Links between parental monitoring and parent-adolescent conflict: A multi-modal test of bidirectional relations

Abstract

During adolescence, youth increase in both independence and conflict with parents. Parents vary in how much they know about their adolescents’ whereabouts and activities, and how they acquire this information (i.e., the sources of what parents know). We probed how parental knowledge of adolescents’ whereabouts and activities―and their information sources―relates to (a) domains of parent-adolescent conflict (fighting about, or having different beliefs about, daily life topics); and (b) parent and adolescent attachment-related behavior during a conflict discussion task. Using the Actor-Partner Interdependence Model, we tested links between parental knowledge/its sources and conflict processes. Eighty-seven adolescents (age_m=15.18; 55% female) and parents completed surveys about parental knowledge and its sources (i.e., parental solicitation of adolescents’ activities, adolescent disclosure to parents about their activities), and separate interviews on conflict domains. A subset of parent-adolescent dyads (n = 65) interacted for 5 minutes about an adolescent-identified conflict topic. Different beliefs about daily life topics related to parental knowledge: parents’ reports of greater different beliefs about daily life topics predicted less knowledge of adolescents’ activities/whereabouts, solicitation, and disclosure, for both parent and adolescent reports of these domains. For adolescents, greater different beliefs related to less solicitation and disclosure. Only adolescent reports of parental knowledge, solicitation, and disclosure predicted attachment-related behaviors both dyad members displayed during the conflict discussion task. Findings reveal links between parental knowledge of adolescents’ activities and conflict processes, and demonstrate dyadic interdependence between parental knowledge of adolescents’ activities and conflict processes.

Youth develop independence and autonomy as they progress through the developmental periods that encompass adolescence. Parent-adolescent relationships also change during this time. Adolescents spend more time away from caregivers, and they may have different perspectives than caregivers on daily life topics—schoolwork, chores, and time with family (De Los Reyes et al., 2012). Accordingly, conflict increases in middle adolescence (Laursen et al., 1998), and is also theorized to facilitate adolescent autonomy development (Smetana et al., 2006). Parents serve as important guardrails during this time by enforcing rules and providing instrumental and emotional support. However, caregivers may become less physically present around adolescents as they become more independent, although behavioral control/monitoring remains important to prevent problematic engagement in risk behaviors (Racz & McMahon, 2011). As such, monitoring tends to occur through communication instead, measured via parental knowledge and how parents obtain knowledge of adolescents’ whereabouts and activities (Kerr & Stattin, 2000). The focus of the current study was to investigate dyadically how parental knowledge of adolescents’ activities is related to parent-adolescent conflict domains about daily life topics, and their behavior when discussing those topics.

Despite increasing autonomy, adolescents still view their parents as important attachment figures during this time, using them as a secure base from which to explore their widening world (Allen & Manning, 2007; Allen et al., 2003; Allen et al., 2007). Caregivers who are sensitive to adolescents’ needs for autonomy while providing emotional and instrumental support facilitate adolescents’ well-being as the relationship transitions to a “supervisory partnership” (Koehn & Kerns, 2018; Waters et al., 2019, pg. 2379). Conversely, rejecting, absent caregiving is associated with worse adjustment (Allen et al., 2018; Allen et al., 2007). Parental rejection and hostility during this time has been linked to a host of negative adolescent outcomes. In a meta-analysis investigating the aspects of parenting behaviors that are associated with adolescent delinquency, the strongest links were those related to hostility and rejection (Hoeve et al., 2009). Mutually hostile behavior between parents and youth relates to greater current and future youth externalizing behaviors (Glatz et al., 2019). Potentially stemming from these factors, difficulties in the parent-adolescent relationship, including conflict and inadequate supervision, have been associated with poor adolescent outcomes, from substance misuse to psychiatric symptoms (Ary et al., 1999; Bray et al., 2022; Weymouth et al., 2016). Historically, the direction of influence studied has been from parents to adolescents (Lionetti et al., 2019; Racz & McMahon, 2011; Smetana et al., 2006). Yet, there remains an understudied dynamic. Specifically, the dynamic systems approach posits that parents and adolescents influence each other—adolescents’ behavior may evoke proximal reactions and future adaptations from a parent and vice versa (Granic & Patterson, 2006). In fact, by the time youth reach adolescence, they and their parents have already accumulated over a decade’s worth of reciprocal, interpersonal interactions that shape each other’s future behaviors. Similarly, Patterson’s coercion theory posits that a process may occur over time by which parents and youth react to each other’s requests and behavior, which may progress into adolescence, with parents withdrawing from their supervision role if adolescents’ responses to parental control become aversive (Patterson, 2016). Informed by these theoretical frameworks, we took a dyadic approach to examining the link between parent-adolescent conflict processes and parental monitoring, using multi-modal, multi-informant data.

Parental Monitoring

Within the category of behavioral control, parental monitoring is one of the most researched facets of parenting and a dynamic issue in the field for several decades (Hayes et al., 2003; Koehn & Kerns, 2018; Lionetti et al., 2019; Racz & McMahon, 2011; Smetana et al., 2006; Stattin & Kerr, 2000). Key definitions include “a set of correlated parenting behaviors involving attention to and tracking of the child’s whereabouts, activities, and adaptations” (Dishion & McMahon, 1998, pg. 61), as well as “parents’ knowledge of the child’s whereabouts, activities, and associations” (Kerr & Stattin, 2000, pg. 368). Parental monitoring is consistently and robustly associated with youth well-being and behavioral concerns (Fosco et al., 2012; Hamza & Willoughby, 2011; Lopez-Tamayo et al., 2016; Rusby et al., 2018). For example, a meta-analysis of 119 studies found a significant, negative relationship between monitoring and adolescent delinquency with a medium effect size (r = −0.26; Hoeve et al., 2009).

During adolescence, autonomy increases across development, and as such, parents are often not direct observers of youths’ behaviors. Consequently, parental knowledge of adolescents’ whereabouts and activities, as well as how parents acquire this knowledge (i.e., sources of knowledge), is included within contemporary conceptualizations of the larger parental monitoring construct (Kerr & Stattin, 2000; Lionetti et al., 2016; Lionetti et al., 2019). For example, one domain is general knowledge about what parents know about where adolescents spend their time, school functioning, and peers. Increased general knowledge relates to increased adolescent school performance and decreased antisocial behavior (Criss et al., 2015). Another domain includes where parents obtain that information—through active attempts to question or learn more about their adolescent (i.e., solicitation), or from adolescents sharing this information with parents (i.e., disclosure). These sources of knowledge―solicitation and disclosure―matter because they vary in their links to adolescent functioning.

