Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Jul 1.
Published in final edited form as: J Pers Assess. 2022 Feb 22;105(4):487–498. doi: 10.1080/00223891.2022.2039165

Are We Thinking about the Same Disorder? A Trifactor Model Approach to Understand Parents’ and Their Adolescents’ Reports of Borderline Personality Pathology

Salome Vanwoerden 1, Veronica McLaren 2, Stephanie D Stepp 1, Carla Sharp 2
PMCID: PMC9393208  NIHMSID: NIHMS1784158  PMID: 35191795

Abstract

Multiple informant assessment is the norm when evaluating borderline personality pathology (BPP) in adolescence, especially by including reports from both parents and adolescents. However, these reports tend to be discrepant, and it is unclear how to integrate. The current study used a trifactor model to isolate sources of variance in parents’ and adolescents’ reports of BPP due to their shared and unique perspectives in a sample of 652 inpatient adolescents (63% female; Mage = 15.31, SD = 1.45) and their parents (81% mothers). Consensus/agreement was characterized by the externalizing features of BPP whereas idiosyncratic views of adolescent BPP covered the full latent BPP construct, suggesting that simple aggregation of parent and adolescent reports is inappropriate. Measurement invariance suggested that unique perspectives were characterized by slightly different operationalizations of BPP and response biases for specific features of BPP. Attachment security and parents’ interpersonal problems predicted shared and unique perspectives differently for female and male adolescents. Lastly, we found that shared and unique perspectives differentially predicted interview based BPP, length of stay, and adolescent mentalizing. In sum, findings replicate previous evidence of parent-child informant discrepancy in youth psychopathology, broadly, and provide insights specific to BPP. Discussion includes practical recommendations for assessment and interpretation of BPP assessment.

Introduction

Multi-informant assessment is the norm when evaluating youth psychopathology (Dirks et al., 2012). The use of multiple informants allows for an assessment of pathology across various settings and counteracts problems of response bias that exist when relying solely on self-reports of psychopathology. This is particularly relevant for the assessment of personality pathology, which is defined by pervasiveness and persistence across contexts (American Psychiatric Association, 2013). However, a multitude of research has demonstrated that informants are discrepant when describing youth psychopathology, with moderate levels of correspondence, at best, which increases through childhood and adolescence (De Los Reyes et al., 2015). This places a challenge on clinical assessment, with limited consensus in the field about how to integrate information provided by different sources.

While informant discrepancies are the rule when assessing youth psychopathology, it is largely understood that separate informants provide unique and valid perspectives on the youth they are rating, with overall greater predictive value when combinations of raters are used (De Los Reyes et al., 2015). Whereas informant discrepancy has often been viewed a factor to be controlled for, or noise, with valid inferences only assumed when perspectives converge (Garner et al., 1956), it is now recognized that informant discrepancies provide information of how pathology may differ across contexts (De Los Reyes et al., 2013). Thus, discrepancies themselves are valuable pieces of information to interpret (De Los Reyes, 2013; Kaurin et al., 2016). To this end, advanced latent variable modeling has the potential to isolate variance attributable to individual raters as well as their convergence to assist in making empirically-driven decisions about integrating information from multiple informants (Bauer et al., 2013; Martel et al., 2017).

The current study took a latent variable modeling approach to understand the assessment of borderline personality pathology (BPP) in adolescence based on reports by parents and adolescents themselves. BPP is a severe mental illness characterized by affective instability, relationship dysfunction, identity disturbance, and impulsive and self-damaging behaviors (American Psychiatric Association, 2013). Included in the relationship dysfunction experienced by adolescents with BPP is high parent-adolescent conflict (Boucher et al., 2017) and parent-child attachment insecurity (Lyons-Ruth et al., 2015). Research suggests that these relationship factors may contribute to greater discordance in parents’ and adolescents’ understanding and reports of adolescent pathology (Borelli et al., 2019; De Los Reyes & Kazdin, 2006; Treutler & Epkins, 2003). In fact, a few studies have now been published demonstrating moderate discordance in parents and youth reports of BPP, that is stronger when using questionnaire measures (Sharp et al., 2011; Wall et al., 2017, 2019).

Using a questionnaire measure of BPP, Sharp et al. (2011) found that, on average, children rated themselves higher on BPP compared to parents. In two separate clinical samples, Wall and colleagues found discordance between parents and adolescents in their reports of BPP using both questionnaire (Wall et al., 2019) and clinician-administered interviews (Wall et al., 2017). Furthermore, among dyads in which there was agreement regarding high levels of BPP, adolescents demonstrated the highest overall levels of psychopathology in the sample (Wall et al., 2019). Despite this progress, it is notable that the number of studies published related to BPP lags far behind those of internalizing and externalizing pathology in youth. Given that BPP accounts for variance shared across all personality disorders (e.g., Sharp et al., 2015) and with both internalizing and externalizing pathology in adolescence (Vine et al., 2020), findings are informative for a range of psychopathology. However, there remain unanswered questions regarding the meaning and clinical relevance of parent-adolescent informant discrepancy regarding reports of BPP.

For one, it is not clear how clinicians may integrate parents’ and adolescents’ reports of adolescent BPP in practice. When considering different informants reporting on the same assessment instruments, it is necessary to test whether these measures function in the same way across informants, or whether informants hold similar operationalizations of pathological features (Olino & Klein, 2015). Furthermore, there are suggestions that informants vary in their ability to accurately report on certain aspects of psychopathology, with less observable pathology (i.e., internalizing) more likely to yield greater discordance (Dirks et al., 2012). On the other hand, externalizing pathology is more observable such that informants may provide incremental information about these forms of pathology (Markon et al., 2013), as well as external criteria like functioning (Miller et al., 2005). Similar patterns of discrepancy have been found when examining mother-child and father-child dyads; however, mothers and fathers are not interchangeable when reporting on their offspring’s pathology (Weitkamp et al., 2013). In addition to identifying the presence of discordance across various features of BPP, which consists of both internalizing and externalizing pathology, there is value in understanding sources or predictors of discrepancy across different informants. In the parent-child informant discrepancy literature, there is some evidence to suggest that parents’ (especially mothers) own pathology drives discrepant reports of their children’s pathology (Clark et al., 2017; Lohaus et al., 2020; Treutler & Epkins, 2003). Specifically, maternal mental health problems predict reports of higher pathology in offspring (although see Olino et al., 2021 for null results). Also, as reviewed previously, quality of the relationship may also predict discordance. Specifically, insecurely attached adolescents may have views of their own psychopathology are less aligned with parents’ views (Borelli et al., 2019) with no differences based on parent gender (Treutler & Epkins, 2003). Considering differences across mothers and fathers, mother-child relationship quality predicted biases in mothers’ reporting of their children’s externalizing problems whereas father-child relationship quality predicted biases in fathers’ reporting of internalizing problems (Treutler & Epkins, 2003). Lastly, given suggestions that different informants provide incremental or unique information regarding external criteria, it is possible that perspectives from different informants predict differential clinical outcomes. Understanding how informant perspectives provide differential predictive value can thus guide interpretation and use of clinical assessment.

