Intraindividual Variability in Adolescent Impulsivity: The Predictive Role of Family and Peer Relationships

Adolescent trait impulsivity is generally regarded as an individual risk factor for a host of problem outcomes. Trait impulsivity is generally defined as acting without forethought and can also include a tendency to accept immediate rewards rather than greater, but delayed rewards, difficulties with planning, and low self-control (van Baal, 2022; Beauchaine et al. 2017). This individual difference factor is recognized as an underlying risk for developmental trajectories of externalizing problems. Impulsivity is recognized as related to deficits in executive functioning, particularly inhibitory control (Leshem & Yefet, 2019) and plays a central role in models of developmental psychopathology, in which is serves as a heritable trait characteristic underlying risk for Attention-Deficit Hyperactivity Disorder (ADHD), Conduct Disorder, poor educational outcomes and problem peer association in adolescence, all of which ultimately set the stage for problematic substance use and antisocial behavior problems in adulthood (Beauchaine et al., 2009). A related model also places trait impulsivity as a key developmental risk factor for borderline personality disorder (Crowell et al., 2009). Trait impulsivity also has been linked with eating disorder risk (Kenny et al., 2019; Wonderlich et al. 2024) as well as non-suicidal self-injury (Riley et al., 2015). Thus, impulsivity is a risk factor of wide-reaching social significance.

Adding to the developmental psychopathology literature documenting a progression across disruptive behavior disorders from early life to adulthood, trait impulsivity also serves as a salient risk indicator at later developmental periods. For example, childhood trait impulsivity is associated with externalizing problems in later developmental periods (e.g., Stewart et al. 2023; Wang et al., 2016). Even in late adolescence, trait impulsivity signals risk for externalizing problems (Ahmad & Hinshaw, 2017) and addiction (Gullo et al. 2023). Thus, trait impulsivity is an important risk indicator across developmental periods.

Historically, impulsivity has been conceptualized as a trait characteristic, but recent evidence suggests that impulsivity also exhibits state-like qualities, fluctuating within-person on a day-to-day (or even moment-to-moment) basis. Studies of adult participants have revealed that state-level measures of impulsivity can reliably capture within-person variability (Tomko et al. 2014; Stevens et al., 2020). These studies disaggregating within- and between-person variability in impulsivity report intraclass correlations ranging from 0.35 to 0.44, indicating 56–65% of the variability in impulsivity may occur within-persons, rather than between-persons (Feil et al., 2020). In disentangling within-person from between-person variance, it is possible to capture both trait-like and state-like features of impulsivity. Trait-like conceptualizations reflect the dispositional aspect of impulsivity, consistent with studies assessing between-person differences in risk, or classification systems for diagnosing individuals with ADHD. State-like qualities refer to ways in which impulsivity changes within-person from day to day. Thus, individuals’ experiences of impulsivity may be more or less than their usual level. Simply stated: people have good and bad days. Because state impulsivity fluctuates within person, it may be malleable to situational factors such as interpersonal relationships, stress, or other experiences. Considering a state-like perspective of impulsivity enables novel insights into the factors that might help manage or exacerbate one’s tendency to be impulsive.

The emerging evidence regarding daily variation in impulsivity has important implications for individual well-being. A handful of studies now link within-person variation in impulsivity with daily mood. On days of higher impulsivity than usual, people tend to feel increased levels of negative affect (Stevens et al., 2020). Other work suggests that impulsivity may have lasting effects into the next day’s mood: days with higher than usual levels of impulsivity predict next-day diminished positive affect (Depp et al., 2016), and next-day elevated negative affect (Titone et al. 2022), although the link with negative affect is sometimes not evident (Depp et al., 2016). Nonetheless, daily variation in impulsivity appears to be important for daily mood and well-being.

State-level impulsivity is often also viewed as episodes of diminished inhibitory control (Chaku et al., 2024), underscoring how state impulsivity is linked to risks. For example, increases in state impulsivity may lead to increased urges to engage in problem behavior, such as the desire to engage in behaviors such as gambling, substance use, risky behavior, and eating (van Baal et al. 2022). Consistent with this work, a study using daily diary methods with a large sample of college students revealed that on days of high state impulsivity, participants reported more sexual risk-taking (Riley et al., 2019). Thus, state-level increases in impulsivity may serve as a signal for risk of engaging in serious problem behavior.

