Should parents apologize to their adolescents to mend their parent–adolescent relationship? A multi‐method, multi‐phase investigation

Abstract

Apologies following interpersonal offenses tend to mend relationships, but the role of parental apologies in the parent–adolescent relationship is understudied. Further, parental apologies may have diametrical effects on the parent–adolescent relationship depending on their phrasing. In this three‐phase study, we recruited 288 adolescents to assess the role of perceived (phase 1), coded (phase 2), and manipulated (phase 3) parental victim‐centered and defensive apologies in the parent–adolescent relationship (i.e., adolescent forgiveness, relationship satisfaction, and perceptions of parental authority). Results from all three phases converged to suggest that victim‐centered apologies predict greater forgiveness, more autonomous and less controlled motivations to forgive, and higher relationship satisfaction, compared to defensive apologies or no apologies. We found limited associations between parental apologies and parental authority.

Healthy parent–adolescent relationships play a central role in adolescents' flourishing. Not only are they predictive of adolescents' social adjustment (Barry et al., 2008), but they are also associated with adolescents' physical and psychological health (Buist et al., 2017; Mello et al., 2020). However, several factors can hamper this relationship during adolescence, including expected increases in parent–child conflicts (Smetana & Rote, 2019), and daily variations in parents' tendency to behave in ways that can hurt or offend their adolescents (e.g., losing their temper, talking more harshly, using psychological control; Mabbe et al., 2018). As such, it is important to identify how parents may repair their relationship with their adolescents after (often minor) situations in which they behaved offensively. In this study, we take a step in this direction by testing across several designs the hypothesis that victim‐centered (vs. defensive or the absence of) parental apologies preserve the parent–adolescent relationship following parental offenses.

Parental apologies within parent–adolescent relationships

Evidence on the benefits of apologies in other relationship contexts suggests that parental apologies hold the potential to safeguard the parent–adolescent relationship following parental offenses (Ruckstaetter et al., 2017). Among the various relationship indicators that may be swayed by apologies, forgiveness and relationship satisfaction are important to consider. First, both indicators inform on the extent to which the relational damage caused by the offensive behavior is repaired by the apology, either directly (i.e., adolescents' relationship satisfaction with their parent inherently taps into their sense of well‐being within the relationship and tends to fluctuate as a function of parent–adolescent conflicts; Van Doorn et al., 2009) or indirectly (i.e., forgiveness implies “conciliation and goodwill for the offender” and “letting go of past hurt and bitterness,” thus representing a willingness to move forward from the offense; Riek & Mania, 2012). Second, both indicators offer benefits to victims and offenders, including enhanced well‐being and health (Riek & Mania, 2012). In part due to these reasons, relationship satisfaction and forgiveness are often examined as correlates of apologies (Fehr et al., 2010).

Thus far, research has found general interpersonal apologies to be positively related to relationship satisfaction (Schumann, 2012) and one of the most powerful predictors of forgiveness (Fehr et al., 2010). Parental apologies, however, remain specifically understudied. Indeed, we could only find four published studies that examined parental apologies, among which only two had examined their role on child or adolescent outcomes. These correlational studies showed that (1) parents who report being more inclined to apologize to their adolescents tend to rate the quality of their parent–adolescent relationship more positively (Ruckstaetter et al., 2017), and that (2) maternal apologies are associated with more prosocial child behaviors and with less internalizing and externalizing child behavioral problems (Lee et al., 2023).

Parental apologies and parental authority: The needs‐based model of reconciliation

These pioneer studies are promising in that they hint at a positive role of parental apologies. However, further research is needed to better understand if and how parents should apologize, for parental apologies could also present drawbacks. Indeed, some theoretical writings suggest that parental apologies may hamper parental authority (i.e., the extent to which adolescents view their parents as having the power and the legitimate right to affect their behaviors; Yaffe, 2017). According to the needs‐based model of reconciliation (Shnabel & Nadler, 2008), individuals surrender their power status to victims when apologizing (e.g., by providing them with the choice to forgive). Though this power reversal is deemed restorative for the victim's well‐being, backfire effects related to power reversals may be feared in hierarchical relationships, including parent–adolescent dyads. Based on this model (Shnabel & Nadler, 2008), one might anticipate that parents who apologize put themselves in a position of weakness and as such compromise their authority. If parental apologies do undermine parental authority—an important factor for healthy parent–adolescent relationships and adolescents' development (e.g., Baumrind, 2013)—the benefits of parental apologies might be less clear.

Contrary to these propositions, research on authoritative parenting style (i.e., parents who are highly responsive to their children's needs while maintaining firm authority over them; Baumrind, 2013) suggests that parents' authority status may be unaffected by their decision to apologize. Indeed, scales on authoritative parenting describe such parents as willing to “admit they made a mistake [when they made a decision in the family that hurt the child]” (Buri, 1991). However, the specific role of parental apologies in parental authority has not been studied.

Victim‐centered and defensive parental apologies: A self‐determination theory perspective

Evidence from research on apologies in other relationship contexts also suggests that the benefits of parental apologies may be influenced by their phrasing (Kachanoff et al., 2017; Schumann & Dragotta, 2020). According to self‐determination theory (SDT), the effects of interpersonal behaviors on relationship outcomes depend on their ability to support (vs. thwart) individuals' basic psychological needs for autonomy (i.e., sense of volition and psychological freedom), relatedness (i.e., sense of connectedness to close others), and competence (i.e., sense of self‐efficacy; Ryan & Deci, 2017). SDT thus posits that parents who behave in ways that are more likely to support the basic psychological needs of their adolescents (i.e., who adopt more need‐supportive behaviors) foster relationship quality, whereas those who behave in ways that thwart their adolescents' basic needs (i.e., who adopt more need‐thwarting behaviors) negatively impact relationships (Vansteenkiste & Ryan, 2013). Need‐supportive behaviors also tend to foster investment in the parent‐relationship for motivations that are autonomous (i.e., self‐endorsed; e.g., identified motivation), whereas need‐thwarting behaviors tend to foster motivations that are controlled (i.e., internally or externally pressured; e.g., introjected motivation, external motivation, and amotivation; Ryan & Deci, 2017). Adolescents' autonomous and controlled motivations to engage in a variety of behaviors have been in turn shown to foster and forestall their psychosocial adjustment, respectively (Soenens & Vansteenkiste, 2010; Vansteenkiste & Ryan, 2013). Based on these propositions, we argue that the ability of parental apologies to mend the parent–adolescent relationship—that is, to foster adolescents' forgiveness, autonomous (vs. controlled) motivations to forgive, and relationship satisfaction with their parents—depends on their need‐supportive and need‐thwarting phrasing.

