Same Family, Divergent Realities: How Triangulation Preserves Parents’ Illusory Harmony while Adolescents Navigate Interparental Conflicts

Abstract

Triangulation is a process in which a child is drawn into conflict between two parents, and is linked to adolescent psychological maladjustment. Although harmful, families may engage in triangulation due to its promotion of diverging realities in which youth become more attuned to interparental conflict, yet parents are distracted from tension within their interparental relationship. Although central to theoretical depictions of triangulation and carrying robust implications for family science and prevention, the phenomenon of diverging realities in triangulating families has received inadequate empirical evaluation. This study utilized data collected from 150 families in which one parent and one adolescent completed baseline surveys and 21 daily diary questionnaires on triangulation, interparental conflict (IPC), and family cohesion. Multi-level models were applied, nesting days within families, to evaluate within-family associations between triangulation and divergent perspectives of family functioning. Results from multi-level models indicated that on days when adolescents experienced elevated triangulation, discrepancies between adolescent and parent reports of IPC and family cohesion increased, with adolescents reporting significantly higher levels of interparental conflict and lower levels of family cohesion relative to their parents. Further probing of the trends driving these discrepancies yielded a distinct pattern of results for IPC and family cohesion. Adolescent involvement in IPC is associated with more negative perspectives of family functioning relative to parents. These findings imply a mechanism through which triangulation confers risk to adolescents, and highlight that divergence in parent and adolescent perspectives of family functioning fluctuates depending on daily processes.

In the family context, triangulation refers to the involvement of a child in parental conflict as a means to defuse tension within the interparental subsystem (Bowen, 1978; Buchanan & Waizenhofer, 2001; Minuchin, 1974; Nichols & Everett, 1986), and generally includes parents’ pressure on children to become involved as well as children’s self-directed attempts to intervene (Bell, Bell, & Nakata, 2001; Buchanan & Waizenhofer, 2001; Kerig, 1995). Research to date has identified triangulation as a source of development risk; children who become involved in their parents’ conflicts experience increased internalizing and externalizing problems (Gerard, Buehler, Franck, & Anderson, 2005; Grych, Raynor, & Fosco, 2004), and emotional reactivity to interparental conflict (Buehler & Welsh, 2009). Triangulation also mediates the link between interparental conflict (IPC) and child maladjustment outcomes (Buchanan & Waizenhofer, 2001; Fosco & Grych, 2008; Grych et al., 2004), indicating that children’s involvement in interparental conflict is a mechanism through which risk from IPC is conferred. Beyond individual adjustment, triangulation also undermines family relationships, drawing on evidence of relations with increased hostility and diminished closeness between parents and adolescents (Fosco & Grych, 2010; Fosco, Lippold, & Feinberg, 2014).

Triangulation in the Home: Divergent Realities and Illusory Harmony

Considering the intuitive and documented harms attributed to triangulating children in IPC, it may be surprising that approximately one-third of families of adolescents are marked by elevated triangulation (Fosco & Bray, 2016). Attempts to understand why triangulation occurs in families have led family theorists to reflect on the function it may have within the family system. Triangulation is thought to be a family’s response to IPC to maintain homeostasis in family harmony (Minuchin, 1974; Minuchin, 1985). Involving children in IPC may serve to defuse the tension between parents who struggle to resolve conflicts on their own (Bowen, 1978; Buchanan & Waizenhofer, 2001; Minuchin, 1974; Nichols & Everett, 1986). Thus, this triadic process may provide temporary stability (Bowen, 1978; Kerig, 1995), maintaining the couple subsystem “in an illusory harmony” (Minuchin, 1974, p. 102) in which couples perceive stability in their relationship while the underlying sources of relational conflict remain unaddressed. Children in these families, however, are thought to be well-aware of the undercurrent of tension in the home (Bell, Bell, & Nakata, 2001; Minuchin, 1974). Indeed, children who are triangulated tend to become more sensitive to later parental conflicts (Buehler & Welsh, 2009), and develop a belief that they are responsible for causing and resolving disagreements as they arise (Fosco & Grych, 2010). The result may be that over time youth in triangulating families become sensitized to the family signals of impending conflicts and engage in efforts to defuse the conflicts while in the early stages of escalation, or perhaps before the parents are even aware of a disagreement that is brewing (Grych, 1998; Davies, Myers, Cummings, & Heindl, 1999; Schermerhorn et al., 2007). Such dynamics have the potential to stabilize into a family’s recurring pattern as parents come to rely on the involvement of their child in defusing couple conflicts (Minuchin, 1974). In families with established interactional dynamics that rely on triangulation to modulate interparental conflict, there may be systematically discrepant perceptions of the family. This divergent realities hypothesis suggests that there may be two lived experiences in the same home: parents who benefit from an “illusory harmony” and perceive stability in their family, and adolescents who feel embroiled in the intermittent conflicts of their parents.