Greater adolescent disclosure to parents relates to greater parental knowledge, better school performance, and less delinquency (Criss et al., 2015; Kerr et al., 2010). Adolescents’ disclosure may provide parents with opportunities to impart advice and support, a notion bolstered by research indicating that increased adolescent-reported disclosure relates to decreased depressive symptoms―an effect that is mediated by parental knowledge (Hamza & Willoughby, 2011). Parental solicitation has been theorized to protect against future antisocial behavior among frequently-unsupervised adolescents―particularly those who question parental authority (Laird et al., 2010). Yet, increased parental solicitation also relates to increased adolescent behavioral concerns (Stattin & Kerr, 2000), and in other cases demonstrates no links to adolescent outcomes (Criss et al., 2015). Although knowledge, solicitation, and disclosure are more indirect routes to monitoring, both direct monitoring and these indirect routes have similar associations of moderate effect with adolescent delinquency, with the largest effect size for youth disclosure (knowledge: r = −.26; disclosure: r = −.31; monitoring: r = −.23; Hoeve et al., 2009).

Parent-Adolescent Conflict

Parent-adolescent conflict, which increases in intensity in middle adolescence around age 14 through age 17 cross-culturally (Laursen et al., 1998; Smetana et al., 2006), has been linked to parental monitoring. In one study, increased adolescent-reported parental supervision and rules related to decreased reports of conflict (Hayes et al., 2004). With regard to parental solicitation from adolescents about their whereabouts and activities, it has been posited that these parental inquiries, especially with increasing frequency, may be perceived by adolescents as intrusive, thereby generating conflict (Hayes et al., 2003). However, conflict is a normal byproduct of parent-adolescent relationships―parents and adolescents normatively vary in their perceptions, and perhaps even have arguments, about daily life events or topics, including how much time adolescents spend with the family, their homework, and chores (Laursen & Collins, 1994). Moreover, conflict is theorized to facilitate both the transition to adolescent/emerging adult autonomy as well as confer negotiation skills (Smetana et al., 2006; Van der Giessen et al., 2014). Yet, how conflict is perceived and discussed, and whether individuals can maintain their connection with each other in the process, may determine whether this conflict confers skill development among adolescents, or alternatively, sets the stage for negative, coercive family exchanges (Allen et al., 2007; Weymouth et al., 2016).

Parent-adolescent conflict has been measured using self-report surveys and laboratory interaction tasks (e.g., Burt & Klump, 2014; Chaplin et al., 2012). Surveys often cover topics about which individuals disagree or fight about to identify frequency and intensity, whereas laboratory tasks leverage specific topics identified by parents and/or adolescents as conflictual, with laboratory personnel instructing parent-adolescent dyads to discuss these topics in the laboratory setting. Assessing parent-adolescent interactions provides the opportunity to observe the interplay between expressing independent viewpoints while maintaining relatedness to the other individual (Allen et al., 1994; Smetana et al., 2006). Although conflict has been a frequent focus of investigations because of its links with adolescent outcomes, studies vary in the instrumentation used, and instruments vary in their ability to directly index conflict. For example, a recent systematic review found that a commonly used measure in studies of conflict assesses dissatisfaction with a dyad member’s behavior and interaction quality, rather than conflict (Marshall et al., 2022). Many studies only assess adolescents’ perceptions of conflict, whereas studies that take a multi-informant, multi-modal approach to assess both parents and adolescents often investigate how divergent perspectives on daily life topics could be related to behavioral concerns. There has been less focus on the distinction between components of these conflict interactions. Indeed, there is measurable distinction between perceived behavioral manifestations of conflict (i.e., do parents and/or adolescents report fighting about daily topics) and perceived differences in how parents and adolescents view these topics, and prior work indicates that each of these domains has implications for understanding conflict (De Los Reyes & Ohannessian, 2016; De Los Reyes et al., 2012). That is, researchers do not always distinguish between topics about which adolescents and parents disagree but do not have fight about, versus topics that do generate conflictual interactions (De Los Reyes et al., 2012; Marshall et al., 2022).

Actor Partner Interdependence Model

A key reason for our limited understanding of the reciprocal interplay between parents and adolescents in their interactions with one another is that studies tend to focus on a limited set of pathways within these interactions (e.g., parental behavior→adolescent outcomes, adolescent well-being→parenting behavior). To understand how dyad members influence one another, and in the context of parent-adolescent interactions, requires analytic techniques that facilitate testing bidirectional relations. The Actor-Partner Interdependence Model (APIM; Cook & Kenny, 2005; Kenny et al., 2006) is a conceptual and statistical model of dyadic relationships that accounts for interdependence in dyads. That is, the APIM facilitates probing how dyad members influence one another, while accounting for each individual member’s effect. The APIM accounts for non-independence of data and allows for simultaneously estimating the influence of each partner’s independent variable on their own dependent variable (i.e., actor effect) and their partner’s dependent variable (i.e., partner effect; Cook and Kenny, 2005). Given the relational dynamics during adolescence, the APIM provides a lens through which to examine parental monitoring and the unique contributions of parents and adolescents in relation to conflict processes.

Current Study

We investigated how conflict processes are related to parental knowledge of adolescents’ activities. The current study was designed to advance knowledge of these domains by addressing limitations of other studies of parent-adolescent relationships/interactions. First, parents and adolescents were assessed on all the same domains across multiple modes of administration: interview, self-report, and observer ratings of behavior. Second, to distinguish domains of conflict (i.e., differentiate differing beliefs from arguments), we conducted structured interviews separately for parents and adolescents (De Los Reyes et al., 2012). Third, we measured parent and adolescent behavior during a conflict discussion task. Finally, we used an analytic model that accounts for the non-independent nature of adolescent and parent data. We first hypothesized that increased ratings on conflict domains (argue; have different beliefs) would be related to decreased ratings of parental knowledge of adolescents’ activities. Next, we predicted that the link between parental knowledge and behavior during the conflict discussion task would vary according to the valence of behavior. That is, we expected (a) increased ratings of parental knowledge and its sources to relate to increased secure base behavior, and (b) increased ratings of parental knowledge and its sources to related to decreased hostile behavior.