The current study addresses these questions regarding parent-adolescent discrepancies in reports of adolescent BPP using an often-used questionnaire that was administered to families upon adolescents’ admission to a psychiatric hospital. Specifically, we used a trifactor model (depicted in Figure 1), which was developed as a method of analyzing multiple rater data (Bauer et al., 2013). This model is useful in isolating sources of variability due to an adolescent’s level of BPP versus variability due to individual informants. It is also useful to identify which items are most influenced by idiosyncratic views of the informant. In practice, variance in item responses by parents and adolescents are decomposed into either shared variance or variance unique to each informant. The model also isolates specific variance due to individual items; however, these parameters are of less interest in the current study. Shared variance (or the consensus factor) can be interpreted as the shared/consensus view of the adolescents’ level of BPP across parent and adolescent, or agreement. Variance unique to each informant (or each rater’s individual perspective) can be interpreted as parents’ and adolescents’ unique/idiosyncratic view of the adolescents’ level of BPP.

Figure 1.

Figure 1.

Open in a new tab

Unconditional trifactor model.

In addition to testing the fit of a trifactor model, we tested measurement invariance for these factors of shared and unique perspectives to determine whether individual items functioned in the same way across raters. Next, we examined predictors of shared and unique perspectives, focused on parents’ interpersonal problems (as a proxy for personality pathology; Pilkonis et al., 2019) and parent-child attachment security. Lastly, we tested the predictive value of shared and unique perspectives for various clinically relevant outcomes, including length of hospitalization, interview based BPP, and adolescent mentalizing capacity, which is an important mechanism for BPP (Fonagy & Bateman, 2008). Results from these analyses have the potential to uncover whether consensus and/or unique perspectives of BPP have incremental predictive value for outcomes, thus pointing to the clinical relevance of collecting individual reports. Notably, we attempted to include external criterion variables measured with modalities other than questionnaire to avoid criterion contamination.

We hypothesized that the consensus factor of BPP would be most strongly identified by items assessing externalizing features such as risky or impulsive behavior and relationship instability whereas unique perspective factors would be most strongly identified by items assessing the more internalizing features such as identity disturbance and affective instability. We expected that parent-child attachment quality and parents’ interpersonal problems would predict unique perspectives of adolescent BPP, such that poor or lower quality of attachment and greater interpersonal problems would be associated with higher levels of adolescent BPP as represented by individuals’ unique perspectives. Given the novelty of this approach, we had no data-driven a priori hypotheses about which sources of variance (consensus or unique perspectives) would predict relevant outcomes. However, given the role that parents play in guiding treatment decisions of their children, we expected that parents’ unique perspectives would uniquely predict length of hospital stay (i.e., higher parent-perspective of BPP associated with longer length of stay). We also expected that adolescents’ unique perspectives of their level of BPP would positively predict interview assessed BPP, given that interviews were conducted with adolescents alone. In sum, this study has the potential to inform clinical practice by determining aspects of BPP that contribute to shared versus unique views between parents and adolescents, and the clinical value of both shared versus idiosyncratic views of adolescents’ BPP.

Materials and methods

Participants

The current study sample was recruited from an adolescent unit in a private inpatient hospital in a large metropolitan area in the southwestern United States. Eight hundred four consecutive admissions of adolescents ages 12–17 and their parents between 2010 and 2016 were approached to participate. Adolescent patients were eligible for study inclusion if they had sufficient fluency in English to complete all research assessments. Patients were ineligible if they had a diagnosis of autism spectrum disorder or any psychotic spectrum disorder, and/or had an IQ below 70. Based on these criteria, 90 patients were excluded, and an additional 62 patients declined participation. Therefore, a total of 652 participants (63% female; Mage = 15.31, SD = 1.45) and their parents (81% female) were included in the current analysis. Of the 563 participants who elected to report their race, 87.8% were White, 3.4% Asian, 2.1% African American or Black, and 6.8% multiracial or other. Additionally, 6.7% of the 567 who elected to report ethnicity were Hispanic. Psychopathology of adolescents was determined using the Diagnostic Interview Schedule for Children (DISC; Shaffer et al., 2000), a fully structured clinical interview conducted with adolescents based on DSM-IV diagnoses upon admission. 57.9% were diagnosed with a depressive disorder, 7.5% with bipolar disorder, 8.7% with an eating disorder, 42.6% with an externalizing disorder, and 58.9% with an anxiety disorder. Two hundred five (33%) of adolescents who completed the Childhood Interview for Borderline Personality Disorder (CI-BPD; Zanarini, 2003), met DSM diagnostic criteria for BPD. Participants were hospitalized for 35.06 days on average (SD = 13.52); before their admission, adolescents had seen an average of 2.95 (SD = 1.82) therapists and 1.91 (SD = 1.30) psychiatrists or other healthcare providers for mental health; and had on average 1.10 (SD = 1.92) acute psychiatric hospital stays and 0.83 (SD = 1.24) extended psychiatric hospital stays.

Measures

Adolescent borderline personality pathology

The Borderline Personality Features Scale for Children (BPFS-C; Crick et al., 2005) is a measure of BPP for youth ages 9 and older with both a parent-report (Sharp et al., 2011) and child-report (Crick et al., 2005) version. The original BPFS-C has 24 items which are rated on a 5-point Likert scale from 1 (Not at all true) to 5 (Always true). Test-retest reliability, internal consistency, and validity of the BPFS-C have all been supported (Crick et al., 2005); however, the original 4-factor structure has never been supported (e.g., Fossati et al., 2016; Haltigan & Vaillancourt, 2016; Sharp et al., 2012). Thus, Sharp and colleagues refined the measure using IRT methods, which resulted in a shortened 11-item version, in which two items (items 14 and 18) are averaged together (Sharp et al., 2014). The structure of this version of the measure has been validated in independent samples of adolescents (Fossati et al., 2016; Vanwoerden et al., 2019). While we administered the full 24-item version; however, for the current analyses, we only included the 11-items for analysis. This is partly due to previous research support for the 11- versus the 24-item version and due to our own failure to find adequate model fit for factor structure using the 24-item version in this sample. Total scores were calculated for descriptive purposes by summing responses across the 10 items (which included the average of items 14 and 18). In the results and tables, item numbers corresponding to the original 24-item version are listed. In the current sample, internal consistency was α = 0.80 for the parent reports and α = .82 for the adolescent reports.

The Childhood Interview for Borderline Personality Disorder (CI-BPD; Zanarini, 2003) is a semi-structured interview that probes each of the nine diagnostic criteria for BPD in the DSM-IV and section II of the DSM-5. Each criterion is rated on a 0–2 scale (0 if the symptom is absent, 1 if the symptom is probably present, and 2 if the symptom is definitely present). The CI-BPD has shown good interrater reliability and internal consistency and associations with other measures of BPP, suggesting strong external validity (Sharp et al., 2012). For the current study, interviews were conducted by trained research staff with adolescents alone and the sum of these ratings on all 9 criteria were calculated to represent severity of BPP. Interrater reliability of the CI-BPD was evaluated with n = 106 randomly selected interviews that were coded independently by a second rater. For the total score, which was used in analysis, ICC was good (.94) and kappa values for diagnostic classification was also adequate (.74).