Divergent perspectives of family functioning have direct implications for family and adolescent well-being. Prior work indicates that discrepancies in parent and adolescent perceptions of the family often reveal problems in the quality of social support, communication, and awareness of adolescents’ experiences (Carlson, Cooper, & Spradling, 1991; De Los Reyes, Ohannessian, & Racz, 2019; Ferdinand, van der Ende, & Verhulst, 2004; Padilla, McHale, Rovine, Updegraff, & Umaña-Taylor, 2016). Longitudinal studies point to elevated parent-adolescent discrepancies in perceptions of family functioning as predicting poorer psychological adjustment in adolescents, as well as lower self-competence and increased delinquent behavior (Carlson et al., 1991; De Los Reyes, 2011; Gaylord, Kitzmann, & Coleman, 2003; Human, Dirks, DeLongis, & Chen, 2016; Ohannessian & De Los Reyes, 2014; Ohannessian, Lerner, Lerner, & Von Eye, 1995; Ohannessian, Lerner, Lerner, & Von Eye, 2000). In families marked by divergent perspectives of family functioning, adolescent’s negative perceptions, relative to their parents, have been identified as stronger predictors of subsequent adjustment (Human et al., 2016). Given the potential for evaluation of discrepancies to provide unique insight into key domains of family functioning and subsequent developmental outcomes, their application in study of triangulation is ideal for capturing divergent perspectives of IPC proposed by the theoretical literature (e.g., Minuchin, 1974).

In Search of Empirical Support for Triangulation and Divergent Realities

Despite general acceptance that triangulation may promote illusory harmony and the developmental and family implications of divergent realities, there is a dearth of empirical inquiry into the links between triangulation and parent-adolescent discrepancies in perceptions of family relationships, such as IPC or family cohesion. Bell, Bell, & Nakata (2001) evaluated patterns of convergence and divergence on measures of family functioning between mothers, fathers, and adolescents, findings that parents’ tendencies to avoid marital tensions were associated with unbalanced family structures resembling parent-adolescents coalitions, scapegoating, and child mediation of the interparental relationship. These results offer promising empirical support for the connection between parents’ general approach to interparental conflict management and unbalanced or discrepant views of family functioning. Beyond this study, the potential for triangulation to promote parent and adolescents’ divergent perspectives of IPC remains untested. Probing for such discrepancies in perceptions of IPC may provide evidence for the theoretically proposed function of triangulation, while illuminating a key link between adolescents’ direct involvement in IPC and developmental maladjustment outcomes.

Triangulation may drive divergent perceptions in domains of family functioning beyond interparental conflict. We sought to evaluate whether triangulation serves to maintain illusory harmony not only in the couple relationship, but also in the family as a whole. Thus, we applied a divergent realities perspective to family-level cohesion, the degree of connectedness, supportiveness, and emotional bonding within a family (Moos, 1974; Olson, Sprenkle, & Russell, 1979). Perception of these broader family qualities may be shaped by one’s role in triangulation. Parents who triangulate as a means to defuse interparental conflict tension may maintain views of the family as connected and supportive; when interparental conflict tensions rise, the family’s response of involving the child serves to restore balance to the family (Minuchin, 1974; Nichols & Everett, 1986). By contrast, a child who is routinely involved in their parent’s disputes may come to view the associated strain and discord as more salient characteristics of the family, diminishing their perception of overall cohesion. Drawing on this divergent realities perspective, parents are likely to perceive the family as more cohesive than triangulated youth.

The unexplored possibility that triangulation may drive divergent realities across key domains of family functioning carries robust implications for family research. Studies often rely on measurements from a single reporter or an average of parent and youth reports to capture a singular family “reality” reflecting the quality of family functioning (Gaylord et al., 2003). Such reports would be problematic in families which tend to experience high levels of divergence between members. Understanding whether discrepancies emerge in triangulating families would underscore the need for including measures of triangulation in family risk assessments, understanding multiple perspectives in the family, and potentially developing intervention content that can rectify discrepancies in the perceived nature of family relationships. Such efforts would be key in promoting improved understanding of developmental risk and serving youth in triangulating families.

Family scientists have long been challenged with modeling the dynamic, recurrent family patterns outlined by theorists (Cox & Paley, 1997; Minuchin, 1985), among which triangulation in particular is viewed as a cyclical pattern of dysfunction which emerges to regulate interparental discord (Minuchin, 1974; Nichols & Everett, 1986). Daily diary methods provide opportunities to capture within-family change processes that address limitations inherent in mean-level, between-family comparisons by facilitating evaluation of daily fluctuations in triangulation and accompanying fluctuations in the degree to which family members’ views of family relationships diverge (Bolger & Laurenceau, 2013). Through frequent assessments of short periods of time (e.g., one day), it is possible to minimize retrospective bias and capture more subtle changes in family functioning as they are experienced by family members (Bolger & Laurenceau, 2013; Bolger, Davis, & Rafaeli, 2013; Schwarz, 2007). Moreover, contemporary methods are capable of disentangling within-person and between-person processes (Bolger & Laurenceau, 2013). As a result it is possible to consider how within-family processes (i.e., evaluating parent and adolescent perspectives of the family on days of high and low triangulation) may have between-family differences as a function of families’ general tendencies to triangulate adolescents. Thus, daily diary methods represent an ideal approach for evaluating potential divergence in parents and adolescents’ perceptions of family functioning on days with elevated triangulation, while also accounting for the influence of a family’s general pattern of triangulation.