Researchers have outlined apology elements that should influence their effectiveness to mend relationships (Kachanoff et al., 2017; Schumann, 2014). From a SDT perspective, these apology elements are also likely to make apologies more need‐supportive or need‐thwarting, which may account for their differential effects. Specifically, eight apology elements have been proposed to increase apologies' effectiveness. Each of these reflects a more need‐supportive, victim‐centered (VC), focus from offenders: (1) expressing remorse (e.g., “I regret what I did”), (2) accepting responsibility for the harm done (“This is my mistake”), (3) admitting the injustice or wrongdoing (“What I did was wrong”), (4) acknowledging the victim's harm (“I see that you are hurt”), (5) forbearing (“I'm taking steps to make sure this will not happen again”), (6) offering repair (“I will pay for the damages”), (7) offering internally oriented explanations (“I lost control”), and (8) making non‐pressuring forgiveness requests (“I hope you can forgive me, though I know that apologizing may not smooth all waters”; Kachanoff et al., 2017; Schumann, 2014). In addition to these eight VC elements, five apology elements that may hamper apologies' effectiveness have been highlighted. Each of these reflects a more need‐thwarting, defensive, stance from offenders: (1) externally oriented excuses (e.g., “I was in a hurry”), (2) justifications (“I did it for your own good”), (3) blame (“You were being so difficult”), (4) minimizations (“It was just a joke”), and (5) pressuring forgiveness requests (“I beg you to forgive me”; Kachanoff et al., 2017; Schumann, 2014). All 13 VC and defensive apology elements are listed in Schumann (2014), apart from the two forgiveness request elements, which were later proposed by Kachanoff et al. (2017). For more theoretical and empirical support for the need‐supportive and need‐thwarting value of these apology elements, see Supporting Information.

Severity of parental offenses

Finally, research on apologies in other relationship contexts suggests that the effects of parental apologies may not only vary as a function of their phrasing but also according to the offenses for which they are offered. Thus far, most research has found apologies to be equally or more effective following less severe (vs. more severe) offenses (e.g., Schumann, 2012; Schumann & Dragotta, 2020). Because of these results, it is important to consider the moderating role of offense severity on parental apologies.

Past limitations and present study

In sum, research and theory suggest strong relational benefits (and sometimes discuss potential pitfalls) of apologies, but these propositions are understudied in the parenting literature. Clarifying the role of parental apologies in the parent–adolescent relationship is important when we consider the key role of healthy parent–adolescent relationships in adolescents' flourishing (e.g., Barry et al., 2008; Buist et al., 2017; Mello et al., 2020) and the commonness of parental mishaps that can harm these relationships (Mabbe et al., 2018). To clarify the role of parental apologies in the parent–adolescent relationship, the following issues need to be addressed. First, because parental apologies may hamper parental authority (e.g., through power‐reversal), research needs to consider their association with this relationship outcome in addition to more commonly assessed aspects of relationships (i.e., forgiveness and relationship satisfaction). Second, because the impact of apologies may diametrically vary according to their VC and defensive elements, parental apologies should be assessed while considering their defining elements. Third, because the effects of apologies may also vary according to the severity of the offense, this potential moderating factor should be considered. Finally, because the two existing studies on parental apologies are correlational, complementary experimental designs are necessary to verify whether parental apologies cause changes in the parent–adolescent relationship (rather than solely vary in frequency as a result of relationship quality). In turn, finding causal effects of parental apologies can inform parenting experts on the relevance of teaching parents how to apologize (e.g., as part of larger parenting programs or within therapy).

To address these issues, we conducted a three‐phase study in which we examined the role of parental apologies in adolescents' (1) forgiveness, (2) relationship satisfaction with their parent, and (3) perceptions of their parent's authority. In phase 1, we assessed the link between adolescents' perceptions of their parent's general tendency to offer VC and defensive apology responses following offensive behaviors and their general assessment of their parent–adolescent relationship. In phase 2, we followed the lead of past research on interpersonal and parental apologies (Adams‐Clark et al., 2021; Lee et al., 2023; Schumann & Dragotta, 2020) and coded the VC and defensive apology elements of a written recalled parental apology response to examine their associations with adolescents' situational assessments of their relationship with their parent. In phase 3, we asked adolescents to evaluate the relational effects of a manipulated (i.e., randomly assigned) hypothetical parental apology response.

Testing our research questions across these three phases allows us to verify the replicability of our findings across methods while leveraging their complementary strengths. Manipulating hypothetical parental apology responses (phase 3) enhances internal validity by assessing the direction of the effects. Coding parental apology responses (phase 2) improves external validity by reflecting real parent–adolescent interactions. Measuring adolescents' perceptions of their parent's general apology tendencies (phase 1) improves ecological validity by assessing whether the situational role of (manipulated or coded) parental apologies translates into general trends.

To make a more stringent test of our assessed associations, we made key controls. First, in phases 1 and 2, we controlled for (1) autonomy‐supportive parenting to ensure that any effects of apologies would not be attributable to other need‐supportive and need‐thwarting parenting behaviors (Mageau et al., 2015), (2) the frequency of parental offenses to consider its potential role in adolescents' appraisal of parental apologies (Schumann & Dragotta, 2020), and (3) sociodemographics known to play a role in the parent–adolescent relationship (i.e., adolescents' age and gender as well as parents' socioeconomic status and gender; Bornstein, 2016). Second, in phase 3, we controlled for participants' perceptions of their parent's behaviors solicited at phase 2 to avoid spillover effects. Third, we considered the moderating role of the severity of parental offenses in the relation between parental apologies and our outcomes in all phases (Schumann & Dragotta, 2020). Finally, we conducted our study with mid‐ to late adolescents given that parent–adolescent conflicts tend to reach their peak around mid‐adolescence (14–17; Hadiwijaya et al., 2017) and that children are most critical of their parent's authority during late adolescence (18 and older; Darling et al., 2008).

We hypothesized that, controlling for the aforementioned covariates, parental VC apology responses would play a positive role in adolescents' (H1) forgiveness (i.e., greater forgiveness, more autonomous and less controlled motivations to forgive), and (H2) relationship satisfaction with their parents, whereas parental defensive apology responses would yield negative associations with these variables. Based on the literature on authoritative parenting (Buri, 1991), we did not expect any relation between parental apologies and (H3) parental authority. Finally, we hypothesized that the magnitude of the effects of VC and defensive parental apologies on our relationship indicators would be equal to or greater following less severe (vs. more severe) offenses.

METHOD

Participants

We recruited 288 mid‐ to late adolescents (M = 16.75 years, SD = 1.67). Participants were high school (81.60%) and first‐year college (18.40%) French speaking students still living with their parents. Most identified as girls (64.43%). The remainder identified as boys (32.41%), non‐binary (1.18%) or “other” (1.98%). Adolescents were mostly born in Canada (89.92%); the remainder were born in one of 16 other countries (1.26% of adolescents or less per country). We conducted our three‐phase, questionnaire‐based, online study in a single 45‐min session. Throughout the study, we asked adolescents to answer questions while thinking about the parent with whom they interact most often (referred to as the “target parent” hereinafter). When prompted, most adolescents selected their mother as the target parent (75.87%); the rest selected their father (23.08%) or a female relative (1.05%). Most target parents were born in Canada (89.45%), with the remaining being born in one of 20 other countries (0.84% of parents or less per country). Target parent socioeconomic status was relatively high. Approximately three fifths had a university diploma (62.20%); the rest had another post‐secondary certification (11.38%), had only their high school diploma (25.20%), or had not finished high school (1.22%). Information on economic status was more limited as 42.57% of adolescents did not know their target parent's annual income. Among those who could answer, an approximate third reported an income (in CAD) higher than $100,000 (31.46%), close to half reported an income between $50,000 and $100,000 (44.76%), and the remainder reported an income between $30,000 and $50,000 (14.69%) or lower than $30,000 (9.09%). For more information on participants' sociodemographic characteristics, see Supporting Information.

