χ2	df	p	RMSEA	CFI	Δχ2	Δdf	p(d)
ERC- Emotion Regulation Abilities
Linear growth model	7.43	8	.49	.00	1.00
Latent basis growth model	7.19	6	.30	.03	.99	.24	2	.89
ERQ- Suppression
Linear growth model	4.45	8	.81	.00	1.00
Latent basis growth model	3.96	6	.68	.00	1.00	.49	2	.78
ERQ- Reappraisal
Linear growth model	21.03	8	.01	.10	.88
Latent basis growth model	13.09	6	.04	.08	.93	7.94	2	.02

Notes: Best-fitting baseline model in boldface. CFI = comparative fit index; RMSEA = root mean square error of approximation; Δχ² = difference in likelihood ratio tests; Δdf = difference in df; p(d) = probability of the difference tests.

For negative parenting, growth mixture modeling (GMM) was used to identify the smallest number of trajectory classes that described repeated measurements of negative parenting. GMM estimates latent factors for intercepts and slopes of developmental trajectories of a variable to test whether there are two or more distinct classes of individuals. GMM is able to capture the heterogeneity in growth trajectories of a construct and determines the optimal class membership for each individual (Kim-Spoon & Grimm, 2016). As such, each class is made up of individuals with similar growth trajectories and uniquely estimated values for intercept and slope. Criteria used to determine the optimal number of classes included (i) sample-size adjusted Bayesian information criterion (SABIC; Sclove, 1987) indicated parsimony goodness of fit based on the log-likelihood adjusted for the number of parameters; Lower SABIC estimates are indicative of better model fit; (ii), the adjusted Lo- Mendell-Rubin Test (ALMR; Lo et al., 2001) is a nested model likelihood ratios test indicative of whether k trajectory classes provide a better fit than k–1 trajectory classes; A small probability (p < 0.05), indicated that a k–1 model should be rejected in favor of a model with at least k classes, and (iii) usefulness of the classes was examined based on the number of individuals in each class (i.e., minimum of 5% or more in each class). Table 4 presents SABIC, ALMR, and entropy estimates for models with a one, two, and three class solution.

Table 4.

Fit Indices for Latent Class Solutions

Number of latent classes	SABIC	ALMR	Entropy
Negative Parenting
1 class	1294.23	-	-
2 class	1271.08	27.72 *	.76
3 class	1260.46	15.47**	.84

Notes: Bold information face indicates best fitting model, SABIC = sample-size adjusted Bayesian criterion, ALMR = adjusted Lo-Mendell-Rubin Test.

^*p < .05,

^**p < .01,

^***p < .001.

After determining the optimal number of negative parenting classes and the most likely membership of each individual by negative parenting trajectory class as well as establishing the best-fitting growth curve models for the ER variables, multiple-group SEM was used to evaluate whether distinctive negative parenting classes exhibited different ER trajectories across adolescence.

Results

Developmental Trajectories of Negative Parenting

The two-class model provided the best fit for negative parenting (see Table 4). While the three-class solution also provided a good fit to the data, the third class contained only 3 individuals (2% of the sample). A residual correlation between Time 2 and Time 3 was added given its significance and the fact that issues with negative slope variance emerged without it. The two classes were characterized as (1) decreasing negative parenting (n = 137, about 82%) and (2) increasing negative parenting (n = 30, about 18%; see Table 5). For Class 1, the decreasing negative parenting class, the mean of the intercept (M = 2.09 SE = 0.07, p < .001) and the negative mean of the slope (M = −0.07, SE = 0.03, p = .009, were significantly different from zero, indicating significant decreases across time. For Class 2, the increasing negative parenting class, the mean of the intercept (M = 2.15 SE = 0.22, p < .001) and the positive mean of the slope (M = 0.43, SE = 0.05, p < .001) were significantly different from zero, indicating significant increases across time. The most likely class membership was then extracted for each participant in order to test whether different patterns of ER development existed for each class. For Class 1 (decreasing), the group means for negative parenting at each time point were: 2.09 at Time 1, 2.02 at Time 2, 1.95 at Time 3, and 1.88 at Time 4. For Class 2 (increasing), the group means for negative parenting at each time point were: 2.15 at Time 1, 2.57 at Time 2, 3.00 at Time 3, and 3.43 at Time 4. These means are presented graphically in Figure 1.