Parent interpersonal problems

The Inventory of Interpersonal Problems (IIP; Horowitz et al., 1979) is a 32-item self-report questionnaire used to screen for interpersonal problems. Items are scored on a 5-point Likert scale ranging from 1 (Not at all) to 5 (Extremely). Items are summed to yield a total score and 8 subscale scores corresponding to the style of interpersonal problems. The total score was used in the current study to estimate overall interpersonal functioning; higher total scores reflect more severe interpersonal problems. Internal consistency, test-retest reliability, and validity of the IIP have been demonstrated (Hopwood et al., 2008; Soldz et al., 1995). In the current sample, Cronbach’s alpha was .91.

Parent-adolescent attachment security

The Child Attachment Inventory (CAI; Shmueli-Goetz et al., 2008) is a semi-structured interview assessing attachment organization via children’s mental representations of their attachment figures. The interview consists of 19 open-ended questions concerning the adolescent’s experiences with primary caregivers with prompts for the adolescent to reflect upon each experience. Interviews were videotaped, transcribed, and then coded by a team of trained coders who were certified by the developers of the instrument based on achieving a pre-designated level of reliability. While several scales of attachment are coded from the CAI, for the current study, we relied on the overall coherence scale. This scale integrates all other scales, providing an estimate of overall attachment security, and has been used in previous research as a dimensional proxy for attachment security (Sharp et al., 2016). The CAI was originally designed to be used with children ages 8–13; however, evidence for validity has been demonstrated in adolescent samples, including adequate interrater reliability, concurrent validity, and convergent validity (Venta et al., 2014). For the present study, n = 123 randomly selected interviews (20% of the full sample) were coded by an independent rater, revealing moderate interrater reliability for the coherence scale (ICC = .51).

Adolescent mentalizing capacity

The Movie Assessment of Social Cognition (MASC; Dziobek et al., 2006) is a video-based assessment that evaluates everyday use of mentalizing. Participants watch a 15-minute film about four people getting together for dinner. The film is stopped at 45 points, during which the participants answer a multiple-choice question regarding a character’s thoughts and feelings. Participants choose from 4 answer options, one of which indicates “accurate” mentalizing. A participant’s total score on the measure is the number of items on which the accurate mentalizing answer was selected. The MASC has been found to be a reliable tool for assessing mentalizing ability in adolescents with psychopathology (Fossati et al., 2018).

Procedures

Participants were approached for study participation on the day they were admitted to an inpatient psychiatric hospital by research staff. Parents were first approached to give signed informed consent for participation and, if provided, adolescents were approached for assent. All assessment procedures were conducted in person and privately by trained research staff within 2 weeks from admission date. All aspects of the study were approved by the human ethics research committee at Baylor College of Medicine (H-23579) and the University of Houston (14238-02-4301).

Missing data

There was a small proportion of missing data that was due to nonresponse or adolescents discharging from the hospital before completing the assessment battery. Proportion missing across measures ranged from 4.4% (n = 29; adolescent-report BPFS-C) to 18.7% (n = 122; IIP). Analyses were conducted to determine whether participants with missing data on any of the six measures differed from those with complete data in adolescent age, gender, or level of BPP. Adolescents who did not complete the CI-BPD tended to be younger (M = 14.43, SD = 1.76) than those who did complete the interview (M = 15.35, SD = 1.42; t(30.87) = 2.81, p = .009; Cohen’s d = 0.64); adolescents who did not complete the MASC rated themselves higher on BPP on the self-report BPFS-C (M = 32.06, SD = 6.83) compared to those who did complete the measure (M = 29.64, SD = 7.70; t(128.27) = −3.03, p = .003; Cohen’s d = 0.32); and lastly, adolescents who didn’t complete the self-report BPFS-C were more likely to be male (62.1%) compared to adolescents who did complete the measure (35.6%; χ2(1) = 8.33, p = .004). In terms of parent report measures, adolescents whose parents did not complete the BPFS-C rated themselves lower on BPP on the self-report BPFS-C (M = 26.96, SD = 7.55) compared to those whose parents did complete the measure (M = 30.26, SD = 7.57); t(621) = 3.01, p = .003; Cohen’s d = 0.44). Similarly, adolescents whose parents did not complete the IIP rated themselves lower on BPP on the self-report BPFS-C (M = 28.59, SD = 8.13) compared to those whose parents did complete the measure (M = 30.30, SD = 7.48); t(621) = 2.15, p = .016; Cohen’s d = 0.23). Adolescents with missing responses on both self- or parent-report BPFS-C were excluded from analyses; however, there were only nine adolescents fitting these criteria and they did not differ from the remainder of the sample in terms of gender or age. Full Information Maximum Likelihood (FIML) estimation in MPlus was used to account for missing data, which is based on the assumption of data missing at random (MAR), which is the case when missingness is not random, but can be fully accounted for by variables with complete information. While it is impossible to statistically test the assumption of MAR, authors have stated that this is a reasonable assumption to make (Little & Rubin, 2019) and that FIML provides unbiased estimates in this case. FIML uses all available information to estimate population parameters that would be most likely, given the available sample data.

Data analytic strategy

Descriptive data analyses and bivariate correlations were run using SPSS (IBM Corp, 2016) and structural equation models were conducted in Mplus 8.1 (Muthén & Muthén, 1998). We used the maximum likelihood estimation method given the normal distribution and continuous nature of item response. Model fit was determined by examining the following fit indices: the root-mean-square error of approximation (RMSEA), for which values of less than .08 indicate reasonable fit and values above .10 suggest poor fit (Browne & Cudeck, 1993); the comparative fit index (CFI; Bentler, 1990), for which values between 0.95 and 1.00 indicate excellent fit and values between 0.90 and 0.95 indicate acceptable fit (Hu & Bentler, 1999); and the standardized root-mean-square residual (SRMR), for which values less than .08 indicate acceptable fit (Hu & Bentler, 1999). Chi-square tests are also reported, but these results are not heavily weighted given the chi-square test’s sensitivity to sample size (Fan et al., 1999).

An unconditional trifactor model was fit to the combined dataset following specifications provided by Bauer et al. (2013) as illustrated in Figure 1. This single (between-person) level model decomposes variance in item responses from two different raters (in our case, parent and adolescent) into variance that is shared by the two raters and variance unique to each rater. To do this, all item responses load onto a consensus factor, which reflects shared variability in item responses across parents and adolescents or the consensus/agreement view of adolescents’ BPP. Consensus views can reflect variability in adolescent BPP as well as other sources of shared variability, such as shared beliefs about appropriate emotional functioning, expression, and behavior (Kenny, 1991). Next, two unique perspective factors consisted of loadings from either the parent or adolescent, with the correlation between these factors fixed to zero. These unique perspective factors represent subjective biases in the reporting of adolescent BPP or idiosyncratic perspectives of BPP. Finally, specific item factors were specified by equal loadings from each informant’s response to that item. These specific item factors were set to be orthogonal to one another and all other factors in the model, thus capturing covariation that is unique to each item. To set the scale of the latent variables, we set the means and variances of each factor to zero and one, respectively.