Recruitment procedure and questionnaire completion order

Following approval from the research ethics board of the first author's university, the first author and trained research assistants (RAs) met with students from three high schools in their classroom to present the study. We targeted high schools based on their proximity to the first author's workplace. We invited college students to complete our study via the university's participant pool management system and, as such, did not meet them. To participate in our study, adolescents had to give their written consent. Parents of high school (but not college) students also signed a consent form. High school students completed the online study in their classroom during a second visit (and as such could ask questions to the first author or the RAs), whereas college students completed the study alone. In all cases, we collected data confidentially.

We presented our measures to participants in the following order: (1) all independent variables (IVs) and parental offense measures from phase 1, (2) all variables from phase 2, (3) all variables from phase 3, (4) all dependent variables (DVs) from phase 1, and (5) autonomy‐supportive parenting and sociodemographics. By temporally separating participants' responses to the IVs and DVs of phase 1, we aimed to reduce artificial inflation of their correlations.

Phase 1: Correlational design (global level)

Procedure

During phase 1, we asked adolescents to answer questions about their target parent and themselves from a global perspective. Thus, adolescents reported on (1) their target parent's general tendency to offer VC and defensive apology responses after offending them, (2) their general tendency to forgive their parent after they offended them, (3) the motivations generally leading them to forgive their parent (i.e., autonomous and controlled), (4) their general relationship satisfaction with their parent, (5) their general perceptions of their parent's authority, (6) their target parent's autonomy‐supportive parenting, (7) the general frequency and severity of their parent's offenses, and (8) sociodemographics.

Measures

Parental apologies

To assess adolescents' perceptions of their target parent's general apology tendencies, we developed the Parental Apology Quality Scale (PAQS). This 13‐item scale lists the eight VC and five defensive apology elements outlined in prior research (i.e., Kachanoff et al., 2017; Schumann, 2014). Adolescents read each apology element and indicated the extent to which they agreed that each element “corresponds to the way their target parent [generally] speaks to them after hurting them or causing them injustice”, on a scale of 1 (do not agree at all) to 7 (very strongly agree). This scale has good psychometric properties. First, as shown in Table 1, exploratory factor analysis revealed a two‐factor solution, with (1) the two highest eigenvalues ≥2.01 (all other eigenvalues ≤0.18), and (2) each apology element only loading onto its theorized factor (all expected loadings ≥ |0.67|; all cross loadings ≤ |0.13|). Second, internal consistency coefficients were satisfactory for both the VC (α = .94) and the defensive (α = .90) subscales. Finally, we found evidence for the theoretical validity of the scale. Specifically, adolescents who perceived their target parent as more autonomy‐supportive tended to rate their parent apologies as more VC, r = .52, p < .001, and less defensive, r = −.54, p < .001.

TABLE 1.

Factor loadings for the exploratory factor analysis with oblique (oblimin) Rotation on the Parental Apology Quality Scale (PAQS).

Apology elements	Factor
In general, after my parent behaves towards me in a way that is hurtful or that causes me injustice, s/he…	Victim‐centered	Defensive
1. Acknowledges that their behavior was hurtful or unfair	0.88	−0.06
2. Commits to better behave in the future	0.86	−0.04
3. Expresses remorse (regrets)	0.86	0.07
4. Offers reparation for their behavior	0.85	0.07
5. Accepts their part of responsibility in the wrong they have caused	0.84	−0.05
6. Acknowledges that I have been hurt or upset	0.82	0.00
7. Tries to explain why they have behaved this way	0.74	0.13
8. Gives me the choice to forgive him/her or not	0.69	−0.06
9. Minimizes their responsibility in their behavior	−0.07	0.91
10. Minimizes the consequences of their behavior	0.01	0.90
11. Blames me for their behavior	−0.04	0.76
12. Tries to defend their behavior	0.07	0.72
13. Pressures me to forgive him/her	0.07	0.67

Note: Factor loadings greater than .30 are written in bold to better illustrate the items loading on each factor.

Forgiveness

To measure adolescents' forgiveness, we adapted Schumann (2012)'s question on forgiveness following offenses and asked adolescents to indicate “how frequently they forgive their target parent when they behave hurtfully or causes them an injustice”, on a scale of 1 (almost never) to 7 (almost always).

Motivations to forgive

To assess adolescents' motivations to forgive their parent, we adapted the Self‐Regulation Questionnaire and Situational Motivational Scale to our setting (Guay et al., 2000; Soenens et al., 2009). This resulted in a 12‐item scale measuring the extent to which adolescents agree that they generally forgive their parents for autonomous and controlled motivations, on a scale of 1 (do not agree at all) to 7 (very strongly agree). To measure autonomous forgiveness, we used the identified subscale (3 items; e.g., “because I personally agree with the idea of forgiving my parent”; α = .80). To measure controlled forgiveness, we used the introjected (3 items; e.g., “because otherwise, I would feel like a bad person”; α = .77), external (3 items; e.g., “because I feel forced to forgive my parent”; α = .81), and amotivation subscales (3 items; e.g., “when I forgive my parent, I am not sure why I do it”; α = .85).

Parent‐adolescent relationship satisfaction

To measure adolescents' relationship satisfaction with their target parent, we used Schrodt and O'Mara (2019)'s 11‐item adaptation of the Marital Opinion Questionnaire to parent–adolescent relationships. This scale measures relationship satisfaction with a 7‐point semantic differential scale for all items (e.g., “gratifying – disappointing”), except for one item assessing global relationship satisfaction on a scale from 1 (highly unsatisfied) to 7 (highly satisfied). Internal consistency was high, α = .93.

Parental authority

To assess adolescents' perceptions of their parental authority, we adapted the Parental Authority Instrument (Yaffe, 2017). This instrument assesses the two main components of parental authority, namely power and legitimacy. In our 3‐item version, adolescents rated the extent to which they agreed with two items regarding power (“My parent can make me obey their rules”; “I am obligated to obey my parent rules”) and one item regarding legitimacy (“My parent has the right to set rules”), on a scale of 1 (do not agree at all) to 7 (very strongly agree). Higher scores imply a more positive perception of parental authority (α = .74).