Table 5.

Intercept and slope estimates for two-class solution of changes in negative parenting

	Class label	Members	Intercept mean	Slope mean
Class 1	Decreasing	137 (82%)	2.09***	−.07**
Class 2	Increasing	30 (18%)	2.15***	.43***

Note:

^**p < .01,

^***p < .001.

Figure 1.

Growth trajectories for negative parenting from Time 1 to Time 4.

Note. See text for mean values at each time point for each class.

Developmental Trajectories of Emotion Regulation

As seen in Table 3, for ER abilities, the linear model provided the best fit to the data. The variance of the intercept (σ² = 0.07, SE = 0.02, p < .001) was significant, whereas the variance of the slope (σ² = 0.00, SE = 0.00, p = .091) was not. Thus, the results indicated significant individual differences in initial levels of ER abilities, but not in growth rates. The mean of the intercept (M = 3.12, SE = 0.03, p < .001), but not the slope (M = 0.01, SE = 0.01, p = .310), was significantly different from zero, indicating that the mean level of ER abilities did not significantly change over time.

For suppression, the linear model provided the best fit to the data. The variance of the intercept (σ² = 0.89, SE = 0.15, p < .001) and slope (σ² = 0.07, SE = 0.02, p = .003) were significant, indicating significant individual differences in both initial levels and rate of change. The mean of the intercept (M = 3.80, SE = 0.09, p < .001), but not slope (M = 0.02, SE = 0.03, p = .603), was significantly different from zero, indicating that the mean level of suppression did not significantly change over time.

For cognitive reappraisal, the latent basis growth model provided the best fit to the data. The variance of the intercept (σ² = 0.33, SE = 0.08, p < .001) was significant, whereas the variance of the slope (σ² = 0.03, SE = 0.07, p = .711) was not, indicating significant individual differences in initial levels of reappraisal but not in the rate of changes. Further, the mean of the intercept (M = 4.94, SE = 0.06, p < .001), but not slope (M = −0.01, SE = 0.03, p = .706), was significantly different from zero, indicating that the mean level of reappraisal did not significantly change over time.

Linking Developmental Trajectories of Negative Parenting with Developmental Trajectories of Emotion Regulation

Next, we tested whether ER trajectories differed depending on the negative parenting class to which the adolescent belonged. A dichotomous variable indicating the two groups of negative parenting was created (Class 1 = 1, Class 2 = 2) and two-group growth models were run separately for ER abilities, suppression, and cognitive reappraisal (see Table 6 for unstandardized parameter estimates). The model test command in Mplus was used to determine whether the means of the intercept and slope of ER were statistically different between negative parenting groups. Wald’s test of parameter constraints was used to test whether imposing the equality constraints significantly degrade model fit or not. Significant Wald’s test suggested that the two groups differed significantly.

Table 6.

Parameter Estimates for Testing the Two-Group Model of Emotion Regulation Development by Negative Parenting

	B	SE	p
Emotion Regulation Abilities
Decreasing group
Intercept mean	3.12	0.03	.000
Slope mean	0.03	0.01	.007
Intercept variance	0.07	0.02	.000
Slope variance	0.00	0.00	.970
Increasing group
Intercept mean	3.13	0.07	.000
Slope mean	−0.06	0.03	.045
Intercept variance	0.10	0.04	.018
Slope variance	0.01	0.01	.085
Suppression
Decreasing group
Intercept mean	3.83	0.10	.000
Slope mean	−0.02	0.04	.684
Intercept variance	0.97	0.17	.000
Slope variance	0.07	0.03	.004
Increasing group
Intercept mean	3.68	0.18	.000
Slope mean	0.14	0.07	.059
Intercept variance	0.48	0.27	.075
Slope variance	0.02	0.05	.630
Cognitive Reappraisal
Decreasing group
Intercept mean	4.88	0.07	.000
Slope mean	0.04	0.03	.287
Intercept variance	0.32	0.10	.001
Slope variance	0.02	0.02	.236
Increasing group
Intercept mean	4.98	0.18	.000
Slope mean	−0.07	0.07	.305
Intercept variance	0.54	0.28	.058
Slope variance	0.01	0.04	.753

Emotion regulation abilities.