Methods

We report below how we determined sample size, data exclusions, manipulations, and study measures. Participants did not give written consent for their data to be shared publicly; due to the small sample size and sensitive nature of the research, supporting data is not available.

Participants and Recruitment

Participants were 87 adolescents between the ages of 14–17 (M _age = 15.18, SD = 1.05; 55% female), and their self-identified primary caregiver (also referred to as “parent”; M _age = 45.97, SD =7.68; range: 31–64). There were two methods of recruitment. For 40 dyads, they were recruited through advertising in the local community at agencies, events, websites (e.g., Craigslist) for a study on family relationships (De Los Reyes et al., 2012). For 47 dyads, adolescents and parents were initially recruited from a multiwave longitudinal study of community adolescents (see Collado et al., 2014 for recruitment details) about the development of youth risk behavior (“parent study”). The community recruitment strategy was the same as for the 40 dyads. Research personnel from the parent study asked participants’ caregivers if they were interested in participating in the present study. Caregivers provided contact information for research personnel affiliated with the current study to contact them, describe study procedures, and screen for eligibility. Research personnel then contacted interested, eligible caregivers to schedule a study visit with their adolescent.

Eligibility requirements for these studies included speaking English; understanding the consent/assent process; having an adolescent living at home without a reported history of substantial developmental, learning, psychotic disorders or cardiac conditions (for interpretability of cardiovascular data outside the scope of the current study for n=47 dyads); and willingness to attend a study visit with both parent and adolescent present. To be included in the current investigation, families had to have completed the measures of interest. Inclusion and exclusion criteria resulted in a final sample of 87 parent-adolescent dyads.

Parents reported adolescents’ race/ethnicity as African American or Black (63.2%); White, Caucasian American, or European (35.6%); Native American (3.5%); Other (3.5%); Hispanic or Latino/a (2.3%); or Asian American (2.3%). Parents reported their own race/ethnicity as African American or Black (63.2%); White, Caucasian American, or European (32.2%); Native American (1.2%); Other (2.3%); Hispanic or Latino/a (2.3%); or Asian American (2.3%). Ethnicity/race totals are greater than 100% because participants could select multiple ethnic/racial categories. Adults were: 13 biological fathers (14.94%), 70 biological mothers (80.46%), one adoptive mother (1.15%), two grandmothers (2.30%), and one aunt (1.15%). Parents reported their current marital status as never married (16.1%); married (51.7%); living together (6.9%); separated (1.2%); divorced (21.8%); or widowed (2.3%). All caregivers except for one reported completing at least a high-school education. According to parent-report of weekly household income, 17.24% of families earned $500 or less; 29.89% earned between $501 and $900; and 51.72% earned $901 or more per week (one family did not report).

Procedures

All study procedures were approved by the institutional review board of the large Mid-Atlantic university at which the study was conducted. After a thorough explanation by research staff, families provided informed consent/assent. Participants completed surveys and parents or caregivers completed a demographic survey. To assess dyadic conflict and identify conflict topics, parents and adolescents separately completed a structured interview (described below). Post-interview, dyads were informed which topic—either a conflict topic or a “control” topic (plan their dream vacation)—they would discuss together in a 5-minute discussion task.

Measures to Address Study Aims

Self-Report Measures

Sources of Parental Knowledge.

We assessed parental knowledge of adolescents’ activities and its sources (Kerr & Stattin, 2000). Adolescents and parents completed a validated 19-item parallel measure in which sets of items are totaled separately to result in three subscales per informant (De Los Reyes et al., 2010; De Los Reyes et al., 2008; Kerr & Stattin, 2000; Kerr et al., 2010). The Parental Knowledge subscale has 9 items assessing what parents know about adolescents’ whereabouts and activities (e.g., “Does your parent normally know where you go and what you do after school?”; possible range: 9–45). The Parental Solicitation subscale has five items about the extent to which parents try to obtain information about how adolescents spend their time away from parents (e.g., “During the past month, how often has your parent started a conversation with you about your free time?” possible range: 5–25). The Adolescent Disclosure subscale has five items that measure the extent to which adolescents tell their parents about their activities (e.g., “If you are out at night, when you get home, do you tell what you have done that evening?” possible range: 5–25). Possible responses range from “1” to “5” (range of frequency; Almost Never to Most of the Time). Higher scores indicate a greater amount of that domain (e.g., parental knowledge). Parent reports are structured such that phrases like “your parent” are replaced with “you”, and “you” replaced with “your teen”. Table 1 presents means and standard deviations. Cronbach alpha reliability estimates ranged from acceptable to excellent: Parental Knowledge: 0.97; parent report of Adolescent Disclosure: 0.94; Parental Solicitation: 0.70; adolescent report of Parental Knowledge 0.94; Adolescent Disclosure: 0.89; adolescent report of Parental Solicitation: 0.76. Correspondence between parent and adolescent reports ranged from medium to very large for knowledge (r = .37, p < .001 ), solicitation (r = .27, p = .011), and disclosure (r = .43, p < .001; see Funder & Ozer, 2019).

Table 1.