Covariates

Autonomy‐supportive parenting

To measure autonomy‐supportive parenting, we asked adolescents to complete the Perceived Parental Autonomy Support Scale (Mageau et al., 2015). This 24‐item scale assesses the extent to which adolescents agree that their parents behave in ways that support (12 items) or thwart (12 items) their need for autonomy, on a scale of 1 (do not agree at all) to 7 (very strongly agree). To create our score of autonomy‐supportive parenting, we combined all autonomy‐supportive and reversed autonomy‐thwarting items (α = .94).

Parental offense frequency and severity

We assessed the general frequency and severity of parental offenses, based on past work operationalizing interpersonal offenses (e.g., Shnabel & Nadler, 2008). To assess frequency, we asked adolescents to indicate “how frequently their parent behaves in a way that is hurtful to them or that causes them an injustice”, on a scale of 1 (almost never) to 7 (almost always). To assess severity, we asked adolescents to rate the extent to which they agree that their target parent's offenses generally make them feel (1) hurt, (2) victim of an injustice, (3) disrespected, and (4) upset, on a scale of 1 (do not agree at all) to 7 (very strongly agree). Internal consistency coefficient was high (α = .89).

Sociodemographics

Finally, we asked adolescents to specify their age, gender, and sex as well as their target parent gender, education level, and annual income. Because an insufficient ratio of participants identified as non‐binary or other diversity of gender, we included adolescents' sex (rather than gender) as a covariate in our analyses. Adolescents rated their parent's education level by rating their parent's highest diploma on a scale from 1 (elementary school) to 5 (graduate studies). They also rated their parent's annual income in CAD$ on a scale from 1 (less than 15,000$) to 6 (100,000$ or more). To create an index of parent socioeconomic status (SES), we averaged the standardized scores of parents' education and income.

Phase 2: Correlational design (situational level)

Procedure

In phase 2, we asked adolescents to recall and describe a parental offense. Borrowing Shnabel and Nadler (2008)'s offense recall procedure, we asked adolescents to “think about a situation they remember well and in which their target parent behaved towards them in a way that was hurtful or caused them injustice.” After recalling the offense, adolescents specified whether their parent had apologized following the offense and then wrote down the apology (or what their parent had said instead of apologizing) so that we could code the presence of VC and defensive apology elements. Adolescents then rated the severity of that offense and the extent to which the described parental apology (or lack thereof) led them to forgive their target parent, affected their relationship satisfaction, and affected their perceptions of their parent's authority.

Measures

Parental apologies

To assess parental apology elements in phase 2, we asked adolescents to write down as accurately as possible the exact words that their target parent used to apologize (or the words they used instead of apologizing) after the recalled offense. Using our developed Parental Apology Coding System, a trained research assistant then coded whether each of the eight VC and five defensive apology elements outlined in the PAQS was included in the described parental response. When an element was present, it received a score of 1. When an element was not present, it received a score of 0. For instance, the following parental response: “I'm sorry, I was angry and I didn't realize what I said was serious and hurtful” was coded as containing (1) three VC elements (expression of remorse [“I'm sorry”], internally oriented explanations [“I was angry”], acknowledging harm [“What I said was serious and hurtful”]), and (2) one defensive element (minimization of responsibility [“I didn't realize”]). To create our variables, we averaged the binary scores of all VC elements and of all defensive apology elements and obtained two ratio scores ranging from 0 (i.e., no VC or no defensive element was present) to 1 (i.e., all VC or all defensive elements were present; for similar procedures, see Adams‐Clark et al., 2021; Lee et al., 2023; Schumann & Dragotta, 2020).

Our coding of parental apologies was based on deductive content analysis, where coders assumed that VC and defensive elements (as outlined in our coding system) would best describe the written responses. Adopting a post‐positivist epistemological relativism stance, coders were also guided by SDT principles, parenting research, and their background in psychology to interpret each response. Coders were all non‐parents, within an age range close to that of participants, though they relied on older researchers who were also parents for guidance.

Because we expected that coders would vary in their interpretations of parental responses, we verified the internal consistency and validity of our coding. To verify internal consistency, a second trained research assistant coded 34.72% of the parental responses (n = 100). Interrater reliability was satisfactory for both the VC ratio scores (ICC = 0.86; 95% CI 0.79–0.92) and the defensive ratio scores (ICC = 0.74; 95% CI 0.60–0.84; Koo & Li, 2016). Given the satisfactory interrater reliability, disagreements between the two coders were not further handled (and as such only codes from the first coder were included in the analyses). To verify validity, we asked adolescents to fill out the PAQS while thinking about the described parental apology. We found positive relations between coded and perceived VC apology elements, r = .61, p < .001, as well as between coded and perceived defensive apology elements, r = .33, p < .001, suggesting good validity. However, 77 responses (out of 288) could not be coded because participants answered inappropriately (n = 12; e.g., unreadable answer), answered that they did not remember the specific words their target parent had said (n = 15), or did not answer (n = 50).

Forgiveness

To assess adolescents' situational forgiveness, we used the same measure as in phase 1. This time, however, we asked adolescents to indicate the extent to which they agreed that the described apology response led them to “forgive their parent”, on a scale of 1 (do not agree at all) to 7 (very strongly agree).

Parent‐adolescent relationship satisfaction

To measure adolescents' situational relationship satisfaction with their target parent, we used the MOQ's single item assessing relationship satisfaction (Schrodt & O'Mara, 2019). Adolescents thus indicated the extent to which they agreed that the described parental response had led them to “feel satisfied in their relationship with their parent”, on a scale of 1 (do not agree at all) to 7 (very strongly agree).

Parental authority

To assess adolescents' situational perceptions of parental authority, we used the same measure as in phase 1 (Yaffe, 2017). This time, however, adolescents rated the extent to which they agreed that the described parental response had undermined their parent's authority, on a scale of 1 (do not agree at all) to 7 (very strongly agree). We reversed adolescent scores so that higher ratings would imply better perceptions of parental authority (α = .87).

Parental offense severity

Finally, we assessed the severity of the described parental offenses. Adolescents completed the same 4‐item offense severity measure as in phase 1 (on the same 7‐point scale), but this time while thinking about the described parental offense (α = .90).

Phase 3: Experimental design (situational level)

Procedure

During phase 3, we followed the lead of past apology studies presenting hypothetical apology responses to real (or ostensibly real) offenses and asked adolescents to “forget for a moment what their target parent said and did after hurting them or causing them an injustice (e.g., apology, lack of apology) and imagine the following situation” (Kachanoff et al., 2017; Schumann & Dragotta, 2020). We then randomly assigned adolescents to one of four conditions, each presenting a parental apology response that varied in terms of VC and defensive phrasing or acted as a control (no apology) condition. (1) In the basic condition, adolescents imagined that their target parent had offered a basic VC apology: “I apologize for what I did.” (2) In the defensive condition, parents offered a defensive apology by including three defensive elements on top of the basic apology: “I apologize for what I did, but what I did was not that bad and besides, I had good reasons. You know, these things can happen in a relationship.” (3) In the VC condition, parents added three VC elements to the basic apology: “I apologize for what I did. It was wrong. I should have behaved better. I'll make efforts to ensure that this does not happen again in the future.” (4) Finally, in the no apology condition, adolescents imagined that “their target parent had the opportunity to offer an apology but said nothing.”