The two-group model demonstrated acceptable fit (χ² = 19.60, df = 16, p = .238, RMSEA = .05, CFI = .98). In the decreasing negative parenting group, the ER intercept mean was significant and the significant slope mean indicated increases in ER over time. The variance of the intercept, but not the slope, was significant, indicating significant individual differences in initial levels of ER, but not in the rate of changes. In contrast, in the increasing negative parenting group, the ER slope mean indicated significant decreases in ER over time. The variance of the intercept, but not the slope, was significant, indicating significant individual differences in initial levels of ER, but not in the rate of changes.

Next, we tested whether imposing the equality constraints on the ER slope means between the decreasing and increasing negative parenting groups would degrade model fit. Wald’s test of parameter constraints indicated that such constraints did significantly degrade model fit, suggesting that the two groups significantly differ with respect to the ER slope means (Wald χ² = 7.85, df = 1, p = .005). Imposing the equality constraints on the intercept means did not significantly degrade model fit, suggesting that the two groups did not differ with respect to the ER intercept mean (Wald χ² = .02, df = 1, p = .881).

Suppression.

The model demonstrated acceptable fit (χ² = 7.47, df = 16, p = .963, RMSEA = .00, CFI = 1.00). In the decreasing negative parenting group, the intercept mean was significant, but the slope mean was not, indicating non-significant changes in suppression over time. The variances of the intercept and slope were significant, indicating significant individual differences in initial levels as well as the rate of changes. In the increasing negative parenting group, the intercept mean was significant, but the slope mean was not significant (p = .059), indicating no growth in suppression over time. Neither the variance of the intercept nor the slope was significant.

Next, we tested whether imposing the equality constraints on the suppression slope means between the decreasing and increasing negative parenting groups would degrade model fit. Wald’s test of parameter constraints indicated that such constraints did not degrade model fit, suggesting that the two groups do not differ with respect to the suppression slope means (Wald χ² = 3.49, df = 1, p = .062). Imposing the equality constraints on the intercept means did not significantly degrade model fit, suggesting that the two groups did not differ with respect to the suppression intercept means (Wald χ² = .57, df = 1, p = .451).

Cognitive Reappraisal.

Though the latent basis univariate growth model for cognitive reappraisal was best (as indicated by the chi-square difference test), there were issues with negative variance in the slope in the two-group growth models of reappraisal with a latent basis growth curve model. Therefore, given that model fits for the linear growth model were just as good and did not produce issues with negative variance in the slope, we selected the linear growth model as the best-fitting model to interpret. The initial model fit was poor (χ² = 28.37, df= 16, p = .029, RMSEA = 1.00, CFI = .88). A residual correlation between cognitive reappraisal at Time 2 and Time 3 was added to improve model fit (based on the modification index). The resulting model fit was acceptable (χ² = 20.09, df= 14, p = .127, RMSEA = .07, CFI = .94). In the decreasing negative parenting group, the intercept mean was significant, but the slope mean was not, indicating no significant changes in cognitive reappraisal over time. The variance of the intercept was significant, but not the slope, indicating significant individual differences in initial levels but not in the rate of changes. In the increasing negative parenting group, the intercept mean was similarly significant, but the slope mean was not significant, indicating no growth in cognitive reappraisal over time. Neither the variance of the intercept nor the slope was significant.

Next, we tested whether imposing the equality constraints on the cognitive reappraisal slope means between the decreasing and increasing negative parenting groups would degrade model fit. Wald’s test of parameter constraints indicated that such constraints did not significantly degrade model fit, suggesting that the two groups did not differ with respect to the slope means of cognitive reappraisal (Wald χ² = 1.92, df = 1, p = .166). Imposing the equality constraints on the intercept means also did not significantly degrade model fit, suggesting that the two groups did not differ with respect to the cognitive reappraisal intercept mean (Wald χ² = .24, df = 1, p = .627).

Discussion

Extant literature has demonstrated that ER continues to develop across adolescence at both behavioral and neurological levels (McRae et al., 2012). Yet, literature examining contextual factors associated with developmental trajectories of ER is limited. Notable neurocognitive maturation, coupled with the unique, fast-changing social contexts associated with adolescence, highlight this developmental period as especially relevant for investigating contextual processes involved in shaping individual differences in ER. Given the continued pervasive influence of parenting on emotional development, the present longitudinal study sought to identify distinctive developmental trajectories of negative parenting and examine how they are differentially related to ER development in adolescence. Overall, our findings provide evidence for the link between negative parenting and maladaptive ER, such that adolescents who experience increases in negative parenting demonstrate decreases in ER ability development.