Next, we tested a set of equality constraints to determine whether parents and adolescents engage in similar processes when rating adolescent BPP. Thresholds of <.015 change in CFI and <.01 change in RMSEA were used for these tests (Chen, 2007). This was conducted for both the consensus factor and the two unique perspective factors, in order. Modification indices were used to free constraints and test partial invariance for each of these sets of parameters. Next, equality constraints were added for item intercepts across informants to determine differences in item difficulty across informants. Following specifications by Bauer et al. (2013), we then allowed the factor mean and variance for the unique parent perspective factor to be estimated in order to estimate whether there were differences between parents and adolescents in their perceived levels of BPP of adolescents.

Lastly, we examined predictors of agreement and disagreement by regressing the consensus and two unique perspective factors on parent interpersonal problems and parent-adolescent attachment in a single model. Equality constraints were used to determine whether these regression paths differed between fathers and mothers and between adolescent males and females. We also examined how consensus and perspective factors predicted reports the outcome variables of length of hospital stay, CI-BPD total scores, and adolescent mentalizing capacity (using the MASC) by regressing these variables onto the consensus and two perspective factors in a single model. The same equality constraints were tested based on parent gender and adolescent gender.

Results

Preliminary analyses

Table 1 lists descriptive statistics and correlations between main study variables. Total scores of the BPFS-C reported by parents and adolescents were only correlated (positively) to a small degree. However, based on a paired samples t-test, there was no significant difference between parent- and child-reports of BPP (mean difference = .26; t(570)=.72, p = .469; Cohen’s d = .03). Thus, despite low correspondence, mean levels were similar across different informants. Adolescent-reported BPFS-C scores correlated strongly in the positive direction with the CI-BPD; however, parent reports correlated positively with CI-BPD scores only to a small degree, likely due to these interviews having been conducted with adolescents alone. Small magnitude correlations were found between BPFS-C total scores and parent interpersonal problems (albeit only with parent-report BPFS-C), parent-adolescent attachment security, and adolescent mentalizing capacity in the expected direction. Specifically, parent-reported levels of BPP correlated positively with parent interpersonal problems and levels of BPP correlated negatively with attachment security and mentalizing capacity. Only adolescent-reported BPFS-C correlated positively with length of stay in the hospital. Lasty, BPFS-C scores were not correlated with parent gender or adolescent age; however, females and minorities had higher scores on the BPFS-C based on both parent- and adolescent-report.

Table 1.

Descriptive statistics and correlations between main study variables.

1 2 3 4 5 6 7 8 9 10 Mean (SD)/% Skew Kurtosis
1. Adolescent report BPFS-C 29.99 (7.62) −0.16 −0.30
2. Parent report BPFS-C .25** 29.95 (6.52) −0.06 −0.37
3. CI-BPD total score .58** .25** 8.42 (4.90) 0.10 −1.07
4. Parent interpersonal problems .06 .17** .03 31.63 (16.08) 0.86 0.84
5. Attachment security −.12** −.12** −.17** −.12* 4.39 (1.87) 0.30 −0.55
6. Adolescent mentalizing −.10* −.09* −.12** −.08 .22** 32.16 (4.98) −0.76 0.98
7. Adolescent gender −.19** −.17** −.26** .00 −.08* −.11** 63.2% female 0.55 −1.71
8. Parent gender .01 −.00 .02 −.02 −.11** −.00 .05 81.4% female 1.68 0.81
9. Adolescent age −.05 −.05 −.05 −.02 .15** .28** .11** .05 15.31 (1.45) −0.54 −0.64
10. Length of hospital stay .13** .02 .13** .10* −.07 −.05 −.09* −.01 −.00 35.06 (13.52) 0.69 0.81
11. Child minority status .12** .11** .12** −.00 .01 −.03 −.15** .09* −.02 .06 14.1% minority 1.85 1.42
Open in a new tab

Note:

*

p < .05

**

p < .01.

Unconditional trifactor model

A trifactor model, which accounts for shared and unique perspectives of adolescent borderline pathology from parents and adolescents, fit the data well (χ2(140) = 422.901, p < .001; RMSEA = .056, CFI = .915, SRMR = .069). A fully constrained model based on parent gender fit reasonably well (χ2(360) = 797.329, p < .001; RMSEA = .062, CFI = .869, SRMR = .085); although the freely estimated model across parent genders fit slightly better (χ2(320) = 756.455, p < .001; RMSEA = .065, CFI = .870, SRMR = .076). However, due to power concerns (there were only 112 fathers in the sample), remaining analyses did not differentiate based on parent gender, although differences in results based on parent gender are noted when relevant. Furthermore, we replicated analyses in the subsample of adolescents who participated with their mothers. Results were not substantially different; however, any differences are noted in the text and tables with values are included in supplemental materials.

Test of differential item functioning between parents and adolescents

Differential item functioning was tested for the consensus factor by constraining factor loadings for the same item across informants to be equal. Change in CFI and TLI were below the pre-defined threshold (see Table 2) suggesting that factor loadings onto the consensus factor were comparable for parent- and adolescent-reported items. We similarly constrained factor loadings for the two unique perspective factors to be equal across informants. Change in CFI and TLI were above threshold, so modification indices were followed to remove equality constraints for factor loadings of item 2 (“I feel very lonely.”) and item 9 (“I feel that there is something important missing about me, but I don’t know what it is.”) onto their respective perspective factors. Last, item intercepts were constrained to be equal across informant; however, model fit was substantially worsened, thus modification indices were followed in succession to free equality constraints for three items: item 11 (“I’m careless with things that are important to me.”), item 16 (“I worry that people I care about will leave and not come back.”), and item 6 (“I want to let some people know how much they’ve hurt me.”). This final model (3c) demonstrated acceptable model fit except for the CFI value, which was .896. It is known that CFI values are lower when factor loadings are low (Clark & Bowles, 2018; Heene et al., 2011). Further, we considered fit indices as a whole; because RMSEA and SRMR values were adequate and CFI was only .004 below suggested thresholds, we considered model fit to be adequate. When conducted in the subsample with mothers only, results were unchanged (see Supplementary material, Table S1 for model fit indices).

Table 2.

Model fit indices for unconditional trifactor model.