Means (M) and Standard Deviations (SD) for Study Variables, and Comparison Between Adolescents and Parents

	Adolescent	Parent		95% CI
Variable	M (SD)	M (SD)	t(df)	LL	UL	p	Cohen’s d
Knowledge	35.93 (6.34)	38.21 (4.14)	−3.44(86)	−3.59	−0.96	<0.001	−0.37
Conflict-only sample	35.08 (6.37)	38.06 (4.28)	−4.07(64)	−4.45	−1.52	<0.001	−0.50
Disclosure	18.89 (4.32)	20.44 (3.31)	−3.48(86)	−2.44	−0.66	<0.001	−0.37
Conflict-only sample	18.38 (4.56)	20.26 (3.30)	−3.54(64)	−2.94	−0.82	<0.001	−0.44
Solicitation	16.25 (4.83)	18.08 (3.84)	−3.22(86)	−2.95	−0.70	0.002	−0.35
Conflict-only sample	15.91 (4.89)	18.29 (3.83)	−3.82(64)	−3.63	−1.14	<0.001	−0.47
TTI Different Beliefs	8.55 (5.34)	12.70 (7.44)	−4.86(86)	−5.85	−2.45	<0.001	−0.52
Conflict-only sample	8.20 (5.34)	13.18 (7.82)	−5.23(64)	−6.89	−3.08	<0.001	−0.65
TTI Argue/Fight	6.80 (4.23)	7.84 (5.27)	−1.85(86)	−2.15	0.08	0.067	−0.20
Conflict-only sample	6.80 (4.42)	8.11 (5.65)	−1.94(64)	−2.66	0.04	0.057	−0.24
Hostility	2.56 (1.82)	2.77 (2.01)	−0.78(62)	−0.51	0.22	0.440	−0.11
Secure Base	4.61 (1.93)	4.89 (1.93)	−1.55(62)	−0.73	0.09	0.126	−0.15

Note. The sample size for the full sample is 87, and for the conflict-only sample is 65.

Interviews

To(may)to-To(mah)to Interview.

To measure domains of conflict related to daily life topics, we used The To(may)to-To(mah)to Interview (TTI; De Los Reyes et al., 2012). The TTI is a structured interview with demonstrated reliability and validity (De Los Reyes, Lerner, et al., 2013; De Los Reyes, Salas, et al., 2013; De Los Reyes et al., 2012) conducted separately with parents and adolescents on 16 daily life topics (e.g., youth’s grades, spending time with the family). Informants answer questions about these topics with regard to 1) behavioral conflict (“How often do you argue or fight with your child about your child getting good grades?”), and 2) perceptions of whether discrepancies exist between dyad members’ beliefs about the topics (“different beliefs”; “Do you think that you and your child have different beliefs about how often children his/her age should get good grades?”) between parent and adolescent, with responses ranging from 0 to 2 (None, Some, A lot). The assessment of different beliefs is a direct measure of the degree to which adolescents and parents disagree in how they perceive daily life topics. Responses for the 16 topics were summed, separately by domain; scores for both domains range from 0–32, with higher scores indicating more endorsement of the conflict domain. Cronbach alpha reliability estimates were as follows, and ranged from acceptable to good: parent report of Behavioral Conflict: 0.85 and Different Beliefs 0.88; adolescent report of Behavioral Conflict: 0.79 and Different Beliefs 0.79. With regard to validity, TTI scores are significantly related to both self-report and performance-based indicators of family functioning and interpersonal perception (De Los Reyes, Lerner, et al., 2013; De Los Reyes, Salas, et al., 2013; De Los Reyes et al., 2012). The TTI was also used in this investigation to choose the discussion topic for adolescents randomly assigned to the conflict discussion condition (described below).

Observational Measures

Discussion Task.

A subset of parents and adolescents (n = 65) completed a widely used 5-minute discussion task to resolve an assigned conflict topic (Donenberg & Weisz, 1997; Gottman, 1979). The task was designed to include factors that elicit a stress response (unpredictability, uncontrollability, and social evaluation; Dickerson & Kemeny, 2004; Gunnar et al., 2009). The conflict topic was chosen from adolescents’ response on the TTI; to prioritize inducing an adolescent stress response, the selection of the discussion topic prioritized the highest intensity topic available (e.g., a topic adolescent reported they argue with their parent about was preferred). To start the discussion, a research staff member stated the conflict topic to the parent and adolescent, invited them to provide a brief description of how this topic comes up for each of them, and instructed the dyad to come to a resolution on the details of the topic for 5 minutes. The research staff member then left the room and returned after 5 minutes. Video recordings using Noldus Observer XT (Zimmerman et al., 2009) were taken for later behavioral coding. The task was completed by a subset of the sample because one sample had a between-subjects design where families randomized to the control condition were assigned the discussion topic of planning a vacation (n = 22; not included in behavioral conflict analyses). For the other sample (n = 40), all families completed a control task and two conflict discussion tasks: one based on each dyad member’s TTI responses. To be consistent across samples, the current analyses only include the discussion tasks in which the adolescent conflict topic was discussed.

Behavioral Coding.

From video recordings, trained research staff coded conflict-related behaviors separately for parents and adolescents using the Conflict Task Coding System, a macro coding system (Ziv et al., 2002). Two scales were used: the Hostile Conflict Scale and Maintaining Secure Relatedness/Secure Base. Briefly, “secure base” is an attachment-related construct describing behaviors (examples below) related to a parent’s role as a secure base from which a child explores the world (Bowlby, 1988). In the context of the discussion task, behaviors are viewed from the lens of how dyad members maintain their relationship despite a potentially stressful discussion. Parent behaviors show their provision of a secure base to explore and be open to challenging discussions while maintaining connection to and understanding of the adolescent (e.g., validating adolescent’s frustration with a household rule yet remaining firm in enforcing the rule); adolescents’ behaviors indicate use of parent as a secure base—willingness to engage in open communication of their thoughts and feelings regarding the conflict, seek advice and comfort, and that their parent understands them. Behaviors (hostility, secure base) were rated from 1 to 7, with one score per behavioral domain given per dyad member for the entire 5-minute interaction; higher scores indicate greater hostility or secure base. Behaviors coded as hostile/rejecting included nonverbal behaviors (e.g., negative facial expressions, eye-rolling, slamming fist on table), and verbal behaviors (yelling, mocking, interrupting, character assassinations, and invalidating what the other person is saying). Behaviors coded as representing secure base also included nonverbal behaviors (e.g., frequent eye contact, nodding, relaxed/open body language) and verbal behaviors (validating adolescents’ statements, refrains from interrupting, receptive/encouraging of adolescents’ views, complimenting adolescent).