While the adolescents in our sample entered adolescence reporting similar levels of negative parenting, our results identified two separate developmental patterns of adolescent perceptions of negative parenting. Class 1 contained the majority of adolescents who reported moderate initial levels of negative parenting that decreased across adolescence. Class 2 contained a smaller group of adolescents who reported moderate initial levels of negative parenting that increased across adolescence. While changes in both groups were significant, the increasing group appeared to have steeper changes than the decreasing group. In general, these findings corroborate current theoretical and empirical work describing notable changes to parent–adolescent relationships during this developmental period (Branje, 2018; De Goede et al., 2009; Ebbert et al., 2018; Eisenberg, et al., 2008; Putnick et al., 2008; Seiffge-Krenke et al., 2010; Steinberg & Silk, 2002). In particular, the current findings based on adolescents’ perceptions of negative parenting complement prior findings of distinct trajectories based on adolescents’ positive and negative feelings towards their parents.

Specifically, using growth mixture modeling analyses of data from German adolescents, Seiffge-Krenke et al. (2010) reported distinctive groups of adolescent-mother relations characterized by 1) a ‘normative’ trajectory in which adolescents reported higher levels of support-closeness with their mothers that decreased slightly alongside low stable levels of negative affect (61%); 2) an ‘increasingly negative’ trajectory in which adolescents reported lower levels of support-closeness with their mothers that continued to decrease alongside increasing levels of negative affect towards their mothers (30%); and 3) a ‘decreasingly negative’ trajectory in which adolescents reported lower levels of support-closeness with their mothers that continued to decrease alongside decreasing levels of negative affect (9%). Similar trajectories were identified for adolescent-father relations with one difference. Similarly, we also identified an increasing trajectory of negative parenting among relatively smaller numbers of adolescents. The ‘normative’ group and ‘decreasingly negative’ group found in Seiffge-Krenke et al.’s study had low or decreasing levels of negative affect towards their parents, and thus seem to be similar to the adolescents in our decreasing negative parenting group, encompassed by the majority in our typically developing sample. Discrepancies are likely due to differences in the measurement of parenting. That is, while Seiffge-Krenke et al. (2010) focused on adolescent’s support and closeness with and affect toward parents, the present study focused on negative parenting practices (i.e., parent behaviors that demonstrated negativity towards the adolescent).

Turning to developmental trajectories of ER, our expectation that ER would show general adaptive trends was not supported. Our results did not demonstrate significant changes over time in any of the ER variables. These results were surprising considering prior work suggesting that increasing adaptive ER abilities should coincide with maturation in prefrontal brain regions that accelerates during adolescence (Lewis & Stieben, 2004; Perlman & Pelphrey, 2011). However, studies examining individual differences in growth trajectories of ER abilities within adolescence are rare. One previous study used adolescent and parent reported emotional self-regulation and reported a positive slope between ages 11 and 22 (Hardy et al., 2019). Another study using latent change score modeling, differentiated between positive constant change in ER abilities versus negative year-to-year change across ages 13 through 17 (Herd et al., 2020). Perhaps, an increasing trend in ER abilities is more readily detectable when considering broad age ranges across pre-to-late adolescence as in Hardy et al.’s study. For instance, ages leading up to and following middle-adolescence may indicate a general increasing trend (i.e., average change) that co-exists with diminishing year-to-year change (as in Herd et al.’s study), resulting in deceleration of change over time. Indeed, extant work has noted a U-shaped dip in self-regulatory traits in late child and early adolescence before increasing again in mid-late adolescence (see Atherton, 2020 for a review).