The Present Study

The current study leverages daily diary methods to evaluate whether there are within-family effects in which triangulation corresponds to changes in parent-adolescent discrepancies in daily reports of interparental conflict and family cohesion. Consistent with a divergent realities hypothesis, we expected that on days when adolescents are triangulated into parental conflicts, there would be increases in parent-adolescent discrepancies in their perceptions of interparental conflict, with higher reports of IPC from adolescents relative to parents (H1a). Taking this further, we expected that a more established pattern of triangulation may moderate the effect of daily instances of triangulation on divergent perspectives of the family. Specifically, we hypothesized that higher global family ratings of triangulation (i.e., reporting higher frequency of triangulation on an established baseline measure) would amplify the within-family association between instances of triangulation and parent-adolescent discrepancies (H1b).

In a second set of analyses, we evaluated whether a divergent realities process would be observed in experiences of family cohesion. Specifically, we hypothesized that daily instances of triangulation in families would correspond to increases in parent-adolescent discrepancies in perceived family cohesion, with higher reports of family cohesion from parents relative to adolescents (H2a). Additionally, we tested whether the daily effect would be moderated by global triangulation, predicting that within-family association would be more robust in families that have more established patterns of triangulation (H2b).

Across both sets of hypotheses, we planned to explore the nature of divergent realities (H1c, H2c). If triangulation is associated with changes in parent-adolescent divergence, we probed the findings to determine whether the pattern of results fit with an “illusory harmony” or “differential disruption” perspective. We conceptualized illusory harmony as a pattern of findings in which triangulation is associated with changes in parent-adolescent discrepancies because of changes in adolescents’ perceptions of family relationship quality (IPC and cohesion, H1c and H2c, respectively), but not with changes in parents’ perceptions of family relationships. A differential disruption perspective would be evident if both parents’ and adolescents’ perceptions of family relationships change in relation to triangulation; however, adolescents’ perceptions change to a greater degree than parents’.

Method

Data were collected as part of the Penn State Family Life Optimizing Well-being (FLOW) study, during which parents and adolescents completed baseline surveys followed by 21 consecutive brief daily reports on family functioning.

Participants

Participants included 150 families with an adolescent in either 9^th or 10^th grade and were recruited through Pennsylvania high schools and referrals from participating families.

In order to be eligible, families were required to meet the following criteria: 1) two-caregiver household status, 2) adolescents resided continuously within a single household , 3) access to internet and ability to complete daily surveys at home, 4) English fluency, 5) the participating adolescent was in either 9^th or 10^th grade at the beginning of the study, 6) participating parent and adolescent agreed to participate (via consent and assent, respectively).

Participating caregivers included 143 females (95%) and 7 males (5%) who were between 30 and 61 years old (M_age = 43.4, SD_age = 6.9), and identified themselves as their adolescents’ mother (92.7%), stepmother (1.3%), aunt (0.7%), foster mother (0.7%), or father (4.6%). Caregivers were primarily White/Caucasian (90%), Black/African American (2.7%), Asian (3.3%), Native American/American Indian (0.7%), Latinx (0.7%), or multi-racial (2.0%). The majority of parents were married (89%), and all lived with a second caregiving adult. Caregivers reported having lived together for 18 years on average (SD = 7.2). Annual family income ranged from less than $10,000 to $125,000 and over, with a median income of $70,000-$79,999.

Participating adolescents were 59% female (n = 89) and 41% male (n = 61), between the ages of 13 and 16 years (M_age = 14.63, SD_age = 0.82), and were identified via parent report as White (83.3 %), African American/Black (4.7%), Native American/American Indian (0.7%), Asian (4.7%), Latinx (0.7%), and multi-racial (5.3%).

Procedure

Families were recruited by emails sent by school principals inviting parents to participate in the study. Parents interested in participating then accessed a study web-site from which they obtained information regarding design and purpose of the study. If the family met all eligibility inclusion criteria, adolescents were emailed study information and given an opportunity to assent to study procedures. If the adolescent assented, they completed a baseline questionnaire; upon completion parents would be sent a link to complete their own baseline questionnaire. When both parent and adolescent completed baseline surveys, links for the daily diary questionnaires were sent to the parent and adolescent in each household each night at 7pm for 21 consecutive nights. Daily questionnaires were identical each day and designed to be completed in 5 minutes or less. Parent compliance on daily surveys was 96.52% (M_Parent = 20.3), whereas adolescent compliance was 90.42% (M_Adolescent = 19.0), yielding 2,848 total observations for data analysis. Upon completion of baseline and 21 daily surveys, all participants were compensated with a choice of gift cards (Amazon.com or Walmart). Each participant received $25 after completing the baseline survey. For the daily surveys, participants received $2.50 for the first four surveys of each week and $5.00 for the final three surveys of each week. Families were able to earn up to $200 for completing all baseline and daily diary surveys. All study procedures were approved and conducted under the supervision The Pennsylvania State University Institutional Review Board (STUDY00000472).

Measures

Study analyses utilized adolescents’ baseline reports of global triangulation, and daily reports of triangulation, interparental conflict, and family cohesion, as well as parents’ daily reports of interparental conflict, and family cohesion. Within-person (R_c) and between-person (R_1F) reliability scores were calculated for scales with multiple items to assess whether scales reliably captured within and between-person change (Bolger & Laurenceau, 2013; Cranford, Shrout, Iida, Rafeali, Yip, & Bolger, 2006).

Daily Triangulation.