Coders were research assistants who had trained on the coding system using archived videos from a similar task. Their average scores on the training videos had to be within less than a 1-point difference from the “true” scores to be a coder in this project. To evaluate whether coders rated behaviors similarly, video recordings were randomly selected as reliability cases (62% of adolescent and 52% of parent videos) and assigned to two coders blind to whether cases were for reliability. Reliability cases were viewed and discussed during weekly meetings. Coders explained their ratings if they did not match. The first author made the final decision on ratings for reliability cases when necessary (e.g., coders disagreed, or it was clear both coders had not rated the behavior accurately). Means and standard deviation estimates are presented in Table 1. Interrater reliability was calculated from original scores on reliability coding cases using intraclass correlation coefficients (ICC) for “average measures” (Cicchetti, 1994; Shrout & Fleiss, 1979). ICCs were as follows and ranged from good to excellent: Parent: secure base: 0.92, p < .001; hostility: 0.90, p < .001; teen: secure base: 0.92, p < .001; hostility: 0.83, p < .001.

Data Analysis

Overview

We used R (R Core Team, 2021) to derive descriptive statistics and conduct preliminary analyses. ICCs were calculated in SPSS 28. Parents’ score on the TTI domain of behavioral conflict had high kurtosis (2.65). A square-root transformation reduced the skewness. We then ran all analyses with the original and square-root transformed versions; all patterns remained the same. For ease of interpretability, all analyses used the untransformed version.

APIM

APIMs accounted for the non-independence of dyads and tested hypotheses about the interdependence of sources of parental knowledge, conflict domains, and behavior during a conflict discussion task (Kenny et al., 2006). APIMs tested the unique association between adolescent and parent characteristics on their own outcome (“actor effect”) and on the outcome of the other member of the dyad (“partner effect”; Figure S1). Actor effect estimates control for partner effects, and vice versa. Partner effects indicate interdependence (Kenny et al., 2006). Because dyad members differed by family role (i.e., adolescent vs. parent), they were treated as “distinguishable”, to test whether effects differ according to a dyad member’s role.

We used Structural Equation Modeling (SEM), the recommended analytic strategy for distinguishable dyads (Ledermann & Kenny, 2017; Stas et al., 2018), via a web-based R analysis using lavaan (“APIM_SEM”, available at https://apimsem.ugent.be/shiny/apim_sem/; Rosseel, 2012; Stas et al., 2018). Actor and partner effects were separate by role (adolescent vs. parent).

APIM_SEM uses Full Information Maximum Likelihood estimation method to account for missing data (e.g., dyad member is missing a score; Ledermann & Kenny, 2017). Effect sizes for actor and partner effects are provided as partial correlations. Goodness-of-fit statistics are not provided for APIM_SEM since these models are fully-saturated (i.e., perfectly fit the date; Ledermann & Kenny, 2017). Missing data were minimal: one parent was missing an age; one adolescent and one parent (separate dyads) were missing behavioral data due to a video issue.

First, we tested for the association between domains of conflict with parental knowledge and its sources (n = 87). Each of these three models included both interview-based domains of conflict (argue/fight and different beliefs). The model including both argue/fight and different beliefs controlled for the influence of the other conflict domain (i.e., the effect presented for different beliefs controls for the influence of argue/fight and vice versa). Next, we tested for the association between parental knowledge, adolescent disclosure, and parental solicitation with behavior during the conflict discussion task (hostility, secure base; n = 65). All models included the three covariates described below (sample, age, gender). For each model we conducted a test of distinguishability based on the dyad member’s role (parent or adolescent) by comparing a model with distinguishable dyads to indistinguishable dyads using a chi square statistic, to verify APIM analyses should treat dyad members as distinguishable.

Covariates

To evaluate whether to include covariates in the model, we tested whether the two recruitment samples differed on any variable of interest. For adolescents, age (t[82.7] = 2.01, p = .048, d = 0.43), knowledge (t[78.14] = −2.69, p = .009, d = −0.58), and disclosure (t[69.80] = −2.16, p = .035, d = −0.47) were significantly different by sample. For parents, TTI Different Beliefs scores were significantly different by sample (t[79.51] = 2.07, p = .041, d = 0.45). Therefore, recruitment sample group was included as a covariate in the APIM_SEM analyses. Parent and child age and gender were significantly related to the sources of parental knowledge variables and TTI; therefore, age and gender were also included as covariates in the models.

Results

The chi square test based on the dyad member’s role was significant for all models (p <= .001), indicating that participants are distinguishable by their role in the dyad (parent or adolescent). There was no effect of recruitment sample for any model. Associations with covariates are reported in the respective model’s table note. See supplemental file for figures.

APIM Analyses

Informant Reports of Conflict Domains with Sources of Parental Knowledge

Parental Knowledge.

For parents, there was a significant, negative actor effect: when parents reported greater different beliefs, they also reported less knowledge (Table 2; Figure S2). There was a significant, negative adolescent partner effect, such that when parents reported greater different beliefs, adolescents reported less parental knowledge. In sum, when parents reported greater different beliefs, both parents and adolescents reported less parental knowledge.

Table 2.

Actor-Partner Interdependence Model Estimates for the Relation between Conflict Domains (x) and Parental Knowledge (y) by Role of Adolescent and Parent (n = 87)

Effect	Role	Estimate	95% CI	p-value	$\widehat{\mathit{\beta }}\left(\mathit{o}\right)$	r
Different Beliefs
Intercept	Adolescents	4.01	[−13.54, 21.56]	0.654
Actor		−0.14	[−0.39, 0.11]	0.269	−0.06	−0.05
Partner		−0.27	[−0.46, −0.07]	0.007	−0.32	−0.33
Intercept	Parents	42.15	[39.88, 44.42]	<0.001
Actor		−0.27	[−0.41, −0.13]	<0.001	−0.33	−0.39
Partner		0.01	[−0.16, 0.18]	0.903	0.01	0.06
Argue/Fight
Intercept	Adolescents	4.01	[−13.54, 21.56]	0.654
Actor		−0.16	[−0.49, 0.17]	0.341	−0.14	−0.16
Partner		0.10	[−0.17, 0.37]	0.474	0.09	−0.15
Intercept	Parents	42.15	[39.88, 44.42]	<0.001
Actor		−0.02	[−0.20, 0.17]	0.866	−0.01	−0.27
Partner		−0.01	[−0.23, 0.22]	0.960	−0.01	−0.08

Note. R² for adolescent: 0.30; R² for parent: 0.28. For age, there was a significant effect on parental knowledge for adolescents (−2.03; p < .001) and parents (−0.19, p < .001). CI = confidence interval; $\widehat{\beta }\left(o\right)$ = a standardized estimate using the overall standard deviation across both parents and adolescents, which enables comparison of these estimates across parents and adolescents; r = represents the partial correlation which provides the effect size for individual actor and partner effects.