We found no clear evidence of systematic developmental changes in ER strategy use, including cognitive reappraisal. This finding is consistent with Gullone et al. (2010)’s study which did not find hypothesized increases in within-person changes of cognitive reappraisal, positing that stability in cognitive reappraisal may develop prior to adolescence and be established as early as age 9. While this explanation seems to query continued prefrontal maturation during adolescence that underlies many higher order cognitive abilities, cognitive reappraisal use may be more stable and trait-like by the beginning of adolescence. However, research has shown that the majority of children can engage in cognitive reappraisal by age 9, but that the skill likely continues to develop into adolescence (Pons et al., 2004). Future research may consider modeling cognitive reappraisal longitudinally over a broader age range, including late childhood and pre-adolescence, to determine whether changes in cognitive reappraisal can be better captured in earlier developmental periods. Regardless, it is important to note that because cognitive reappraisal is likely not uniformly adaptive, we may not be able to expect that the use of cognitive reappraisal would consistently increase with age simply because the prefrontal cortex matures and self-control abilities improve. As suggested by McRae (2016), it is important to consider contextual factors that determine when cognitive reappraisal is an effective or adaptive strategy and when it is not (e.g., an individual’s level of control over the stressor, availability of cognitive resources, the nature of the emotion, an individual’s goal, conditions of psychobiological or environmental adversity), which may influence an adolescent’s decision to engage it.

Turning to suppression, our data revealed significant individual differences in use of suppression strategies, but no significant change over time. While general developmental patterns may not necessarily exist, individual differences do, such that depending on context, adolescents decide which strategies to use, when to do so, and how to express them. In particular, given the number of unique challenges associated with adolescence (e.g., pervasive influence of the limbic system, reorganization of relationships, new and increased intensity in emotions; McRae et al., 2012; Steinberg, 2005), some adolescents may use suppression to mitigate the cause of the distressing emotion (Gross & Thompson, 2007). Indeed, reluctance to express emotions is positively correlated with ER abilities (Riley et al., 2019).

Importantly, discussion of ER development is not complete without considering the environment in which development is unfolding. That is, while ER development is derived partly from neurobiological maturation and behavioral capabilities, it also depends greatly on interactions between emotional arousal and environmental responses. Given the continued pervasive influence of parenting on adolescent emotion development, the present study considered how ER development may differ depending on adolescents’ exposure to particular patterns of change in negative parenting. While unconditional growth modeling indicated no significant changes in ER development as a whole sample, we found evidence of significant developmental changes in ER across varying patterns of negative parenting. This highlights the advantage of our person-oriented approach which identified subgroups of negative parenting and elucidated that the parenting context is vital for uncovering significant developmental changes in ER.

The differential developmental patterns of negative parenting evidenced in the present study may reflect how the familiar parenting context becomes somewhat novel to adolescents during this period, with unique implications for adolescent emotional development. Our results demonstrated that developmental patterns of negative parenting are associated with differential patterns of ER development across adolescence. Although the adolescents in both groups of negative parenting exhibited similar levels of ER abilities at the beginning of adolescence, their developmental trajectories varied notably. Specifically, adolescents in the decreasing negative parenting group exhibited significant increases in their ER abilities overtime, whereas adolescents in the increasing negative parenting group exhibited significant decreases in their ER abilities overtime. Our results support the body of literature documenting the continued importance of optimal parenting practices throughout adolescence in order to promote adaptive ER development (Morris et al., 2007). Although initial elevation of negative parenting at the beginning of adolescence may be considered relatively typical (Branje, 2018), for adolescents that see normative decreases in negative parenting, ER continues to develop along an adaptive trajectory. Indeed, our findings are consistent with a recent study documenting that adolescent reported decreases in paternal control predicted increasing ER for adolescents (Van Lissa et al., 2019). On the other hand, suboptimal parenting characterized by hostile behaviors may hinder adaptive ER development by constraining adolescents’ opportunities to learn, acquire, and practice adaptive strategies in a supportive context.

Furthermore, regarding strategy use, results indicated that both adolescents’ suppression and cognitive reappraisal use did not show significant changes overtime regardless of parenting group. For suppression, this result was somewhat surprising given that youth who experience negative parenting are more likely to view their emotions negatively and avoid expressing them (Morris et al., 2007). That is, while suppression often does not serve to lessen unpleasant emotions (Gross & John, 2003), it may be adaptive in the short-term to discourage continued interactions of negative parenting. Regarding cognitive reappraisal it may be the case that it is fairly stable by adolescence (Gullone et al., 2010) and thus, less susceptible to change as a result of the parenting environment. Overall, ER strategy use seems less sensitive to the parenting environment as ER abilities.