Model χ 2 df p RMSEA CFI SRMR ΔRMSEA ΔCFI
1 Trifactor – Unconstrained 422.901 140 <.001 0.056 0.915 0.069
2 Trifactor – Constrain consensus loadings 455.744 150 <.001 0.056 0.909 0.071 0.000 0.006
2.1 Trifactor – Constrain perspective loadings 532.921 160 <.001 0.060 0.888 0.095 0.004 0.027
2.1a Free perspective loading for item 2 502.154 159 <.001 0.058 0.897 0.082 0.002 0.018
2.1b Free perspective loading for item 9 486.809 158 <.001 0.057 0.902 0.077 0.001 0.013
3 Trifactor – Constrain intercepts 841.328 166 <.001 0.080 0.798 0.087 0.023 0.104
3a Free threshold for item 11 712.229 165 <.001 0.072 0.836 0.083 0.015 0.066
3b Free threshold for item 16 603.318 164 <.001 0.065 0.869 0.082 0.008 0.033
3c Free threshold for item 6 510.672 163 <.001 0.058 0.896 0.077 0.001 0.006
Open in a new tab

Note. Change in RMSEA and CFI are calculated based on the most constrained model in the previous class (e.g., all model 2’s are compared against model 1 and all model 3’s are compared against model 2.1b).

Unstandardized and standardized factor loadings and item intercepts are displayed in Table 3. There were two items that did not significantly load onto the consensus factor and two additional items that had negative loadings onto the consensus factor. The strongest loadings onto the consensus factor were items 2 (negatively), 11, 15, and 20 suggesting that parent-adolescent consensus about adolescents’ borderline pathology was largely characterized by recklessness and impulsivity. For the most part, factor loadings onto each of the unique perspective factors were larger in magnitude relative to loadings onto the consensus factor, statistically significant, and in the same direction. Thus, unique perspectives were characterized by the broader DSM conceptualization of BPD, which includes a similar emphasis on inter- and intra-personal distress and dysfunction as well as impulsive and reckless behavior. The two items (2 and 9) with differential factor loadings onto the parent and adolescent perspective factors demonstrated stronger loadings for adolescent-report. Thus, the unique perspective factor was defined more strongly by intrapersonal distress for adolescents compared to parents. Examination of differential item intercepts also suggests that at the same level of borderline pathology, adolescents rated themselves higher on items 6 (let people know they’ve hurt me) and 16 (worry that people will leave and not come back) whereas parents rated adolescents higher on item 11 (careless with important things). This suggests that adolescents item responses were more biased toward the endorsement of interpersonal distress whereas parents were more likely to endorse reckless behavior. In the subsample with mothers only, overall interpretation of the final model was not changed although the statistical significance of two factor loadings onto the consensus factor (item 6: “I want to let some people know how much they’ve hurt me.” and item 9: “I feel that there is something important missing about me, but I don’t know what it is.”) changed. However, magnitude of these factor loadings were not substantially different from the model in which mothers and fathers were combined (see Supplementary material, Table S2 for factor loadings and item intercepts).

Table 3.

Factor loading and thresholds for final constrained unconditional trifactor model.

Consensus factor Adolescent perspective Parent perspective Intercept
Adolescent report:
 2. Lonely −.34 (.05)* −.30 .72 (.05) * .63 3.40
 6. Let people know they’ve hurt .10 (.05)* .09 .52 (.04)* .42 3.29
 8. Feelings are strong .19 (.05)* .16 .72 (.04)* .58 3.50
 9. Something important missing about self −.10 (.06) −.08 .90 (.05) * .67 3.31
 11. Careless with important things .35 (.05)* .31 .59 (.04)* .51 2.29
 13. Let down by people −.06 (.05) −.05 .63 (.03)* .54 3.24
 14/18. Go back and forth between feelings .11 (.05)* .10 .68 (.03)* .65 2.99
 15. Get into trouble for doing things without thinking .71 (.07)* .55 .63 (.04)* .49 2.96
 16. Worry that people will leave and not come back −.13 (.06)* −.09 .75 (.04)* .55 3.11
 20. Friends are really mean to each other .23 (.04)* .22 .49 (.03)* .47 2.06
Parent report:
 2. Lonely −.34 (.05)* −.37 .36 (.04) * .39 3.56
 6. Let people know they’ve hurt .10 (.05)* .10 .52 (04)* .47 2.72
 8. Feelings are strong .19 (.05)* .17 .72 (.04)* .63 3.50
 9. Something important missing about self −.10 (.06) −.09 .64 (.05) * .55 3.52
 11. Careless with important things .35 (.05)* .30 .59 (.04)* .51 2.85
 13. Let down by people −.06 (.05) −.05 .63 (.03)* .59 3.24
 14/18. Go back and forth between feelings .11 (.05)* .11 .68 (.03)* .72 2.99
 15. Get into trouble for doing things without thinking .71 (.07)* .55 .63 (.04)* .49 2.96
 16. Worry that people will leave and not come back −.13 (.06)* −.10 .75 (.04)* .58 2.40
 20. Friends are really mean to each other .23 (.04)* .24 .49 (.03)* .52 2.06
Open in a new tab

Note:

*

p < .05; Unstandardized factor loadings (SE) and standardized factor loadings displayed; Abbreviated item content displayed; Bolded and italicized values represent parameters allowed to differ across informants.

In this final, most justifiably constrained factor, we freely estimated the mean and variance of the unique parent perspective factor to estimate differences compared to the child perspective factor. Results demonstrated that parents’ perspective of adolescent BPP (factor mean = 0.04 (.06), p = .517) was not significantly different from the child perspective factor. Thus, when accounting for differential functioning of item responses between parents and adolescents, perceived severity of BPP in adolescents were largely similar. This result was replicated in the subsample with mothers alone.

Predictors of agreement and disagreement between parents and adolescents

We next examined predictors of consensus and idiosyncratic views of parents and adolescents by regressing the consensus factor and both unique perspective factors onto attachment security and parent interpersonal problems. Multi-group models were run to determine whether regression paths differed based on parent gender; however, non-significant change in model fit suggested no differences between these models (χ2difference(6) = 38.34, p = .121) and the more parsimonious model was retained (i.e., effects did not differ based on parent gender). Similarly, multi-group models were run to test whether there were differences in regression paths based on adolescent gender. Results from these models suggested that the effect of attachment security on both consensus and adolescent perspective factors differed based on adolescent gender and the effect of parent interpersonal problems on the parent perspective factor differed based on adolescent gender. Regression effects are displayed in Table 4. Parent-adolescent attachment security was a negative predictor of the consensus factor in females and a negative predictor of the adolescent perspective factor in males. This suggests that attachment security predicts lower levels of BPP when considering the shared perspective of adolescent BPP among parent-female adolescent dyads but predicts adolescent males’ unique perspectives of their levels of BPP (reflected in lower levels of adolescent-perceived BPP). Parent interpersonal problems also uniquely predicted parents’ perspectives of their adolescent boys’ BPP as reflected in higher levels of parent-perceived BPP. In the subsample with only mothers included, we additionally found that the effect of attachment security on mothers’ unique perspectives of their son’s BPP was negative and significant (see Supplementary material, Table S3) such that when considering mother-son dyads, more secure attachment was associated with less severe BPP according to the unique perspectives of both mothers and adolescent boys.