Adolescents rated that day’s level of triangulation into interparental conflict relative to how they usually felt using a slider scaled −5 (“Less than Usual”) to 5 (“More Than Usual”). Adolescents responded to three items, “How much did you feel CAUGHT IN THE MIDDLE OF [parent 1] and [parent 2]’s disagreements today”, “How much did you feel INVOLVED in your [parent 1] and [parent 2]’s disagreements today?”, and “How much did your [parent 1] or [parent 2] want you to BE ON THEIR SIDE today?” Values from the three items were averaged to create a single score for daily triangulation. Daily scores ranged from −5 to 5 (M = −2.11, SD = 2.39), and demonstrated meaningful within-person (R_c = 0.85) and between-person (R_1F = 0.96) variation.

Daily IPC Discrepancies.

Adolescents and parents rated that day’s level of IPC by responding to one item on a slider scaled 0 (“Not At All”) to 10 (“A Lot”). Adolescents responded to the item, “My [parent 1] and [parent 2] were angry or mad at each other today”, and parents responded to the item, “My partner and I were angry or mad at each other today”. Daily scores for adolescent reports of interparental conflict ranged from 0 to 10 (M = 0.92, SD = 2.04), while parent scores ranged from 0 to 10 (M = 0.87, SD = 1.95). Intraclass correlations for adolescent (ICC = .46) and parent (ICC = .25) reports of IPC indicated approximately 54% and 75% of the variance was evident at the within-family level, respectively. Daily discrepancies scores were then calculated by subtracting day’s parent report of IPC from day’s adolescent report of IPC, and ranged from −10 to 10 (M = .06, SD = 2.26). Positive values indicate days on which an adolescent’s report of IPC was greater than their parent’s report. Intraclass correlations for interparental conflict discrepancies (ICC = .28) suggested a meaningful degree of within-family variation (72%), demonstrating sufficient within and between-family variation for multi-level analysis.

Daily Family Cohesion Discrepancies.

Adolescents and parents responded to 5 items regarding that day’s level of family cohesion, using a slider scaled 0 (“Not At All”) to 10 (“A Lot”). An example item was “Family members really helped and supported one another.” Items were averaged to create a single daily score. Daily scores for adolescent family cohesion ranged from 0 to 10 (M = 7.87, SD = 2.41), while parent scores ranged from 0 to 10 (M = 7.83, SD = 2.18). Daily family cohesion demonstrated meaningful within-person and between person variation for adolescents (R_c = 0.76; R_1F = 0.92; ICC = .71) and parents (R_c = 0.82; R_1F = 0.88; ICC = .57).

Daily discrepancies for family cohesion were calculated by subtracting day’s adolescent report from day’s parent report, and ranged from −10 to 9.6 (M = −.03, SD = 2.54). Positive values indicate days on which a parent’s report of family cohesion was greater than their parent’s report. Intraclass correlations for family cohesion discrepancies (ICC = .59) suggest a meaningful degree of within-family variation (41%), demonstrating sufficient within and between-family variation for multi-level analysis.

Global Triangulation.

Adolescents’ general patterns of triangulation into interparental conflict was measured at baseline as an average of five items from the Children’s Perceptions of Interparental Conflict Scale, a commonly used scale with sound psychometric properties with adolescents (Grych et al., 1992; Grych et al., 2004). Adolescents indicated their agreement with statements describing patterns of triangulation into interparental conflict on a 5-point scale (1= Strongly Disagree to 5 = Strongly Agree). Scores ranged from 1 to 5 (M = 1.79, SD = 0.79). Item statements were contextualized as occurring during conflict, allowing measurement of global patterns of triangulation. A sample item was “I feel caught in the middle when my parents argue”. Cronbach’s alpha was 0.84.

Data Analysis

Multi-level models were constructed with days (Level 1) nested within families (Level 2) to permit evaluation of within-family variance in daily triangulation. Between-family variables include global triangulation, adolescent age, and adolescent gender provided during the baseline assessment, and are time in-variant within the current study (Bolger & Laurenceau, 2013). All between-family variables were grand mean centered (i.e., adolescent age, gender, and global triangulation). Within-family variables include adolescents’ daily reports of triangulation, adolescent and parents’ daily reports of interparental conflict and family cohesion, as well as daily IPC discrepancies and daily family cohesion discrepancies. Adolescents were instructed to rate daily triangulation as compared to usual levels of triangulation, and thus reported scores already represented daily deviation.

We conducted separate multi-level models for IPC discrepancies, adolescent-reported IPC, parent-reported IPC; and family cohesion discrepancies, adolescent-reported cohesion, parent-reported cohesion). All models controlled for adolescent gender, age, and day in the study.

Interparental Conflict Models.