To enhance the external validity of our between‐person manipulation, we increased participants' immersion level by asking them to read their assigned apology response directly after phase 2 (i.e., after recalling the offense for which the assigned apology response was given). To further enhance validity, we implemented key controls. First, we controlled for participants' perceptions of the recalled offense and the recalled apology response to avoid spillover effects (see plan of analyses). Second, we ensured that the VC and defensive conditions (French version) had the same number of apology elements (5), words (34), and characters (185) to enhance the likelihood that any difference would be due to their VC and defensive phrasing.

Measures

To assess the effects of our manipulated parental apology responses on the parent–adolescent relationship, we asked adolescents to answer the same outcome measures as in phase 2, but this time while thinking about their assigned parental apology response. Thus, adolescents rated the extent to which they agreed that their assigned parental apology response would lead them to forgive their parent, lead them to feel satisfied in their relationship with their parent, and influence their perceptions of their parent's authority (α = .90), on a scale of 1 (do not agree at all) to 7 (very strongly agree). As a manipulation check, we also asked adolescents to rate the extent to which they agreed that their assigned response led them to “consider that their parent had apologized”, on a scale of 1 (do not agree at all) to 7 (very strongly agree).

Plan of analyses

Phase 1: Correlational design (global level)

We conducted all our analyses on R 4.3.1. In line with past research on apologies, we assumed linear associations between parental apologies and our outcomes of interest (Fehr et al., 2010; Lee et al., 2023; Ruckstaetter et al., 2017; Schumann & Dragotta, 2020). Data and R script for our analyses are available on Open Science Framework at DOI: 10.17605/OSF.IO/C4Z2A.

Preliminary analyses

For the preliminary analyses of phase 1, we examined adolescents' ratings of their parent's VC and defensive apology elements to document their prevalence following offenses. We then examined the correlations between our main variables.

Main analyses

For the main analyses of phase 1, we conducted a series of multivariate regressions using the ML estimator and FIML (to handle missing data). In our first model, we examined whether adolescents' perceptions of parental VC and defensive apology elements interacted with the severity of parental offenses to predict adolescents' general forgiveness, motivations to forgive (i.e., identified, introjected, external and amotivation), relationship satisfaction, and perceptions of parental authority, while adjusting for autonomy‐supportive parenting, the frequency of parental offenses, and sociodemographics (i.e., adolescents' age and sex as well as parents' gender and SES). Provided significant interactions, we intended to examine the predictive value of parental apologies at one SD above and below the average severity score. Provided non‐significant interactions, we intended to examine the main effects of parental apologies while modeling offense severity as an additional covariate.

Phase 2: Correlational design (situational level)

In phase 2, we followed the same analytical plan as in phase 1. In our preliminary analyses, we examined our coding of parental apologies to document the frequency of their VC and defensive elements and then examined correlations between our variables. In our main analyses, we examined interaction effects between the severity of the described parental offense and our coded measures of VC and defensive apology elements, while controlling for the same covariates as in phase 1. We then pursued examining simple or main effects.

Phase 3: Experimental design (situational level)

In phase 3, we first conducted a manipulation check using an ANOVA followed by Tukey post hoc tests to verify that adolescents in all apology conditions agreed to a greater extent than those in the no apology condition that their assigned response was an apology. For our main analyses, we examined whether our manipulated parental apology responses caused differences in adolescents' intent to forgive, relationship satisfaction, and perceptions of their target parent's authority with a MANCOVA (to evaluate the significance of multivariate differences). We then conducted a series of ANCOVAs (to identify the specific outcome variables responsible for the significant effect at the multivariate) and Tukey post hoc tests (to examine pairwise comparisons between conditions for each significant outcome variable). In all analyses, we controlled for potential spillover effects related to the recalled parental offense and apology response from phase 2. Specifically, we modeled as covariates (1) adolescents' perceptions of the VC and defensive apology elements in the recalled parental apology response (using the PAQS), and (2) the corresponding outcome assessed following the recalled parental apology response (i.e., all outcomes in the MANCOVA and the same outcome as the assessed DV in each ANCOVA).

To ensure the validity of our findings, we verified three main assumptions for our MANCOVA, namely (1) multivariate normality, (2) multicollinearity, and (3) homogeneity of covariance. Provided a significant violation of the homogeneity of covariance, we intended to use Pillai's Trace (rather than Wilk's Lambda) as the primary statistic to determine the significance of our results. To obtain effect sizes, we calculated partial η² for all effects. In line with the analytical plan of prior phases, we first tested interaction effects with offense severity (i.e., phase 2 measure, dichotomized at the mean) and then examined simple effects for significant interactions and main effects for non‐significant interactions.

RESULTS

Phase 1: Correlational design (global level)

Preliminary analyses

Global parental apology responses

Descriptive statistics and correlations between the variables used in phase 1 are shown in Table 2 and in Supporting Information. Examining adolescents' perceptions of their target parent's general apology responses suggested that target parents tend to offer apology responses that are more VC (4.18) than defensive (3.13), p < .001.

TABLE 2.

Mean (M), standard deviations (SD), and correlations among the main variables of phases 1 and 2.

Note: *p < .05. Phase 1 correlations. Phase 2 correlations. Global offense frequency and autonomy‐supportive parenting are used as covariates in phases 1 and 2 and, as such, are included in both triangles. For the correlations between the sociodemographic variables and all other variables, see Supporting Information.

Abbreviation: VC, victim‐centered.

Correlations

In line with our main hypothesis, correlations revealed that parental apologies with more VC and less defensive elements were related to adolescents' greater forgiveness and relationship satisfaction with their target parent. Higher levels of VC apology elements and lower levels of defensive apology elements were also linked to adolescents' higher global identified motivation to forgive and lower controlled types of motivation to forgive, except for one non‐significant relation between VC apology elements and introjected motivation, p = .185. We found no relation between parental apologies and parental authority, ps ≥ .573.

Main analyses

Main analyses first revealed no interaction effects between the severity of parental offenses and parental apology elements on our outcomes, ps ≥ .123, except for one non‐significant trend between defensive apology elements and offense severity on external motivation to forgive, p = .052. Given our assessment of 14 interactions, we interpreted this trend as type I error and thus did not examine it further.

Main effects revealed similar associations as those in correlational analyses, though fewer were statistically significant. Thus, as shown in Table 3, main effects revealed that beyond autonomy‐supportive parenting, the general frequency and severity of parental offenses, and sociodemographics, VC apology elements were related to (1) greater forgiveness and (2) higher relationship satisfaction, but did not predict (3) perceptions of parental authority, p = .574. Defensive apology elements did not predict any of these three outcomes, ps ≥ .174.

TABLE 3.

Standardized beta coefficients (standard errors) for the main analyses of phases 1 and 2.