Parental Solicitation.

There were significant actor effects for both parents and adolescents for different beliefs and parental solicitation: a higher rating of different beliefs was associated with lower reports of parental solicitation (Table 3; Figure S3). There was a significant adolescent partner effect: parental reports of greater different beliefs were associated with adolescent reports of less parental solicitation.

Table 3.

Actor-Partner Interdependence Model Estimates for the Relation between Conflict Domains (x) and Parental Solicitation (y) by Role of Adolescent and Parent (n = 87)

Effect	Role	Estimate	95% CI	p-value	$\widehat{\mathit{\beta }}\left(\mathit{o}\right)$	r
Different Beliefs
Intercept	Adolescents	2.96	[−11.33, 17.25]	0.685
Actor		−0.26	[−0.46, −0.06]	0.013	−0.16	−0.11
Partner		−0.20	[−0.36, −0.05]	0.011	−0.30	−0.20
Intercept	Parents	21.90	[19.74, 24.06]	<0.001
Actor		−0.31	[−0.44, −0.18]	<0.001	−0.47	−0.38
Partner		0.02	[−0.14, 0.17]	0.845	0.02	0.08
Argue/Fight
Intercept	Adolescents	2.96	[−11.33, 17.25]	0.685
Actor		0.18	[−0.09, 0.44]	0.198	0.19	0.02
Partner		0.20	[−0.02, 0.42]	0.073	0.22	−0.03
Intercept	Parents	21.90	[19.74, 24.06]	<0.001
Actor		0.14	[−0.04, 0.32]	0.118	0.15	−0.11
Partner		0.01	[−0.20, 0.22]	0.945	0.01	−0.07

Note. R² for adolescent: 0.21; R² for parent: 0.25. For age, there was a significant effect on parental solicitation for parents (−0.14, p = 0.011). For gender, there was a significant effect on parental solicitation for adolescents (−1.21; p = 0.010) and parents (1.26, p = 0.020). CI = confidence interval; $\widehat{\beta }\left(o\right)$ = a standardized estimate using the overall standard deviation across both parents and adolescents, which enables comparison of these estimates across parents and adolescents; r=represents the partial correlation which provides the effect size for individual actor and partner effects.

Adolescent Disclosure.

Parents and adolescents had significant actor effects for different beliefs and youth disclosure: higher ratings of different beliefs were associated with less youth disclosure (Table 4; Figure S4). For adolescents, there was a significant partner effect: parent reports of greater different beliefs were associated with adolescent reports of less disclosure.

Table 4.

Actor-Partner Interdependence Model Estimates for the Relation between Conflict Domains (x) and Adolescent Disclosure (y) by Role of Adolescent and Parent (n = 87)

Effect	Role	Estimate	95% CI	p-value	$\widehat{\mathit{\beta }}\left(\mathit{o}\right)$	r
Different Beliefs
Intercept	Adolescents	1.10	[−10.67, 12.87]	0.855
Actor		−0.19	[−0.36, −0.02]	0.033	−0.21	−0.12
Partner		−0.19	[−0.32, −0.06]	0.004	−0.32	−0.35
Intercept	Parents	22.90	[21.13, 24.67]	<0.001
Actor		−0.17	[−0.27, −0.06]	0.002	−0.28	−0.41
Partner		0.02	[−0.11, 0.15]	0.710	0.04	−0.01
Argue/Fight
Intercept	Adolescents	1.10	[−10.67, 12.87]	0.855
Actor		0.02	[−0.20, 0.25]	0.850	0.03	−0.09
Partner		0.03	[−0.16, 0.22]	0.735	0.04	−0.20
Intercept	Parents	22.90	[21.13, 24.67]	<0.001
Actor		−0.08	[−0.22, 0.07]	0.309	−0.09	−0.32
Partner		−0.14	[−0.31, 0.04]	0.129	−0.23	−0.22

Note. R² for adolescent: 0.28; R² for parent: 0.32. For age, there was a significant effect on youth disclosure for adolescents (−1.13; p = 0.004), and parents (−0.12, p = .006). CI = confidence interval; $\widehat{\beta }\left(o\right)$ = a standardized estimate using the overall standard deviation across both parents and adolescents, which enables comparison of these estimates across parents and adolescents; r = represents the partial correlation which provides the effect size for individual actor and partner effects.

Sources of Parental Knowledge with Conflict Discussion Behavior

Parental Knowledge.

We observed a significant actor effect for adolescents. Adolescent reports of greater parental knowledge were related to greater adolescent secure base behavior (Table 5; Figure S5a). There were no significant actor or partner effects for hostility.

Table 5.

Actor-Partner Interdependence Model Estimates for the Relation between Parental Knowledge (x) and Conflict Behavioral Domains Secure Base (y; Model 1) and Hostility (y; Model 2) by Role of Adolescent and Parent (n = 65)

Effect	Role	Estimate	95% CI	p-value	$\widehat{\mathit{\beta }}\left(\mathit{o}\right)$	r	R 2
Model 1
Intercept	Adolescents	10.77	[5.33, 16.22]	<0.001			0.10
Actor		0.08	[0.00, 0.17]	0.045	0.59	0.21
Partner		0.01	[−0.11, 0.13]	0.889	0.02	−0.00
Intercept	Parents	3.81	[2.75, 4.87]	<0.001			0.14
Actor		−0.02	[−0.13, 0.10]	0.786	−0.05	−0.08
Partner		0.07	[−0.01, 0.15]	0.078	0.20	0.25
Model 2
Intercept	Adolescents	−1.54	[−6.22, 3.15]	0.521			0.11
Actor		−0.07	[−0.15, 0.00]	0.056	−0.65	−0.23
Partner		−0.01	[−0.12, 0.10]	0.888	−0.02	0.02
Intercept	Parents	4.02	[2.96, 5.08]	<0.001			0.12
Actor		−0.01	[−0.13, 0.12]	0.934	−0.02	0.04
Partner		−0.07	[−0.15, 0.02]	0.113	−0.19	−0.20

Note. For age, there was a significant effect on secure base for adolescents (0.40; p = 0.025) and parents (0.05, p = .039), and a significant effect on hostile behavior for parents (−0.08; p = 0.001). CI = confidence interval; $\widehat{\beta }\left(o\right)$ = a standardized estimate using the overall standard deviation across both parents and adolescents, which enables comparison of these estimates across parents and adolescents; r = represents the partial correlation which provides the effect size for individual actor and partner effects.