Daily variation in impulsivity may be a risk factor for substance use. For example, on days of elevated impulsivity, young adults were more likely to use alcohol (Griffin & Trull, 2021; Riley et al., 2019), with other work suggesting that days of high impulsivity may increase risk for heavy episodic drinking (Stamates et al., 2019; Stevens et al., 2020). At the same time, some studies fail to find a link between impulsivity and substance use either in the form of null findings in same-day analyses (e.g., Griffin et al., 2021; McGowan et al., 2022), other studies conducting lagged analyses have yielded results suggesting a reverse direction of effects in which substance use predicts elevated impulsivity on the following day (Stamates & Lau-Barraco, 2020). However, although findings regarding the link between state impulsivity and substance use are somewhat mixed at this early stage of research, they do suggest that daily variation in impulsivity may hold important implications for alcohol use risk.

To date, much of the work studying within-person variability in impulsivity has been focused on clinical populations (e.g., participants with borderline personality disorder; Tomko et al. 2014) and samples of late adolescents and adults (studies typically sampling participants that are 18 years or older). As a result, less is known about the degree to which impulsivity fluctuates for adolescents. However, if impulsivity does exhibit meaningful within-person fluctuations in adolescence, it would suggest that adolescents have days of higher and lower risk for problem behaviors. Understanding the factors that drive these within-person fluctuations could guide interventions around how to reduce short-term risk for adolescents’ distress, engagement in substance use, problem behavior, or even suicide.

An important question remains: what are factors that explain day-to-day variability in adolescents’ impulsivity? By identifying daily experiences that elicit within-person changes in adolescent impulsivity, we may be better able to characterize the conditions under which adolescents are at elevated or decreased risk for problem behaviors. Further, by using within-person methods, we can identify malleable processes that may guide interventions aimed at reducing adolescent risk (Molenaar, 2004).

Little is known about the predictors of daily variation in adolescent impulsivity; to our knowledge, the work to date has also focused on late adolescents and adults (i.e., ages 18 and older). However, existing work suggests that negative affect and stressful experiences are important risks associated with variability in state impulsivity. In clinical and non-clinical samples, experiences of negative affect correspond to increases in impulsivity (Feil et al., 2020; Tomko et al., 2015); lagged analyses link elevated negative affect with increases in next-day impulsivity (Depp et al., 2016). Finally, a recent study evaluated daily stressors, positive mood, and daily mood as unique predictors of state impulsivity in a clinical sample. In this study (Sharpe et al., 2021), unique effects were found for each predictor: impulsivity was higher on days when negative affect was higher than usual, when participants experienced more stressors than usual, and when positive affect was lower than usual. Taken together, this literature suggests that identifying stressors in adolescents’ lives may be a fruitful avenue for identifying factors that account for daily variation in impulsivity.

Family and peer relations represent salient domains of adolescent daily experiences (Steinberg, 2001). The current study was designed to capture daily fluctuations in family, parent-adolescent, and peer relationships, and evaluate whether they explain within-person variability in impulsivity. As no studies have been conducted on this question to date, we turned to the developmental literature for preliminary evidence. Across family-level, parent-adolescent, and peer domains, longitudinal evidence suggests these interpersonal relations are associated with later impulsivity. For example, longitudinal panel model data indicate reciprocal relations between family conflict and trait impulsivity between late childhood (ages 9–10) and early adolescence (ages 11–12) (Yin et al., 2025). However, family conflict predicted late adolescent trait impulsivity, but reciprocal effects were not found during this period (Elam et al., 2016). Other work has found a longitudinal link between positive parent-adolescent relationship quality and later adolescent delay discounting (Kahn et al., 2015). Likewise, reciprocal relations between peer rejection and ADHD symptoms are evident in childhood (Stenseng et al., 2016), and again in adolescence (Stenseng et al., 2024) – a finding that was evident in early to middle adolescence (Ji et al., 2019). Thus, during the adolescent developmental period, there is evidence for prospective, reciprocal relations between impulsivity and interpersonal relationships; however, there seems to be a slight indication that in middle- to late-adolescence, interpersonal relations may be somewhat more of a driving factor in these processes.

The current study was designed to address two research aims. First, we sought to replicate work documenting meaningful within-person variability in impulsivity among adults in a non-clinical sample of adolescents. Using a 21-daily diary design, we collected adolescent self-report data on their daily impulsivity. We expected to find that these daily reports of impulsivity exhibited meaningful within-person variability. We first tested this by calculating an intraclass correlation, and expected to find that 20% or more of the variance would be attributed to within-person variability, suggesting that we can evaluate predictors of daily variation in this construct (Bolger & Laurenceau, 2013).