Variables	Forgiveness	Identified forgiveness	Introjected forgiveness	External forgiveness	Amotivation to forgive	Relationship satisfaction	Parental authority
Phase 1
VC apology elements	0.24 (0.07)*	0.36 (0.06)*	0.10 (0.07)	−0.12 (0.05)*	−0.10 (0.06)	0.15 (0.05)*	−0.04 (0.08)
Defensive apology elements	0.00 (0.07)	−0.22 (0.07)*	0.12 (0.07)	0.33 (0.06)*	0.29 (0.06)*	−0.07 (0.05)	−0.02 (0.08)
Autonomy‐supportive parenting	0.15 (0.09)	0.03 (0.08)	−0.20 (0.09)*	−0.34 (0.07)*	−0.24 (0.07)*	0.44 (0.05)*	−0.12 (0.09)
Parental offense frequency	−0.08 (0.08)	−0.06 (0.08)	0.00 (0.09)	0.03 (0.07)	0.21 (0.07)*	−0.31 (0.05)*	−0.12 (0.09)
Parental offense severity	0.26 (0.08)*	0.10 (0.07)	0.06 (0.08)	0.02 (0.06)	0.01 (0.06)	0.02 (0.05)	0.10 (0.08)
Adolescents' age	−0.01 (0.06)	.06 (0.06)	−0.03 (0.06)	−0.03 (0.05)	−0.08 (0.05)	0.02 (0.04)	−0.16 (0.07)*
Adolescents' sex	−0.08 (0.07)	0.02 (0.07)	−0.09 (0.07)	−0.10 (0.06)	−0.07 (0.06)	0.06 (0.04)	0.03 (0.08)
Parent's gender	0.11 (0.06)	0.03 (0.06)	0.00 (0.06)	−0.08 (0.05)	−0.08 (0.05)	−0.03 (0.04)	−0.04 (0.07)
Parent's socioeconomic status	0.08 (0.09)	0.03 (0.07)	−0.07 (0.09)	0.10 (0.07)	0.14 (0.07)*	0.04 (0.05)	0.10 (0.09)
Phase 2
VC apology elements	0.33 (0.07)*	–	–	–	–	0.23 (0.06)*	0.22 (0.08)*
Defensive apology elements	−0.02 (0.07)	–	–	–	–	0.00 (0.06)	−0.08 (0.08)
Autonomy‐supportive parenting	0.19 (0.08)*	–	–	–	–	0.31 (0.06)*	0.20 (0.08)*
Parental offense frequency	−0.10 (0.08)	–	–	–	–	−0.31 (0.06)*	0.01 (0.08)
Parental offense severity	−0.06 (0.08)	–	–	–	–	−0.06 (0.06)	−0.04 (0.08)
Adolescents' age	−0.02 (0.06)	–	–	–	–	−0.07 (0.05)	0.13 (0.07)
Adolescents' sex	−0.02 (0.07)	–	–	–	–	−0.00 (0.06)	−0.05 (0.07)
Parent's gender	−0.02 (0.06)	–	–	–	–	−0.06 (0.05)	−0.03 (0.06)
Parent's socioeconomic status	−0.04 (0.08)	–	–	–	–	0.00 (0.06)	−0.08 (0.08)

Note: *p < .05. Adolescent sex: 0 = Girl, 1 = Boy. Parent gender: 0 = Mother, 1 = Father.

Abbreviation: VC, victim‐centered.

Examining the associations between parental apologies and adolescents' motivations to forgive revealed that higher levels of VC apology elements and lower levels of defensive apology elements predicted more identified motivation, less external motivation, and less amotivation to forgive (though this last association was not significant for VC elements, p = .090). Parental apologies did not predict introjected forgiveness; however, ps ≥ .102.

Phase 2: Correlational design (situational level)

Preliminary analyses

Coded parental apology responses

Descriptive statistics of coded parental apologies are shown in Table 2 and in Supporting Information. In line with phase 1, parents were coded as including more VC apology elements (1.07) than defensive ones (0.75), p = .006. Our coding also revealed that more parents omitted using VC apology elements (46.39%) than defensive elements (34.02%), p = .050. Taken together, these results suggest that though fewer parents use VC apology elements following offenses, those who include VC elements tend to use several. This interpretation is consistent with the fact that 68.27% of parental responses including VC elements had more than one, whereas only 9.25% including defensive elements had more than one. For information on the prevalence of each apology element, see Supporting Information.

Correlations

Descriptive statistics and correlations between the variables of phase 2 are shown in Table 2 and Supporting Information. In line with phase 1, apology responses coded as including more VC and fewer defensive elements were related to adolescents' greater forgiveness and relationship satisfaction with their target parent. This time, however, such apologies were positively associated with adolescents' perceptions of their parent's authority.

Main analyses

Our main analyses revealed similar results as in phase 1. First, coded parental apology elements did not interact with the severity of the described parental offenses to predict our outcomes, ps ≥ .303. Second, as shown in Table 3, main effects revealed that beyond autonomy‐supportive parenting, the global frequency and situational severity of parental offenses, and sociodemographics, coded VC apology elements were related to (1) greater forgiveness, (2) higher relationship satisfaction, and this time (3) more positive perceptions of parental authority. As in phase 1, coded defensive apology elements did not predict any outcome, ps ≥ .317.

Phase 3: Experimental design (situational level)

Preliminary analyses

Manipulation check

Conducting an ANOVA to examine our manipulation success revealed a significant difference in adolescents' perceptions of their assigned apology response, F (256, 3) = 57.08, p < .001, η² = .40. As expected, Tukey post hoc tests revealed that adolescents in all apology conditions (Ms ≥4.01, SEs ≤ .20) agreed to a greater extent than those in the no apology condition (M = 2.63, SE = 0.21) that their assigned response was an apology, ps < .001.

Main analyses

Multivariate effect

Table 4 and Figure 1 present the results of the main analyses of phase 3. Conducting a MANCOVA to examine the multivariate effect of our manipulated parental apology responses revealed (1) no issue with multivariate normality (kurtoses ≤ |.98| and skewness ≤ |.59|, linear associations, homoscedasticity), (2) no issue with multicollinearity (with correlations among dependent variables ranging between 0.14 and 0.70), but (3) a violation of the homogeneity of covariance (as calculated with Box's test of Equality of Covariance Matrices), p < .001. We thus examined the statistical significance of our results using Pillai's Trace.

TABLE 4.

Means (standard errors) of adolescents' forgiveness, relationship satisfaction, and perceptions of parental authority according to manipulated parental apology response.

Dependent variables	ANCOVA	Partial η2	No apology	Defensive apology	Basic apology	VC apology
Forgiveness	F (247, 3) = 27.62, p < .001	.25	3.54 (0.20)a	4.38 (0.18)b	5.63 (0.18)c	5.56 (0.19)c
Relationship satisfaction	F (246, 3) = 28.65, p < .001	.26	3.46 (0.19)a	4.03 (0.17)a	5.45 (0.17)b	5.26 (0.18)b
Parental authority	F (247, 3) = 1.65, p = .178	.02	3.38 (0.14)a	3.36 (0.13)a	3.33 (0.14)a	3.01 (0.14)a

Note: Means in the same row with a different subscript (a, b, c) differ at p < .05, with Tukey pairwise comparisons.