Parental Solicitation.

We observed actor effects for adolescents for both secure base and hostility. Adolescent reports of greater parental solicitation related to greater adolescent secure base behavior (Table 6; Figure S6a) and less adolescent hostile behavior (Figure S6b).

Table 6.

Actor-Partner Interdependence Model Estimates for the Relation between Parental Solicitation (x) and Conflict Behavioral Domains Secure Base (y; Model 1) and Hostility (y; Model 2) by Role of Adolescent and Parent (n = 65)

Effect	Role	Estimate	95% CI	p-value	$\widehat{\mathit{\beta }}\left(\mathit{o}\right)$	r	R 2
Model 1
Intercept	Adolescents	10.08	[4.95, 15.21]	<0.001			0.16
Actor		0.15	[0.05, 0.25]	0.003	0.77	0.34
Partner		−0.00	[−0.12, 0.12]	0.951	−0.01	−0.04
Intercept	Parents	3.68	[2.67, 4.70]	<0.001			0.15
Actor		0.04	[−0.09, 0.16]	0.551	0.09	0.07
Partner		0.08	[−0.02, 0.18]	0.099	0.19	0.20
Model 2
Intercept	Adolescents	−1.18	[−5.61, 3.26]	0.603			0.16
Actor		−0.12	[−0.21, −0.03]	0.009	−0.71	−0.31
Partner		−0.01	[−0.12, 0.10]	0.849	−0.03	0.02
Intercept	Parents	4.04	[3.02, 5.05]	<0.001			0.14
Actor		−0.01	[−0.14, 0.12]	0.874	−0.02	−0.02
Partner		−0.08	[−0.18, 0.02]	0.110	−0.19	−0.20

Note. For age, there was a significant effect on secure base for adolescents (0.35; p = 0.038) and parents (.05, p = .033), and a significant effect on hostile behavior for parents (−0.08; p = 0.001). CI = confidence interval; $\widehat{\beta }\left(o\right)$ = a standardized estimate using the overall standard deviation across both parents and adolescents, which enables comparison of these estimates across parents and adolescents; r = represents the partial correlation which provides the effect size for individual actor and partner effects.

Adolescent Disclosure.

We observed significant actor effects for adolescents for secure base and hostility. Greater adolescent-reported disclosure was related to greater adolescent secure base behavior (Table 7; Figure S7a) and less adolescent hostile behavior (Figure S7b). There were also significant partner effects for parents, indicating that greater adolescent-reported disclosure was associated with greater parental secure base behavior and less parent hostile behavior. In other words, adolescent reports of greater disclosure were associated with more secure base behavior and less hostile behavior for both adolescents and parents.

Table 7.

Actor-Partner Interdependence Model Estimates for the Relation between Youth Disclosure (x) and Conflict Behavioral Domains Secure Base (y; Model 1) and Hostility (y; Model 2) by Role of Adolescent and Parent (n = 65)

Effect	Role	Estimate	95% CI	p-value	$\widehat{\mathit{\beta }}\left(\mathit{o}\right)$	r	R 2
Model 1
Intercept	Adolescents	10.28	[5.08, 15.47]	<0.001			0.16
Actor		0.13	[0.02, 0.24]	0.017	0.52	0.25
Partner		0.07	[−0.08, 0.21]	0.376	0.14	0.10
Intercept	Parents	3.84	[2.80, 4.87]	<0.001			0.18
Actor		−0.01	[−0.16, 0.14]	0.906	−0.02	−0.04
Partner		0.13	[0.02, 0.23]	0.020	0.26	0.31
Model 2
Intercept	Adolescents	−0.91	[−5.50, 3.69]	0.698			0.14
Actor		−0.11	[−0.22, −0.01]	0.029	−0.54	−0.26
Partner		−0.02	[−0.16, 0.11]	0.736	−0.05	−0.04
Intercept	Parents	3.97	[2.95, 4.98]	<0.001			0.19
Actor		−0.02	[−0.17, 0.13]	0.773	−0.05	0.02
Partner		−0.14	[−0.25, −0.03]	0.013	−0.29	−0.32

Note. For age, there was a significant effect on secure base for adolescents (0.37; p = 0.030) and parents (.05, p = .047), and a significant effect on hostile behavior for parents (−0.08; p = 0.001). CI = confidence interval; $\widehat{\beta }\left(o\right)$ = a standardized estimate using the overall standard deviation across both parents and adolescents, which enables comparison of these estimates across parents and adolescents; r = represents the partial correlation which provides the effect size for individual actor and partner effects.

Discussion

We evaluated the dyadic association between parental knowledge of adolescents’ activities and whereabouts and conflict processes. Strengths of the investigation include multi-modal assessment, multiple informants, a diverse sample, and dyadic analyses. First, we found significant negative associations between the conflict domain of different beliefs about daily life topics with parental knowledge of adolescent activities and its sources: greater perceptions of discrepancies on beliefs about daily life topics were associated with less knowledge of youth’s activities (and vice versa) and the sources of this information. Second, reports of parental knowledge and its sources were related to observed behavior during a conflict discussion task as hypothesized—positive associations between sources of knowledge and secure base yet negative associations between sources of knowledge and hostility. Third, in several models we observed dyadic interdependence: parental knowledge and its sources with different beliefs about daily life topics, as well as youth disclosure of their activities with secure base and hostility behaviors.