Second, we evaluated family and peer predictors of daily variation in adolescent impulsivity. In the family domain, we examined family-level and parent-adolescent relationship quality. Specifically, we examined family-level cohesion and conflict, as well as parent-adolescent closeness and conflict. All four of these constructs have been found in prior research to (a) exhibit meaningful within-person variability, and (b) they each are linked with adolescent negative or positive affect (Bai et al., 2017; Chung et al., 2011; Coffey et al., 2020; Fosco et al., 2019, 2020; Fosco & Lydon-Staley, 2020; Lippold et al., 2016). Thus, we expected that, on days when family conflict and parent-adolescent conflict were higher than usual, adolescents would experience increased impulsivity. Conversely, we evaluated whether increases in family cohesion or parent-adolescent closeness would be associated with decreases in adolescent impulsivity.

In the peer domain, we considered peer rejection and peer relationship satisfaction as indicators of stressful (i.e., rejection) and pleasant (i.e., satisfaction) daily peer experiences. A robust literature documents that these aspects of peer relationships in adolescence exhibit meaningful day-to-day variability and also are linked with adolescents’ daily mood and well-being (Bai et al., 2017; Ha et al., 2019; Mayfield & Fosco, 2021; Morrow et al., 2014; Weinstein et al., 2006; Xu et al., 2023). Similar to analyses with family processes, we expected that experiences of peer rejection would correspond to increases adolescents’ impulsivity and days of high peer relationship satisfaction would correspond to decreases in adolescents’ impulsivity.

Because of the dearth of research studying daily family and peer factors in relation to adolescent impulsivity, we did not have specific hypotheses about which family or peer domains would have the most robust effects. Thus, we first examined these relations by domain (i.e., separate models for family-level cohesion and conflict, parent-adolescent closeness and conflict, and peer relationship satisfaction and rejection). We then used a model-building approach to test the unique effects of the significant predictors of adolescent impulsivity in an integrative model. Finally, given the evidence in the developmental literature for reciprocal relations, we conducted follow-up lagged analyses to test the direction of effects in our models.

Table 1 presents the descriptive statistics and correlations for the study variables at both within-person and between-person levels; demographic data are provided in Supplemental Table S1. To address our first research question, we conducted a variance decomposition analysis to examine the within-person variability in adolescent impulsivity. The ICC was 0.6018, indicating that 60.18% of the daily variability in adolescent impulsivity can be attributed to between-person differences and 39.82% can be attributed to within-person variability. Based on guidance from Bolger and Laurenceau (2013), this surpasses the acceptable level (20%) of within-person variability needed to proceed with analyses of within-person process. Thus, our hypothesis that there would be meaningful within-person variability was supported. An illustrative plot of daily variation in adolescent impulsivity of the sample is presented in Fig. 1.

In the next step, we computed three sets of multilevel models separately to evaluate whether variability in family-level, parent-adolescent, and peer relations corresponded to changes in adolescents’ state impulsivity. Within each domain, we evaluated main effects and interactive effects. If interactions were not statistically significant, they were omitted from the final (presented) model for parsimony. Models evaluating family-level and parent-adolescent relations, models were computed separately for adolescent-report and parent-reports of family relations. Parents did not report on peer relationships, so only one set of analyses was conducted on peer relationships. Thus, five multilevel models were computed for the first set of analyses.

Table 2 reports the findings from the analyses of family-level and parent-adolescent relationships. In three of the four models (Models 1–3), no significant interaction effects were found; thus, only the main effects are reported. We first examined the models examining family-level conflict and cohesion in relation to adolescent impulsivity. In the adolescent-report model (Model 1), family conflict was a risk, such that on days of higher family conflict, adolescents reported increases in their impulsivity (\(\:\beta\:\) = 0.11, p <.01); no relation was found for family cohesion. In the model evaluating parent-reported family processes, neither family conflict nor family cohesion were associated with adolescent impulsivity (Model 2).