Abbreviation: VC, victim‐centered.

FIGURE 1.

Adolescents' forgiveness, relationship satisfaction, and perceptions of their parental authority according to the manipulated parental apology response. Bars represent 95% confidence intervals. For each variable, scores with a different subscript (a, b, c) differ at p < .05, with Tukey pairwise comparisons. VC, victim‐centered.

Our MANCOVA revealed (1) no interaction effect with the severity of the offense at the multivariate level, p = .410, but (2) a significant medium main effect of our manipulation at the multivariate level, Pillai's Trace = 0.35, F (9, 687) = 10.18, p < .001, partial η² = .12. We thus proceeded with ANCOVAs to examine the main effects for each dependent variable.

Forgiveness

Conducting an ANCOVA to assess the impact of our manipulation on adolescents' forgiveness revealed a large main effect, F (247, 3) = 27.62, p < .001, partial η² = .25. Tukey post hoc tests showed that adolescents in all apology conditions had greater intentions to forgive their parents than those in the no apology condition, M = 3.54, SE = 0.20, 95% CI = 3.15–3.93, 0.84 ≤ M _differences ≤2.09, ps ≤ .012. Adolescents in the basic condition, M = 5.63, SE = 0.18, 95% CI = 5.27–5.99, and in the VC condition, M = 5.56, SE = 0.19, 95% CI = 5.19–5.93, also had greater intentions to forgive than those in the defensive condition, M = 4.38, SE = 0.18, 95% CI = 4.02–4.74, 1.18 ≤ M _differences ≤1.25, ps < .001. We did not find a difference between the basic and VC conditions; however, p = .993.

Relationship satisfaction

Conducting an ANCOVA to assess the impact of our manipulation on adolescents' relationship satisfaction revealed a large main effect, F (246, 3) = 28.65, p < .001, partial η² = .26. Tukey post hoc tests showed that adolescents in the no apology condition, M = 3.46, SE = 0.19, 95% CI = 3.09–3.83, and in the defensive condition, M = 4.03, SE = 0.17, 95% CI = 3.69–4.37, anticipated lower relationship satisfaction following the parental response than those in the basic condition, M = 5.45, SE = 0.17, 95% CI = 5.11–5.79, and in the VC conditions, M = 5.26, SE = 0.18, 95% CI = 4.91–5.60, −1.99 ≤ M _differences ≤ −1.23, ps < .001. We found no difference between the no apology and the defensive conditions, p = .122, nor between the basic and the VC conditions, p = .861.

Parental authority

Conducting an ANCOVA to assess the impact of our manipulation on parental authority revealed no significant effect, p = .176. Thus, adolescents perceived their target parent's authority similarly regardless of their assigned apology response.

DISCUSSION

The quality of parent–adolescent relationships plays a key role in adolescent healthy development. However, maintaining high‐quality parent–adolescent relationships is far from effortless, for these relationships are often marked by conflicts (Smetana & Rote, 2019) and situations in which parents, despite having their adolescents' best interest at heart, behave offensively (Mabbe et al., 2018). In this study, we tested the general hypothesis that parental VC (vs. defensive) apologies following parental offenses preserve the parent–adolescent relationship by fostering (H1) forgiveness (including healthier motivations to forgive) and (H2) relationship satisfaction. We also considered their potential negative impact on (H3) parental authority as well as potential interaction effects with the severity of offenses.

Overall, our main results supported our general hypothesis. In phases 1 and 2, they showed that, even after adjusting for autonomy‐supportive parenting, the severity and frequency of parental offenses, and sociodemographics, VC parental apology elements following parental offenses were linked to (H1) greater forgiveness (including more autonomous and less controlled motivations to forgive) and (H2) greater relationship satisfaction. Further, in phase 3, basic and more elaborate VC apologies were found to induce (similarly) higher levels of forgiveness and relationship satisfaction than a lack of apology. In contrast, defensive apology elements only seemed to minimally benefit the relationship. Indeed, though they elicited (H1) greater forgiveness intentions than no apology in phase 3, they were associated with poorer forgiveness intentions than basic and VC apologies in phase 3, less autonomous and more controlled motivations to forgive in phase 1, and (H2) lower relationship satisfaction than basic and VC apologies in phase 3 (i.e., similar to a lack of apology). Thus, to effectively repair their relationship with their adolescents, parents may not only need to apologize to their adolescents, but they may also need to ensure that their apologies are VC rather than defensive.

In addition to highlighting key benefits of VC apology responses and drawbacks of defensive apology responses, our results tempered concerns that parental apologies could (H3) undermine parental authority. Apologies have been proposed to cause a shift in terms of power balance (Shnabel & Nadler, 2008), but we did not find that such a shift compromised parental authority. Rather, we found that coded parental VC apology elements were positively associated with parental authority, and no other significant associations. Further, supplemental analyses revealed that both dimensions of parental authority (i.e., legitimacy and power) were (un)associated with parental apologies in the exact same pattern as their aggregate score (see Supporting Information). Overall, these results are in line with the literature on authoritative parenting style, which suggests that parents can admit their mistakes to their children while retaining their authority (Buri, 1991). Given the key role of parental authority in adolescents' development (Baumrind, 2013), observing orthogonality between apologies and perceived authority offers further evidence that VC parental apologies may be recommended to parents.

Our results also offered preliminary evidence for the need‐supportive and need‐thwarting value of the assessed apology elements. First, VC and defensive parental apology elements were associated with adolescents' motivations to forgive in theoretically expected ways, with VC apologies mostly fostering autonomous forgiveness, and defensive apologies mostly fostering controlled forgiveness (Vansteenkiste & Ryan, 2013). Second, we observed strong relations between VC and defensive parental apologies and other need‐supportive and need‐thwarting parental behaviors (Mageau et al., 2015). Taken together, these results raise the possibility that VC and defensive apology elements could respectively be need‐supportive and need‐thwarting.

Finally, our results suggest that the effectiveness of parental apologies to mend relationships does not vary as a function of the severity of the offenses for which apologies are offered. Indeed, across 20 tested interaction effects, we did not find any significant one. It might nonetheless be important to note that the overall level of offense severity was somewhat low in our sample (i.e., Ms ≤ 3.90 on a 7‐point scale). It thus remains possible that interaction effects do exist, but only when considering the full spectrum of offense severity. The lower level of offense severity may have also played a role in the absence of differences between the basic and VC conditions in phase 3. Studies in other interpersonal settings suggest that including more VC apology elements increases forgiveness particularly after a more severe offense (Kirchhoff et al., 2012). Alternative explanations for the non‐significant differences between the VC and basic conditions include (1) the existence of other moderators (e.g., adolescents' characteristics, such as their beliefs in a just world, Nudelman & Nadler, 2017), and (2) our usage of generic sentences in the VC condition, which may have been less meaningful (and thus less impactful) compared to sentences specifically tailored to each recalled offense. This latter alternative explanation could also lead one to expect that our usage of generic sentences in the defensive condition softened adolescents' unfavorable assessment of this form of apology (as manifested by their greater intentions to forgive, compared to adolescents in the no apology condition).