Table 4.

Regression paths for predictors and outcomes of consensus and unique perspective factors.

Predictors: Consensus factor
 Adolescent perspective
 Parent perspective
Male Female Male Female Male Female
Attachment security 0.01 (.04) .01 −0.08 (.03)* −.16 −0.13 (.03)* −21 −0.03 (.02) −.05 −0.04 (.02) −.07 −0.04 (.02) −.08
Parent interpersonal problems 0.00 (.00) .06 0.00 (.00) .06 0.00 (.00) .02 0.00 (.00) .02 0.02 (.00)* .41 0.00 (.00) .05
Outcomes: Unstd. Std. M/F Unstd. Std. M/F Unstd. Std. M/F
Length of hospital stay −1.40 (.72) −.11/−.12 1.87 (.63)* .14/.17 0.56 (.77) .04/.04
CI-BPD total score 0.75 (.27)* .16/.16 2.96 (.20)* .62/.65 0.92 (.24)* .17/.16
Adolescent mentalizing −0.56 (.28)* −.12/−.12 −0.43 (.24) −.09/−.09 −0.62 (.30)* −.12/−.11
Open in a new tab

Note:

*

p < .05; Unstandardized (Unstd.) regression estimates (SE), and standardized (Std.) regression estimates displayed for males and females in model with predictors; Unstandardized regression estimates in model with predictors were constrained to be equal across adolescent gender with standardized estimates displayed for males and females, respectively.

Predictive value of consensus and unique perspective factors

Lastly, we were interested in the utility of consensus and unique views between parent and adolescent as an index of the clinical utility of collecting these reports of adolescent BPP. We regressed length of stay in the hospital, CI-BPD symptom severity, and scores of adolescent mentalizing capacity onto the consensus factor and both unique perspective factors. We constrained these regression paths to be equivalent based on adolescent gender, which did not result in worsened model fit, thus findings are reported for the model in which regression paths are constrained to be equal across adolescent gender. Results are listed in the bottom half of Table 4, which demonstrate that higher levels of BPP based on shared/consensus parent-adolescent views and parents’ unique perspective predicted higher CI-BPD symptoms and lower adolescent mentalizing capacity. Only adolescents’ unique perspectives predicted length of stay in the hospital, with higher levels of BPP predicting longer length of stay. Adolescents’ unique perspectives of their BPP similarly predicted higher levels of CI-BPD symptoms, to a stronger degree than other factors. Thus, while adolescents’ experiences of their own symptomatology appeared to be the driving force in predicting length of hospital stays, adolescents and parents provided both unique and shared information regarding clinician-mediated accounts of adolescent BPP. In the subsample of mothers only, the only difference in results was that the effect of the shared/consensus factor on CI-BPD scores was no longer statistically significant. However, the magnitude of this effect was similar (standardized effect = .10; see Supplementary material, Table S3).

Discussion

The current study applied a relatively novel latent variable modeling approach to understand shared and unique views of parents and adolescents in ratings of adolescent BPP. Our findings both replicate previous findings of moderate convergence across parent and adolescent ratings, but also provide new insights regarding how parents and adolescents uniquely operationalize and report on different features of BPP and how these unique perspectives are both predicted by and predict external variables that are clinically relevant for adolescents and families.

First, we identified three independent factors representing consensus and uniqueness in parents’ and adolescents’ perspectives of adolescent BPP. As hypothesized, consensus or shared perspective was largely characterized by externalizing features of BPP such as reckless and impulsive behavior. This converges with the broader literature on informant discrepancies that states that behaviors that are more observable, like acting out, are more likely to reach convergence across informants (De Los Reyes et al., 2015; Dirks et al., 2012). On the other hand, parents’ and adolescents’ unique perspectives reflected the broader construct of BPP, with similarly strong loadings across intra- and inter-personal dysfunction (including internalizing pathology) and features of recklessness and impulsivity. However, there were some differences in how these unique perspective factors were defined by parents and adolescents. Specifically, identity-related features including loneliness and emptiness were more strongly related to adolescents’ unique perspectives of their BPP relative to parents’. Thus, identity impairment may be a more central construct for adolescents when identifying and reporting on their BPP. This aligns with evidence and theory that identity is a core feature of BPP, and personality pathology in general (Bogaerts et al., 2021; Sharp, 2020; Sharp & Wall, 2021). Further, while identity impairment is a likely driver of adolescents’ experience of their own BPP, this may not be readily observable by parents. Rather, identity impairment may be expressed through inconsistent behavior and interpersonal conflict (Kaufman & Crowell, 2018) that may interfere with parents’ awareness of the underlying intrapersonal cause for these behaviors.

We also found different tendencies of rating items between parents and adolescents when holding the level of latent BPP constant. Specifically, we found that it was easier for adolescents to rate items related to interpersonal or rejection sensitivity whereas it was easier for parents to rate an item relating to recklessness with important possessions (i.e., these items may be endorsed even at low levels of the latent trait of BPP). These differences may reflect discrepancies in personal values and priorities across parents and adolescents that are intuitive from a typical developmental framework. Forming intimate connections with peers and romantic partners is central to socioemotional development during adolescence (Lapsley, 1993; Schwartz et al., 2008; Selman, 2003). As part of this, adolescents are more attuned and responsive to social cues and feedback (Crone & Dahl, 2012). On the other hand, parents are more likely to focus on indicators of self-reliance, self-directedness, and responsibility in their adolescents’ transition to adulthood such that destruction of personal belongings and recklessness will be more salient and endorsed more easily by parents.

Regarding levels of BPP, we found that latent means of the unique perspectives were not statistically different between parents and adolescents, suggesting that within dyads, parents and adolescents are actually in agreement in terms of the severity of BPP in adolescents, despite the differences in how BPP are defined. It is possible that this agreement is due to the particular context in which these data were collected. In this sample, hospital admission was often catalyzed by life threatening behavior (e.g., either suicide attempt or ideation). These situations may have led to increased communication within the family (and with the hospital staff) about adolescents’ symptoms and distress, which may temporarily increase agreement between parents and adolescents regarding levels of BPP. Previous research has found that parent-child informant discrepancies are weaker in magnitude when measured in more restrictive settings (Handwerk et al., 1999) suggesting that greater communication about symptomatology during clinical intake or with increased monitoring during crisis events may provide parents with a greater understanding of their adolescents’ experience.

Despite this, these findings also suggest that parents’ and adolescents’ perspectives of BPP represent unique constructs. Furthermore, parents’ and adolescents’ unique perspectives of adolescent BPP demonstrated different patterns of effects on important clinical outcomes, which suggests that it is clinically useful to have both sets of reports of adolescent BPP. That being said, it is not uncommon for parent and adolescent reports of adolescent BPP to be combined using simple aggregation; however, based on our results, we would recommend keeping parents and adolescent reports of adolescent BPP separate. The exception to this recommendation would be for the externalizing features of BPP, which demonstrated equivalent loadings onto the shared/consensus factor of adolescent BPP. Thus, either single reports or averages across raters would be appropriate as they yield overlapping information about impulsivity and recklessness that comprise BPP. Future research should focus on creating scoring algorithms by comparing percentage variance explained by combinations of reports for relevant clinical correlates (see Martel et al., 2021 for an example). These results provide an important first step in understanding discrepancies in reports of BPP.