In the models examining parent-adolescent relationships, adolescent report of parent-adolescent conflict was associated with impulsivity, but parent-adolescent connectedness was not (Model 3). Specifically, on days when adolescents perceived higher than usual levels of conflict with their parents, they experienced increases in their impulsivity (\(\:\beta\:\) = 0.12, p <.01). In the model evaluating parent-reported data (Model 4), there was a statistically significant interaction between parent-adolescent conflict and connectedness (\(\:\beta\:\) = − 0.03, p <.01). To probe this interaction, we conducted a Johnson-Neyman analysis, which indicated that on days when parent-adolescent conflict was higher than usual, adolescents reported higher levels of impulsivity, but only when parent-adolescent connectedness was low (See Fig. 2).

We then computed a model of peer rejection and peer relationship satisfaction in relation to daily impulsivity (Table 3). Peer relationship satisfaction was not associated with impulsivity; however, on days when peer rejection was higher than usual, adolescents reported increased impulsivity (\(\:\beta\:\) = 0.19, p <.01). A similar pattern emerged in the between-person effects: average peer rejection was positively associated with daily adolescent impulsivity (\(\:\beta\:\) = 0.68, p <.01), but no relation was found regarding satisfaction with peer relationships.

In the final step, we integrated all statistically significant predictors of impulsivity into a comprehensive model (See Table 4) and conducted the analyses separately for adolescent-reported family processes and parent-reported family processes. The adolescent-report model (Model 6) largely replicated findings from the separate models described above. On days when adolescents perceived higher family conflict (\(\:\beta\:\) = 0.09, p <.05), parent-adolescent conflict (\(\:\beta\:\) = 0.05, p <.05), and peer rejection (\(\:\beta\:\) = 0.16, p <.01) than usual, they reported increases in impulsivity. Regarding between-person effects, average parent-adolescent conflict (\(\:\beta\:\) = 0.41, p <.05) and peer rejection (\(\:\beta\:\) = 0.51, p <.05) were positively associated with adolescent impulsivity. The parent-report model (Model 7) also replicated results from models testing domains separately, described above. Indeed, the interaction between parent-adolescent conflict and connectedness was observed again (\(\:\beta\:\) = − 0.03, p <.01), as was the effect for daily peer rejection (\(\:\beta\:\) = 0.14, p <.01), average peer rejection (\(\:\beta\:\) = 0.64, p <.01), and average parent-adolescent conflict (\(\:\beta\:\) = 0.44, p <.05) positively predicted daily impulsivity.

Random intercepts and slopes were also estimated for day-level variables. The random intercept captured individual differences in adolescents’ average levels of impulsivity. Across models, the random intercept variances (e.g., \(\:{\sigma\:}^{2}\)=2.49 to \(\:{\sigma\:}^{2}\) =2.89) indicated significant individual differences in adolescents’ average levels of impulsivity. Random slopes reflect individual differences in the magnitude of within-person associations observed in the day-level predictors and impulsivity. In other words, it allowed the effect of daily experiences on impulsivity to vary across adolescents. For example, in Model 1 (Table 2), the random slope variances for daily family cohesion (\(\:{\sigma\:}^{2}\)=0.05) and family conflict (\(\:{\sigma\:}^{2}\)=0.03) suggested that adolescents varied in the association between daily family relationship and impulsivity.

Adolescent trait impulsivity has long been recognized as an underlying risk characteristic in ADHD, Conduct Disorder, Substance Use Disorder, and Antisocial Personality Disorder (Beauchaine et al., 2009), while other work links trait impulsivity to borderline personality disorder (Crowell et al., 2009), eating disorders (Kenny et al., 2019; Wonderlich et al., 2024) and non-suicidal self-injury (Riley et al., 2015). Recently, we have begun to expand our notion of impulsivity beyond conceptualizing it as a trait characteristic, to consider within-person fluctuations in state impulsivity. The current study had two goals. First, we sought to replicate prior work studying adults to determine whether there is meaningful within-person variability in impulsivity among adolescents. Second, we sought to identify family and peer processes that contribute to adolescents’ fluctuations in impulsivity. Ultimately, by examining daily processes underlying within-person variability in impulsivity, we hoped to better understand the conditions under which adolescents are at more or less at risk for poorer mental health or risky behaviors.

Adolescent trait impulsivity is a robust risk indicator for a wide range of psychopathology outcomes. Our study documents that adolescent impulsivity exhibits both trait-like and state-like qualities, calling for an expanded conceptualization of adolescent impulsivity. In addition, our findings suggest that conflict in interpersonal relationships – family-level, parent-adolescent, and peer – appear to be key risk factors for change in adolescent state impulsivity. These findings suggest that interventions may fruitfully target conflict in close relationships to reduce adolescent risk for impulsivity.