Theoretical and practical contributions

Our study contributes to the scientific literature on several grounds. First, it contributes to the limited, but growing (Lee et al., 2023; Ruckstaetter et al., 2017), literature on parental apologies by further documenting a key strategy that parents may use with their adolescents in situations that are both typical and potentially damaging to their relationship. Given the key role of the parent–adolescent relationship in adolescents' development (e.g., Barry et al., 2008), it is crucial to identify ways for parents to repair this relationship when hindered. Our study also extends past parental apology research by being the first to directly examine their potential drawbacks on parental authority. In our study, parental authority did not generally relate to parental apologies, apart from one positive association with coded VC apology elements.

Second, our study offers tangible information on how parents should apologize. Overall, results from phases 1 and 2 suggest that (1) including more VC and less defensive elements is preferable, while results from phase 3 nuance that (2) a basic apology can be similarly effective as a more elaborate VC apology, at least for the types of offenses recalled in this study.

Third, our study provides relevant information on the prevalence of VC and defensive elements in parental apology responses. Consistent with the few studies that have coded parental apologies (e.g., Adams‐Clark et al., 2021), our results suggest that parents—like people in other relationship contexts (e.g., Schumann, 2014)—tend to have ample room to grow in the quality of the apologies they offer. For instance, though parents were perceived (phase 1) and coded (phase 2) as including more VC elements than defensive ones following offenses, adolescents only moderately agreed on average that their parents included VC elements in their apologies in phase 1 (M = 4.01 on a 7‐point scale). Further, we only coded 1.07 out of 8 VC elements per parental response on average in phase 2, and coded no VC element for almost half of our sample (see Supporting Information). Finally, two thirds of our sample were coded as including at least one defensive apology element following an offense in phase 2. These results, paired with research showing that teaching individuals how to apologize can improve their apologies (e.g., Gillies et al., 2011), highlight the potential clinical relevance of teaching parents how to apologize in a victim‐centered manner. This contribution is particularly important if we consider that parenting programs do not tend to include parental apologies in their curriculum (Leijten et al., 2019).

Strengths, limitations, and future research

The contributions of our research are strengthened by methodological assets. First, our study is the first to experimentally examine parental apologies, thereby providing preliminary evidence on their potentially causal effects on the parent–adolescent relationship. Second, our research presents two novel, exhaustive, and complementary measurements of (perceived and coded) VC and defensive parental apologies. Given the good psychometric properties of our measures, future research should benefit from relying on the Parental Apology Quality Scale and Parental Apology Coding System to study parental apologies. Finally, our study offers convincing evidence for the positive association between VC parental apologies and healthy parent–adolescent relationships by relying on a multi‐method, multi‐phase approach and replicating the same pattern of results across phases, measures of parental apologies, and indicators of parent–adolescent relationship quality.

Alongside these strengths reside limitations that need to be considered when interpreting our results. First, we did not examine whether our apology elements predict basic needs satisfaction and frustration. As such, our discussion of VC and defensive apology elements as being respectively need‐supportive and need‐thwarting should be interpreted cautiously until further evidence supporting this assumption can be provided. Second, though we assumed (and found evidence for) linear relations between parental apologies and relationship outcomes, it remains possible that non‐linear relations better capture the dynamic at play between parental apologies and adolescent outcomes. In line with our results in phase 3 (where we found no difference between the basic and VC conditions), one may argue that the effectiveness of VC apologies could reach its peak after a given number of elements.

Third, using a single sample for our three phases (rather than independent samples for each phase) may have artificially enhanced the similarity between our results and induced a variety of biases within each phase (especially for phase 3, which was presented after phase 2 and part of phase 1). Fourth, our sample was rather homogeneous (i.e., mostly born in Canada, relatively high socioeconomic status, mostly choosing mothers as target parents), which prevents the generalizability of our findings to families and fathers in less favorable situations and with other cultural belief systems. Finally, other noteworthy limitations and biases include (1) exclusive reliance on adolescents' self‐reports to assess our outcome variables and their associated risk of perceptual biases (especially in phase 1 where we solely used self‐report measurements), (2) false memory or recollection biases (especially in phase 2 where approximately one fourth of parental apology responses could not be coded), (3) external validity issues (especially in phase 3 where we asked participants to imagine their reactions to hypothetical parental apology responses), (4) likelihood of type I error (given that we conducted over 40 inferential tests across phases), and (5) the epistemological and ontological positionality of our research team, which necessarily influenced the coding of parental responses in phase 2.

In addition to addressing these limitations, future research should investigate whether parental apologies can predict changes in the parent–adolescent relationship over short‐ and long‐term periods. Given the body of evidence suggesting that adolescents' behaviors and adjustment can predict changes in parents' behaviors (e.g., Pinquart, 2017), such longitudinal research could also provide valuable insights into potential bidirectional associations between parental apologies and adolescents' responses to these apologies (e.g., using actor‐partner‐interdependence models with parent–adolescent dyads).

CONCLUSION

Parents tend to have their adolescents' best interests at heart but can behave hurtfully towards them at times. Though typical, these offenses can nonetheless hamper the parent–adolescent relationship and in turn adolescents' development. Across three complementary phases, our study highlights the relevance for parents to apologize to their adolescents after behaving offensively towards them to preserve the parent–adolescent relationship. Our study also stresses the importance of favoring VC over defensive apology elements and the potential relevance of teaching parents how to do so. Parenting interventions could benefit from including parental apologies in their larger curriculum as a tool to help parents navigate potential parent–adolescent conflicts and their aftermath.

FUNDING INFORMATION

The Faculté des Études Supérieures et de la Recherche de l'Université de Moncton and the Social Sciences and Humanities Research Council of Canada (SSHRC) funded and facilitated this research through grants to the first author. Funding sources were not involved in any part of the research project.

CONFLICT OF INTEREST STATEMENT

All authors declare no conflict of interest.

CONSENT STATEMENT

All participants signed a consent form. Parents of high school (but not college) students also signed a consent form.

Supporting information

Data S1.

Robichaud, J.‐M. , Schumann, K. , Kil, H. , Koestner, R. , & Mageau, G. A. (2025). Should parents apologize to their adolescents to mend their parent–adolescent relationship? A multi‐method, multi‐phase investigation. Journal of Research on Adolescence, 35, e70024. 10.1111/jora.70024

DATA AVAILABILITY STATEMENT

Data and R script for our analyses are available on Open Science Framework at DOI: 10.17605/OSF.IO/C4Z2A