Our study found that having different beliefs―but not actually arguing or fighting―about daily life topics was significantly, negatively associated with parental knowledge and its sources. This result aligns with a study showing when adolescents rated parental supervision and rules high, their ratings of conflict with parents were low (Hayes et al., 2004). For parental knowledge, we found a significant parent actor effect and significant adolescent partner effect with different beliefs: when parents reported more different beliefs from their adolescent, both parents and adolescents reported less parental knowledge of adolescents’ activities and whereabouts. For parental solicitation and adolescent disclosure, both parents’ and adolescents’ reports of different beliefs were associated with their own reports of parental solicitation and adolescent disclosure, as well as a partner effect of parents’ reports of differing beliefs with adolescents’ reports of parental solicitation and adolescent disclosure. Partner effects demonstrate interdependence for different beliefs on daily life topics with parental knowledge and its sources. Curiously, arguing or fighting about these same topics was not a significant conflict domain related to sources of parental knowledge; rather, it was reports of having different beliefs about daily life topics that was significantly related to parental knowledge and its sources. Prior work indicates that measurements of different beliefs of daily life topics can be considered a direct measurement of informant discrepancies in perceived family functioning (De Los Reyes, Salas, et al., 2013). As such, this investigation adds to the growing body of literature supporting the importance of assessing informant discrepancies in relation to parent-adolescent relationships (De Los Reyes & Ohannessian, 2016; De Los Reyes et al., 2019).

Theories of attachment processes during adolescence help frame the results regarding behaviors observed during the conflict discussion task. Parents’ ability to know what adolescents are doing when they are not with them is dependent on communication—either parents asking adolescents directly or adolescents voluntarily informing them—which in turn likely relies on the extent of their attachment (Koehn & Kerns, 2016, 2018). In the current study, all domains of parental knowledge and its sources were associated with observed attachment behavior during the conflict discussion task. For adolescent reports of parental knowledge, parental solicitation, and adolescent disclosure of activities, there were significant, positive relations with adolescent secure base use, indicating that when adolescents reported greater parental knowledge, parental solicitation, and their own disclosure about their activities, they exhibited behaviors demonstrating they were open to discussing their thoughts and feelings about the conflict topic, proactively problem-solve, and/or seeking help or advice from their parent. For adolescent reports of parental solicitation of information and adolescent disclosure of activities, there were significant negative links to adolescent hostile behavior. Finally, there was interdependence in reports of adolescent disclosure for both secure base and hostility: adolescent reports of their disclosure of activities was positively associated with parents’ secure base provision and negatively associated with parent hostile behavior. Therefore, these findings appear to support the assertion that dyad members’ ability to maintain relatedness in the context of conflict may facilitate parents’ awareness of what adolescents are doing when they are not together.

Youth disclosure of their activities to parents is one of the most important contributors to parental knowledge/monitoring, and the best source of information of their rule breaking (Criss et al., 2015; Keijsers et al., 2010; Kerr et al., 2010; Stattin & Kerr, 2000). Parent-youth relationship quality has been theorized to explain adolescent disclosure’s prominence. Our results confirm that adolescent disclosure is associated with the attachment indices of both secure base and hostility. In fact, adolescent disclosure of their activities was the only source of parental knowledge that had partner effects with conflict behaviors, demonstrating dyadic interdependence on sources of parental knowledge and attachment-related behavior during the conflict discussion. A supportive relationship in which adolescents feel comfortable speaking openly about conflict topics and obtaining parental advice may also be one in which adolescent disclosure occurs naturally, facilitating parental knowledge by positive relationship qualities.

Findings have been mixed regarding whether parental solicitation of information about adolescents’ activities is an effective strategy (Criss et al., 2015; Keijsers et al., 2010; Laird et al., 2010; Stattin & Kerr, 2000). Greater parental solicitation has been associated with more adolescent behavioral concerns, potentially in response to youth’s norm-breaking; however, in general, it is presumed the link may also be interpreted that adolescents do not like feeling controlled (Stattin & Kerr, 2000). Conversely, less parental solicitation (measured as supervision) was associated with greater adolescent behavioral concerns (Hayes et al., 2004). Our findings indicated that parental solicitation of information about adolescents’ activities was negatively associated with the conflict domain of different beliefs about daily life topics for both parents and adolescents, while for adolescents only their ratings of parental solicitation were negatively associated with their own hostile behavior and positively associated with their own secure base use behavior. That is, parent reports of soliciting information about adolescents’ activities was not significantly associated with parents’ own attachment-related behavior during the conflict discussion. A possible interpretation is that in the context of similar viewpoints about daily life and a parent-adolescent relationship characterized by secure base use/provision, parental solicitation about adolescents’ activities may be either perceived as positive parental interest as well as occurring naturally in a relationship where this information is already shared.

This investigation had several limitations. We used specialized statistical models in which family role was treated as a moderator and cannot accommodate additional moderators like gender or demographic characteristics. For example, nearly two-thirds of the sample identified as African American or Black, and there may be differences in parenting strategies and parental monitoring (e.g., Lopez-Tamayo et al., 2016) or in conflict domains, frequency and intensity, and familial perceptions between racial and ethnic groups (e.g., Ehrlich et al., 2011; Ehrlich et al., 2016; Smetana & Gaines, 1999). Cultural considerations of parental discipline practices and emotional expression/socialization are important to consider when interpreting the results (Reaume et al., 2022). Future work on parenting factors in larger samples should address the socio-cultural, structural, and systemic contexts that may contribute to parent-adolescent conflict and monitoring. The fact that several models evaluating behavior only had adolescent effects may be because the main study design prioritized eliciting an adolescent stress response, and not because the phenomenon is only important for adolescents (vs. parents). Our study was cross-sectional and not designed to test the causal associations or temporal ordering among these factors. Variable order in models was based on the temporal ordering of study tasks.

Our study evaluated parent-adolescent dyadic associations between parental knowledge of adolescent activities and conflict processes. Advancing understanding of parenting factors like obtaining knowledge of adolescents’ activities and conflict processes is crucial because of links to adolescent adjustment. Our findings affirm parental knowledge and its sources are related to dyads’ perceptions of different beliefs about daily life topics. Also, parental knowledge, adolescent disclosure, and parental solicitation were associated with attachment behaviors of secure base and hostility during a conflict task. We observed interdependence, confirming the influence dyad members have on one another. Future research should replicate and extend these findings using longitudinal designs, and alternative approaches to modeling rater discrepancies (e.g., polynomial regression, latent class analyses; De Los Reyes et al., 2023).