Four models with interparental conflict as the outcome were constructed to test three hypotheses (i.e., H1a-H1c). The first model was for H1a - day’s triangulation and global triangulation (i.e., two main effects) and control variables (i.e., adolescent age, gender, day in study) were included in the model to evaluate the main effect of daily triangulation and global triangulation on the outcome variable, accounting for the effect of control variables (Bolger & Laurenceau, 2013):

Level 1:

$$IPC\_Dis{c}_{it}= {\beta }_{0i}+ {\beta }_{1i}Day\text{'}sTriangulatio{n}_{it}+ {\beta }_{2i}Da{y}_{it}+e_{it}$$

Level 2:

$${\beta }_{0i}= {\gamma }_{00}+ {\gamma }_{01}GlobalTriangulatio{n}_i+ {\gamma }_{02}Gende{r}_i+ {\gamma }_{03}Ag{e}_i+ u_{0i}$$

$${\beta }_{1i}= {\gamma }_{10}+u_{1i}$$

$${\beta }_{2i}= {\gamma }_{20}$$

where $IPC\_Dis{c}_{it}$ represented the discrepancy between adolescent and parent report of interparental conflict in family i on day t. At level 1 (day-level variables) $Day\text{'}sTriangulatio{n}_{it}$represented adolescent reports of triangulation in family i on day t respectively, and Day_i represented time centered to the midpoint of the study (day 10.5) for family i. At level 2 (family-level variables) GlobalTriangulation_i, Gender_i, and Age_i represented levels of baseline global triangulation, adolescent gender, and adolescent age for family i, respectively.$u_{0i}$ and $u_{1i}$ represented the random effects, while $e_{it}$ represented a within-family error component that was allowed to auto-correlate (AR1). Fixed effects for different components in equations were represented by γs. Specifically,${\gamma }_{00}$ represented the intercept for IPC discrepancy;${\gamma }_{10}$ represented the effect of day’s triangulation on outcome;${\gamma }_{20}$ represented the effect of day in study on IPC discrepancy; ${\gamma }_{01}$, ${\gamma }_{02}$ ,${\gamma }_{03}$ represented the effects of global triangulation, adolescent gender, and adolescent age on IPC discrepancy, respectively.

The second model was for H1b. In this model, a cross-level interaction between global triangulation and day’s triangulation was introduced to the previous model to evaluate the effect of global triangulation on daily associations between day’s triangulation and IPC discrepancies:

Level 1:

$$IPC\_Dis{c}_{it}= {\beta }_{0i}+ {\beta }_{1i}Da{y}^{\prime }sTriangulatio{n}_{it}+ {\beta }_{2i}Da{y}_{it}+e_{it}$$

Level 2:

$${\beta }_{0i}= {\gamma }_{00}+ {\gamma }_{01}GlobalTriangulatio{n}_i+ {\gamma }_{02}Gende{r}_i+ {\gamma }_{03}Ag{e}_i+ u_{0i}$$

$${\beta }_{1i}= {\gamma }_{10}+{\gamma }_{11}GlobalTriangulatio{n}_i+ u_{1i}$$

$${\beta }_{2i}= {\gamma }_{20}$$

where ${\gamma }_{11}$ represented the interaction effects of global triangulation by day’s triangulation on IPC discrepancy. All other terms remained identical to those listed in step one.

The third and fourth models were constructed to evaluate the “illusory harmony” and “differential disruption” perspectives outlined in H1c. In these two models, all terms are identical to those in the first model, except for the daily IPC discrepancy outcome variable was replaced by adolescent (Adolescent_IPC_it) and parent (Parent_IPC_it) daily reports of interparental conflict, respectively.

Family Cohesion Models.

Four models addressing the family cohesion outcomes were specified and evaluated in the same format as those described for the interparental conflict outcome. Cohesion_Disc_it, Adolescent_Cohesion_it, Parent_Cohesion_it outcome terms were substituted for IPC_Disc_it, Adolescent_IPC_it, and Parent_IPC_it outcome terms in the models presented above, respectively.

Results

Descriptive statistics and correlations for all study variables are presented in Table 1.

Table 1.

Descriptive Statistics and Bivariate Correlations

						Correlations
		1	2	3	4	5	6	7	8	9	10
1	Ad Gender	--	--	--	--	--	--	--	--	--	--
2	Ad Age	.05	--	--	--	--	--	--	--	--	--
3	Global Tri	−.01	.02	--	--	--	--	--	--	--	--
4	Daily Tri	.09	.08	.04	--	.17	.05	−.09	−.04	.05	.04
5	Ad IPC	.00	.01	.40	.04	--	.32	−.24	−.17	.48	.05
6	P IPC	.06	.20	.24	.02	.47	--	−.12	−.25	−.66	−.09
7	Ad FCoh	.02	−.09	−.36	−.14	−.34	−.29	--	.23	−.04	−.53
8	P FCoh	−.04	−.13	−.20	−.07	−.24	−.23	.47	--	.13	.62
9	IPC Disc	−.06	−.18	.17	.02	.54	−.49	−.06	−.02	--	.13
10	FCoh Disc	−.06	−.02	.19	.08	.13	.09	−.61	.42	.05	--
	M	--	14.63	1.79	−2.11	0.92	0.87	7.87	7.83	0.06	−.03
	SD	--	0.82	0.79	2.39	2.04	1.95	2.41	2.18	2.26	2.54

Note. N = 150; M = mean; SD = standard deviation; Ad = adolescent; P = parent; Tri = triangulation; IPC = interparental conflict; FCoh = family cohesion; Disc = discrepancy. Between-person correlations (below diagonal) are calculated using individual mean scores across 21 days, except for global triangulation, adolescent gender, and adolescent age which used the baseline assessments. Between-person correlations .17 or greater were statistically significant (p<.05). Within-person correlations (above diagonal) reflect a sample average of within-person correlations between daily variables; no tests of statistical significance are available.