Another aim of the current study was to identify predictors of shared and unique perspectives of adolescent BPP, which differed based on adolescent, but not parent gender. Specifically, we found that for female adolescents, attachment security predicted lower levels of BPP in the consensus factor. Thus, insecure attachment is related to how parents and female adolescents collectively view adolescent BPP. However, attachment and parent interpersonal problems were only significant predictors of the unique perspective factors for dyads with male adolescents. Secure attachment predicted lower BPP based on male adolescents’ unique perspectives and parents’ interpersonal problems predicted their unique perspectives of higher BPP in adolescents. This suggests that for male adolescents only, unique views of their BPP are a function of the parent-child attachment relationship (we also found that higher attachment security predicted lower ratings of BPP based on mothers’ unique perspectives). Furthermore, within parent-son dyads, parents’ interpersonal problems (viewed as a proxy for personality pathology) may function to inflate ratings of sons’ BPP. Research on BPP development and parent-child relationships related to BPP has focused on females to a much greater extent than males, with some research suggesting that developmental trajectories of BPP differ across gender (Goodman et al., 2013). It is also the case that differences in parent-son versus daughter relationships are likely to contribute to different factors related to informant discrepancies for adolescent pathology. Greater research is needed to understand the complexities of gender in the parent-child relationship in the context of BPP (Franz & McKinney, 2018).

In terms of the clinical utility of shared and unique perspectives of BPP, we found that consensus regarding BPP predicted both interview based BPP and lower mentalizing abilities in adolescents. Thus, identifying agreement in parents’ and adolescents’ views of BPP has clinical utility. In terms of unique perspectives, both adolescents’ and parents’ perspectives of adolescent BPP predicted incremental variance in interview based BPP. This suggests that interview based assessment of BPP reflects both idiosyncratic and collective views of BPP, even though interviews were conducted with adolescents alone. This points to the utility of interview based assessments in cases when parents may be unavailable to provide information. We also found that length of stay in the hospital was solely predicted by adolescents’ unique perspectives. This was contrary to our expectations given that parents often make treatment-related decisions for their children; however, our findings may suggest that while this is the case for hospital admission, discharge decisions are largely driven by adolescents’ own experiences of their pathology. Thus, adolescents’ feelings of self-efficacy and commitment to safety may be the driving factor in planning discharge. Future research may focus on additional indicators of clinical outcome including re-hospitalization or suicide attempts following discharge. Lastly, we found that parents’ unique perspectives predicted adolescents’ mentalizing abilities whereas adolescents’ unique self-perspectives were unrelated to mentalizing capacity. While the former finding is consistent with the broader idea that increased psychopathology associates with reduced mentalizing capacity, we were surprised by the lack of association between mentalizing and adolescents’ unique perspectives. Because mentalizing capacity is understood to subserve the intrapersonal aspects of personality function (Fonagy et al., 2002; Sharp & Wall, 2021), we would have expected mentalizing capacity to relate to adolescents’ unique perspectives given that the latter were found to be best characterized by features of intrapersonal psychopathology. However, it may be that this association was driven by the association between mentalizing capacity and parent-adolescent consensus on BPP.

There are several limitations to acknowledge in interpreting these findings. For one, our sample was largely White and high SES, which limits generalizability of these results. Our preliminary results demonstrated that adolescents who identified with a minority ethnoracial status had higher scores on the BPFS-C based on parent and adolescent reports. There has been a relatively limited amount of research on differences in BPP across ethnoracial groups, with mixed findings regarding prevalence and severity of BPP across groups (McGilloway et al., 2010 ; Newhill et al., 2009; Swartz et al., 1990; Tomko et al., 2014). However, there are justifiable hypotheses that experiences of BPP may differ across ethnoracial groups due to cultural differences in how emotional symptoms are tolerated, accepted, and understood (Selby & Joiner, 2008) and coping mechanisms (Haliczer et al., 2020). Additionally, overt and systematic racism likely has an important effect on not only how symptoms are expressed, but also reported (Hunter & Schmidt, 2010). However, given the paucity of research to this effect (and the small number of minority adolescents in the current sample), it is critical that this research be replicated and extended with ethnically underrepresented adolescents. A second limitation, as mentioned previously, was that ratings of BPP were collected during hospital admission, often following crisis events. Similar evaluations should be conducted in community and outpatient settings to understand the generalizability of these findings. Third, while we attempted to include external variables that were not measured with self-report to avoid criterion contamination, this was not possible for parents’ interpersonal problems and interview based BPP (which was only completed with adolescents). Future research should examine more diverse indicators of parent pathology collected with multiple modalities. Lastly, mothers were overrepresented in our sample compared to fathers, which may have limited power to detect gender differences in parent perspectives. We conducted analyses in the full sample as well as with the subsample of mothers alone, and results were largely unchanged. However, it is unclear whether this is due to the overrepresentation of mothers in the full sample or because of a true lack of differences across mothers and father. No previous studies on informant discrepancies of adolescent BPP examined mothers and fathers separately, thus this should be addressed in future research.

Despite these limitations, these findings suggest that parents and adolescents hold strong unique perspectives of adolescent BPP that contributes to informant discrepancies observed in practice. Defining features of BPP differ to some extent across parents and adolescents and we identified additional evidence of item response bias. Furthermore, predictors and outcomes of parents’ and adolescents’ unique perspectives differed suggesting there is value in collecting both sources of information when assessing for BPP in adolescents. However, our findings also provide initial considerations when integrating information provided by parents and adolescents that should be expanded upon in future research.

Supplementary Material

Supplementary Material

Funding

This work was supported by the NIMH under Grant F32MH126510; and the McNair Family Foundation (no grant number).

Footnotes

Conflict of interest

No potential competing interest was reported by the authors.

Supplemental data for this article is available online at https://doi.org/10.1080/00223891.2022.2039165.

Data availability

The data that support the findings of this study are available from the corresponding author, SV, upon reasonable request.