Though prior research has demonstrated that negative parenting (including control, punishment, and criticism) may be associated with less adaptive ER and poorer emotional adjustment (Balan et al., 2017; Cui et al., 2014; Luebbe et al., 2014; Rogers et al., 2019), it is important to recognize that just as an environment characterized by negative parenting may deprive adolescents of opportunities and supportive contexts for learning and practicing adaptive ER, it may also be the case that increased adolescent emotion dysregulation creates tension and conflict that elicits increases negative parenting. Indeed, a study by Van Lissa et al., (2019) reported that, though lower levels of perceived paternal control predicted increases in ER, adolescent-reported ER also predicted adolescent-perceived parenting, such that adolescents reporting increases in ER felt that they were more supported and controlled less. Perhaps this is a feedback loop in which parents and adolescents continually respond to each other’s behavior. In both the current study and that by Van Lissa et al., (2019), it is possible that changes in adolescent ER abilities influenced adolescents’ perceptions of parents’ behavior. For example, adolescents demonstrating increases in ER abilities may also be better able to tolerate negative parenting that may be somewhat typical of this developmental period, leading them to perceive their parents’ negativity as decreasing, and vice versa. Future research should attempt to elucidate this potential bidirectional association.

Nonetheless, the association between negative parenting and ER is a concern given its potential for cascading risk into further maladjustment. For example, for adolescents experiencing difficulties in ER, it is not unreasonable to expect engagement of maladaptive ER strategies, such as rumination, that are linked with an array of clinical diagnoses (Aldao et al., 2010). Thus, the present study identifies negative parenting as a potential intervention point for adolescents at risk of developing maladaptive outcomes associated with difficulties in ER. Intervention efforts may be able to focus on parental education of the developmental period of adolescence, including increased parental stress, and teach skills related to stress management and optimal parenting practices (e.g., quality communication), such that negativity relatively typical of this period does not foster decreases in ER development.

Limitations and Future Directions

The present study should be interpreted in light of several limitations with implications for avenues of future research. First, given our relatively small sample coupled with the complexity of the analyses, there is a need for replication of our findings, particularly in more ethnically and culturally diverse community populations. Furthermore, though it was beyond the scope of the current study to explore, there may have been significant baseline differences between adolescents in the two negative parenting groups that affected ER development (e.g., temperament, personality, psychopathology) and thus, replication is necessary. Second, all variables were assessed using self-report. Consequently, associations among the variables might have been artificially inflated by method variance due to single informant or mono-method bias. Future studies should consider using data from multiple informants (e.g., parents, peers, and teachers) and multiple methods (e.g., observation and interview). Relatedly, adolescent reports on the Parent-Child Relationship Scale may have been limited in terms of capturing the most crucial aspects of the parent-child relationship that affect ER. For example, how consistent adolescent-reported parent-child relationship and observed parenting behaviors may be, and whether observed parenting behaviors may better capture some nuances of parenting that contribute more readily to adolescent ER, will be important questions to answer in future studies. Furthermore, ER research in general will be strengthened by recognizing that since ER can be expressed differently across various dimensions (e.g., abilities, multiple different strategies) single indicators of ER may not be the best means of capturing the complexity of ER. Measurement of ER that jointly considers multiple dimensions will strengthen how we capture it as a construct and be helpful in aligning ER measurement with its inherently complex definition. Finally, it is important to consider child effects on negative parenting during adolescence given that research has demonstrated that youths’ emotion dysregulation is associated with negative parenting and parental distress and has suggested a bidirectional relationship between child ER and parental reactions (King et al., 2013; Van Lissa et al., 2019).

Conclusions

The developmental period of adolescence is unique for both adolescents and their parents. Key developmental changes associated with this period, including cognitive and physical changes stemming from puberty, necessitate a reorganization of the ways in which parents and their adolescents relate. Our findings are consistent with the idea that perceptions of negative parenting can be amplified during the developmental transition into adolescence, but that for many adolescents, these perceptions decrease across time. Nonetheless, adolescents’ experiences of negative parenting seem to have important implications for their ER development. The present study is the first to systemically test how parenting behaviors and ER are coincidentally changing across adolescence, illustrating the unique influence parents have on their adolescents’ ER development, whereby persistent high levels of negative parenting may be disruptive to adaptive ER development. Therefore, intervention efforts aimed at mitigating risk for future maladjustment for adolescents with difficulties in ER can meaningfully target negative parenting patterns.