Interparental Conflict Discrepancies

The first set of analyses pertaining to parent-adolescent discrepancies in daily IPC are depicted in Table 2. Families with higher global triangulation also tended to have more discrepant parent and adolescent perceptions of IPC (${\gamma }_{01}$ = .15, p < .01). Relevant to Hypothesis 1a, there was a statistically significant within-family finding for day’s triangulation corresponding to changes in IPC discrepancies (${\gamma }_{10}$ = .16, p < .01). Specifically, on days when triangulation was higher than usual, discrepancies in parents’ and adolescents’ perception of IPC increased. Moreover, the positive coefficient indicated that adolescents’ ratings of IPC were higher than those of their parents, as expected in Hypothesis 1a.

Table 2.

Interparental Conflict and Family Cohesion Models

	Interparental Conflict Models				Family Cohesion Models
	Discrepancy main effect (H1a)	Discrepancy interaction effect (H1b)	Ad. report main effect (H1c)	Parent report main effect (H1c)	Discrepancy main effect (H2a)	Discrepancy interaction effect (H2b)	Ad. report main effect (H2c)	Parent report main effect (H2c)
	$\beta$ (SE)				$\beta$(SE)
Intercept (${\gamma }_{00}$)	.00 (.04)	.00 (.04)	.05 (.05)	.02 (.05)	.00 (.06)	−.03 (.06)	−.02 (.06)	−.01 (.06)
Gender$({\gamma }_{02}$)	−.05 (.04)	−.05 (.04)	−.05 (.04)	.03 (.04)	−.12 (.06)	.07 (.06)	.07 (.06)	−.07 (.06)
Age$({\gamma }_{03}$)	−.05 (.04)	−.04 (.04)	−.03 (.04)	.04 (.04)	−.04 (.06)	.00 (.06)	.00 (.06)	−.04 (.06)
Day (${\gamma }_{20}$)	.00 (.02)	.00 (.02)	−.02 (.01)	−.03 (.02)	−.02 (.01)	.02 (.01)	.02 (.01)	−.00 (.01)
Global_Tri $({\gamma }_{01})$	.15 (.04) **	.19 (.04)**	.20 (.04)**	.10 (.04)*	.15 (.06)*	−.36 (.06)**	−.33 (.06) **	−.21 (.06)**
Day’s_Tri $({\gamma }_{10})$	.16 (.04)**	.15 (.04)**	.25 (.05)**	.06 (.04)	.08 (.04)*	−.15 (.04)**	−.16 (.03)**	−.09 (.03)**
Day’sTri*Global_Tri$({\gamma }_{11}$)	--	.10 (.04)**	--	--	--	−.05 (.03)	--	--

Note.

^*p < .05;

^**p < .01.

Tri = triangulation; Ad. = adolescent.

In the second model, the cross-level interaction between global triangulation and day’s triangulation was added to test hypothesis 1b. The cross-level interaction term was significant (${\gamma }_{11}$ = .10, p < .01), indicating that the association between day’s triangulation and IPC discrepancies was stronger in families with higher levels of global triangulation. To probe this interaction, we examined simple slopes for day’s triangulation at high (+1 SD) and low (−1 SD) values of global triangulation. As seen in Figure 1a, in families with higher levels of global triangulation (+1 SD), day’s triangulation was more strongly associated with increases in IPC discrepancies ($\beta$ = .25, p < .01). However, in families with lower levels of global triangulation (−1 SD), daily variation in triangulation did not correspond to IPC discrepancies ($\beta$ = .05, ns). Regions of significance tests indicated that day’s triangulation was statistically significantly associated with IPC discrepancies when standardized baseline levels of triangulation were greater than values of −0.60 (n = 97; 65% of sample; Figure 1b).

Figure 1.

N = 150. Global triangulation moderates the association between daily triangulation and IPC discrepancies. 1a presents simple slopes for day’s triangulation and day’s IPC Discrepancy at high (+1 SD) and low (−1 SD) levels of global triangulation. 1b presents the region of significance for the simple slope. Daily triangulation is associated with increases in IPC discrepancies at values of global triangulation above −0.60 SD (n = 97; 65% of sample).

To better understand these findings, we re-computed models predicting adolescent-reported and parent-reported IPC, respectively. Across both models, between-family effects were similar: families with higher triangulation also tended to have higher average levels of IPC (adolescent:${\gamma }_{01}$ = .20, p < .01; parent:${\gamma }_{01}$ = .10, p < .05). Turning to within-family effects germane to questions of illusory harmony or differential disruption, different findings emerged for adolescents and parents. In the adolescent-report model, triangulation was associated with IPC. On days when adolescents experienced elevated levels of triangulation, adolescents reported statistically significant increases in IPC (${\gamma }_{10}$ = .25, p < .01). However, in the parent model, the association between day’s triangulation and IPC was not statistically significant. Collectively, results from these two models support the illusory harmony perspective over the differential disruption perspective.

Family Cohesion Discrepancies

Families with higher global triangulation tended to have more discrepant parent and adolescent perceptions of family cohesion (${\gamma }_{01}$ = .15, p < .01). Regarding Hypothesis 2a, there was also a statistically significant within-family finding for day’s triangulation corresponding to changes in family cohesion discrepancies (${\gamma }_{10}$ = .08, p < .05). On days when triangulation was higher than usual, discrepancies in parents’ and adolescents’ perceptions of family cohesion increased. As expected in hypothesis 2a, parents’ ratings of family cohesion were higher than those of adolescents.

In the second model, a cross-level interaction between global triangulation and day’s triangulation was added to test hypothesis 2b. This interaction term was non-significant, indicating that the association between day’s triangulation and family cohesion discrepancies was not moderated by level of global triangulation. Thus, H2b was not supported.

Once again we probed the pattern of discrepancies by re-computing models predicting adolescent-reported and parent-reported family cohesion, respectively. Between-family effects across both models were similar: families with higher triangulation also tended to have lower average levels of family cohesion (adolescent:${\gamma }_{01}$ = −.30, p < .01; parent:${\gamma }_{01}$ = −.21, p < .01). Evaluating within-family effects relevant to questions of illusory harmony or differential disruption also exposed a similar pattern; in both adolescent-report and parent-report models, triangulation was significantly associated with family cohesion. On days when adolescents experienced elevated levels of triangulation, adolescents reported statistically significant decreases in family cohesion (${\gamma }_{10}$ = −.16, p < .01) as did parents (${\gamma }_{10}$ = −.09, p < .01), offering support for the differential disruption perspective over the illusory harmony perspective.

Discussion

Insight into the function of triangulation in families has largely been guided by the clinical wisdom of family therapists, postulating that triangulating youth into IPC serves to help return balance in families with distressed interparental relationships (Bowen, 1978; Minuchin, 1974; Nichols & Everett, 1986). Recurrence and stabilization of this family pattern is thought to be driven by divergent perspectives of family functioning between triangulating parents and children; however, empirical evaluation of this fundamental aspect of triangulation is largely absent. The present study addressed these empirical gaps by leveraging daily diary methods to capture within-family change processes in triangulation and parent and adolescents’ divergent perspectives of IPC and family cohesion. In doing so, this study evaluated within-family dynamics associated with triangulation to determine whether it offers an “illusory harmony” to parents in conflict (Minuchin, 1974, p. 102).

We evaluated two sets of hypotheses regarding the influence of triangulation on divergent perspectives of IPC and family functioning. For each outcome, we hypothesized that a) daily fluctuations in triangulation would be associated with divergent perspectives of family functioning between adolescents and parents, and b) these daily associations would be of greater magnitude within families who reported greater levels of global triangulation. We then probed the degree to which these daily discrepancies reflected an illusory harmony or differential disruption perspective in families.

Same Family, Divergent Realities

Across both sets of analyses focusing on discrepancies in IPC and in family cohesion, a consistent message emerged. Daily experiences of triangulation correspond to increasingly divergent views of the family for parents and adolescents. Despite these common findings, unique aspects to IPC and family cohesion also emerged, providing evidence of illusory harmony in the context of IPC discrepancies and differential disruption in relation to family cohesion.

Interparental Conflict.

First we discuss interparental conflict. On days with elevated triangulation, adolescents reported significantly higher levels of IPC than their parents did. Furthermore, this daily link between triangulation and divergent perspectives of IPC was more robust within families which tended to engage in higher levels of triangulation on a regular basis. This pattern of findings aligns with prior empirical work which suggests that exposure to triangulation may sensitize adolescents to their parents’ conflict (e.g., Buehler & Welsh, 2009; Fosco & Grych, 2010), in this case promoting heightened perception of conflict intensity as patterns of IPC and triangulation recur. Finally, evaluation of the main effects of day’s triangulation on both adolescent and parent reports of IPC revealed the pattern driving discrepancies more closely resembled the illusory harmony perspective, rather than the differential disruption perspective. Elevated triangulation was accompanied by significant increases in reports of IPC from adolescents, but not their parents – suggesting that parents in triangulating families may experience a disconnect from their conflict as a result of their adolescent child’s involvement.

These findings collectively suggest that adolescent involvement in IPC may in fact function to defuse distress experienced in the interparental relationship. Adolescent involvement may thus be reinforced, as it yields temporary relief from the mounting tensions, thereby increasing the likelihood that the family will engage in a similar pattern during subsequent interparental disputes. Adolescents embroiled in such a pattern over time are likely to develop sensitivity to the cues and signals of impending conflict and becoming involved in ways which sustain “illusory harmony” between parents, and thus the likelihood that such dysfunctional triadic patterns will be maintained by the family system.

Family Cohesion.

Triangulation also was associated with discrepancies in family cohesion. As hypothesized, when families engaged in triangulation, discrepancies in parents’ and adolescents’ perceptions of family cohesion grew. However, in contrast to findings of illusory harmony related to IPC, both parents’ and adolescents’ perceptions of cohesion changed in relation to triangulation; however, adolescents experienced a more marked decrease in their perceptions of cohesion on those days. In addition, the degree to which parent-adolescent discrepancies in cohesion changed in relation to daily triangulation were not tempered by global family tendencies to triangulate adolescents. In other words, divergent perspectives of family cohesion were just as likely to occur in families in which triangulation was a seldom irregularity as it was in families with more frequent, well-established patterns of triangulation. Taken together, it seems that discrepancies in IPC and cohesion fluctuate in distinct ways in families. With regard to family cohesion, our findings generally supported a differential disruption perspective.

Understanding how and why these processes operate differently for IPC and cohesion discrepancies calls for thoughtful unpacking of theory in this new light. One possibility lies in parents’ perceptions that are potentially shaped by the manner in which triangulation occurs within a family. For example, triangulation may involve adolescent hostility or disruptive behavior to distract parents from conflicts (e.g., Davis et al., 1998), or a cross-generational coalition with the opposite parent that creates distance and tension in the parent-adolescent relationship (Christensen & Margolin, 1988; Haley, 1967). These processes, while serving to distract from IPC tension specifically, come at the cost of creating tension within other domains of the family (Fosco & Grych, 2010; Fosco, Lippold, & Feinberg, 2014). As a result, triangulation would foster an illusory harmony in the couple relationship, and lead parents to mis-identify problems in their child or in other relationships in the family (Minuchin, 1974). Thus, as a whole, these findings seem to support the function of triangulation as preserving parents’ sense of marital harmony, but coming at the cost of the broader family relationships.

Implications for family science

In addition to providing empirical support for family systems concepts, the current study yields broader implications for assessments within family science. Prior work has acknowledged that discrepant views of family functioning between parents and adolescents are indicative of dysfunction and also hold implications for adolescent development (De Los Reyes, 2019; Human et al., 2016). However, the present findings illustrate that families experience daily fluctuation in the degree to which individual members agree on key domains of family functioning. Crucially, within-family variability in agreement on such domains looks to be influenced by commonplace family level processes, such as triangulation (Fosco & Bray, 2016). Of further note is that triangulation was linked to discrepancies in reports of family cohesion across family types, even those who do not habitually involve their children in IPC. These trends suggest that family researchers should take care to collect data from multiple reporters when assessing and modeling family functioning to account for multiple perspectives. Furthermore, congruence between members is not guaranteed to be a fixed, time-invariant characteristic of families, and experiences which promote fluctuations in congruence should also be carefully evaluated.

Implications for family prevention

Triangulation is consistently associated with adolescent maladjustment (e.g., Buchanan & Waizenhofer, 2001; Grych, Raynor, & Fosco, 2004; Fosco, Lippold, & Feinberg, 2014). The present findings indicate that triangulating parents are less likely to be attuned to dimensions of family distress relative to their adolescent children who regularly become involved in their interparental disputes. As result, such families may continue engaging in dysfunctional patterns of interparental conflict resolution, placing their adolescents at further developmental risk. Family-based prevention programs have demonstrated efficacy in supporting parents’ constructive IPC behaviors, and in turn, improved adolescent well-being (e.g., Miller-Graff, Cummings, & Bergman, 2016). Programs targeting IPC may stand to benefit from assessing triangulation in addition to dyadic conflict and including content which supports families in recognizing and reducing maladaptive conflict patterns involving adolescents. Furthermore, evaluations of program efficacy have relied on parent report of adolescent adjustment (Miller-Graff, et al., 2016). Given that triangulation shapes discordant perspectives within the family, inclusion of adolescent reports of adjustment may strengthen evidence for program effects on developmental outcomes.

Limitations and Future Directions

The current study was not without limitations which should be considered and addressed in future research. First, participating families were primarily White and reported annual incomes comparable to national averages, limiting generalizability to more diverse families, in terms of race, family structure, sexual orientation, socio-economic status, and geographical location. Family processes identified in this study should be replicated in more diverse samples to determine whether IPC and triangulation are similarly experienced by parents and adolescents across cultural and demographic groups. Additionally, some debate exists regarding the optimal implementation of discrepancy scores within regression-based models to yield statistically reliable information (Laird & De Los Reyes, 2013). This topic has yet to be fully explored within a within-person format, thus the extent to which the solutions offered by Laird and De Los Reyes (2013) apply to multi-level models remains unclear, representing a domain in need of further attention.

Furthermore, triangulation is a broad term covering a number of patterns of child involvement in parents’ conflicts (e.g., scapegoating, coalitions, pressure to side, child mediation; Buchanan & Waizenhofer, 2001; Kerig, 1995). The brief daily measure for triangulation used in this study did not offer the level of specificity offered by theoretical work in this domain. Future studies may probe whether distinct patterns of triangulation (i.e., scapegoating, pressure to side with one parent over the other) produce similar discrepancies within families.

Finally, triangulation and family members’ perceptions of IPC were measured within the same day; therefore, study analyses are unable to offer insight into the direction of effects. Although theory articulates that triangulation functions by promoting discrepant views of IPC (e.g., Minuchin, 1974), the possibility remains that parents’ and adolescents’ discrepant views of IPC yield triangulation instead. However, to our knowledge, this alternative interpretation is without theoretical or empirical precedent. Collecting more granular data (e.g., multiple measurements per day) would provide an opportunity to evaluate the theoretically proposed direction of effects.

Conclusions

Triangulation is relatively common in families and consistently correlated with adolescents’ problem outcomes. The present study offers novel insight into the process of triangulation within families, demonstrating within-family links between triangulation and divergence in parents and adolescents’ perceptions of family functioning on a daily timescale. The findings suggest that triangulation may indeed serves to promote a sense of illusory harmony for parents’ in distressed relationships. Moreover, on days when triangulation occurs, both parents and adolescents perceive decreases in family cohesion; however, adolescent perceptions exhibited more pronounced changes. Together, these findings provide empirical support for previously unexamined theoretical principles of triangulation, while carrying meaningful implications for measurement within family science.