References

  1. American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (DSM-5®). American Psychiatric Pub. [Google Scholar]
  2. Bauer DJ, Howard AL, Baldasaro RE, Curran PJ, Hussong AM, Chassin L, & Zucker RA (2013). A trifactor model for integrating ratings across multiple informants. Psychological Methods, 18(4), 475–493. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bentler PM (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107(2), 238–246. 10.1037/0033-2909.107.2.238 [DOI] [PubMed] [Google Scholar]
  4. Bogaerts A, Luyckx K, Bastiaens T, Kaufman EA, & Claes L (2021). Identity impairment as a central dimension in personality pathology. Journal of Psychopathology and Behavioral Assessment, 43(1), 33–42. 10.1007/s10862-020-09804-9 [DOI] [Google Scholar]
  5. Borelli JL, Palmer A, Vanwoerden S, & Sharp C (2019). Convergence in reports of adolescents’ psychopathology: A focus on disorganized attachment and reflective functioning. Journal of Clinical Child & Adolescent Psychology, 48(4), 568–581. 10.1080/15374416.2017.1399400 [DOI] [PubMed] [Google Scholar]
  6. Boucher M-È, Pugliese J, Allard-Chapais C, Lecours S, Ahoundova L, Chouinard R, & Gaham S (2017). Parent-child relationship associated with the development of borderline personality disorder: A systematic review. Personality and Mental Health, 11(4), 229–255. 10.1002/pmh.1385 [DOI] [PubMed] [Google Scholar]
  7. Browne MW, & Cudeck R (1993). Alternative ways of assessing model fit. In Bollen KA & Long JS (Eds.), Testing structural equation models (pp. 136–162). SAGE. [Google Scholar]
  8. Chen FF (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modeling: A Multidisciplinary Journal, 14(3), 464–504. 10.1080/10705510701301834 [DOI] [Google Scholar]
  9. Clark DA, & Bowles RP (2018). Model fit and item factor analysis: Overfactoring, underfactoring, and a program to guide interpretation. Multivariate Behavioral Research, 53(4), 544–558. 10.1080/00273171.2018.1461058 [DOI] [PubMed] [Google Scholar]
  10. Clark DA, Durbin CE, Donnellan MB, & Neppl TK (2017). Internalizing symptoms and personality traits color parental reports of child temperament. Journal of Personality, 85(6), 852–866. [DOI] [PubMed] [Google Scholar]
  11. Crick NR, Murray-Close D, & Woods K (2005). Borderline personality features in childhood: A short-term longitudinal study. Development and Psychopathology, 17(04), 1051–1070. 10.1017/S0954579405050492 [DOI] [PubMed] [Google Scholar]
  12. Crone EA, & Dahl RE (2012). Understanding adolescence as a period of social-affective engagement and goal flexibility. Nature Reviews. Neuroscience, 13(9), 636–650. [DOI] [PubMed] [Google Scholar]
  13. De Los Reyes A (2013). Strategic objectives for improving understanding of informant discrepancies in developmental psychopathology research. Development and Psychopathology, 25(3), 669–682. 10.1017/S0954579413000096 [DOI] [PubMed] [Google Scholar]
  14. De Los Reyes A, Augenstein TM, Wang M, Thomas SA, Drabick DAG, Burgers DE, & Rabinowitz J (2015). The validity of the multi-informant approach to assessing child and adolescent mental health. Psychological Bulletin, 141(4), 858–900. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. De Los Reyes A, & Kazdin AE (2006). Informant discrepancies in assessing child dysfunction relate to dysfunction within mother-child interactions. Journal of Child and Family Studies, 15(5), 643–661. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. De Los Reyes A, Salas S, Menzer MM, & Daruwala SE (2013). Criterion validity of interpreting scores from multi-informant statistical interactions as measures of informant discrepancies in psychological assessments of children and adolescents. Psychological Assessment, 25(2), 509–519. 10.1037/a0032081 [DOI] [PubMed] [Google Scholar]
  17. Dirks MA, Reyes ADL, Briggs-Gowan M, Cella D, & Wakschlag LS (2012). Annual research review: Embracing not erasing contextual variability in children’s behavior-theory and utility in the selection and use of methods and informants in developmental psychopathology. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 53(5), 558–574. 10.1111/j.1469-7610.2012.02537.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Dziobek I, Fleck S, Kalbe E, Rogers K, Hassenstab J, Brand M, Kessler J, Woike JK, Wolf OT, & Convit A (2006). Introducing MASC: A Movie for the Assessment of Social Cognition. Journal of Autism and Developmental Disorders, 36(5), 623–636. [DOI] [PubMed] [Google Scholar]
  19. Fan X, Thompson B, & Wang L (1999). Effects of sample size, estimation methods, and model specification on structural equation modeling fit indexes. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 56–83. 10.1080/10705519909540119 [DOI] [Google Scholar]
  20. Fonagy P, & Bateman A (2008). The development of borderline personality disorder—A mentalizing model. Journal of Personality Disorders, 22(1), 4–21. [DOI] [PubMed] [Google Scholar]
  21. Fonagy P, Gergely G, Jurist EL, & Target M (2002). Affect regulation, mentalization and the development of the self. Other Press. [Google Scholar]
  22. Fossati A, Borroni S, Dziobek I, Fonagy P, & Somma A (2018). Thinking about assessment: Further evidence of the validity of the Movie for the Assessment of Social Cognition as a measure of mentalistic abilities. Psychoanalytic Psychology, 35(1), 127–141. 10.1037/pap0000130 [DOI] [Google Scholar]
  23. Fossati A, Sharp C, Borroni S, & Somma A (2016). Psychometric properties of the Borderline Personality Features Scale for Children-11 (BPFSC-11) in a sample of community dwelling Italian adolescents. European Journal of Psychological Assessment. [Google Scholar]
  24. Franz AO, & McKinney C (2018). Parental and child psychopathology: Moderated mediation by gender and parent–child relationship quality. Child Psychiatry and Human Development, 49(6), 843–852. [DOI] [PubMed] [Google Scholar]
  25. Garner WR, Hake HW, & Eriksen CW (1956). Operationism and the concept of perception. Psychological Review, 63(3), 149–159. [DOI] [PubMed] [Google Scholar]
  26. Goodman M, Patel U, Oakes A, Matho A, & Triebwasser J (2013). Developmental trajectories to male borderline personality disorder. Journal of Personality Disorders, 27(6), 764–782. [DOI] [PubMed] [Google Scholar]
  27. Haliczer LA, Dixon-Gordon KL, Law KC, Anestis MD, Rosenthal MZ, & Chapman AL (2020). Emotion regulation difficulties and borderline personality disorder: The moderating role of race. Personality Disorders: Theory, Research, and Treatment, 11(4), 280–289. 10.1037/per0000355 [DOI] [PubMed] [Google Scholar]
  28. Haltigan JD, & Vaillancourt T (2016). The Borderline Personality Features Scale for Children (BPFS-C): Factor structure and measurement invariance across time and sex in a community-based sample. Journal of Psychopathology and Behavioral Assessment, 38(4), 600–614. 10.1007/s10862-016-9550-1 [DOI] [Google Scholar]
  29. Handwerk ML, Larzelere RE, Soper SH, & Friman PC (1999). Parent and child discrepancies in reporting severity of problem behaviors in three out-of-home settings. Psychological Assessment, 11(1), 14–23. 10.1037/1040-3590.11.1.14 [DOI] [Google Scholar]
  30. Heene M, Hilbert S, Draxler C, Ziegler M, & Bühner M (2011). Masking misfit in confirmatory factor analysis by increasing unique variances: A cautionary note on the usefulness of cutoff values of fit indices. Psychological Methods, 16(3), 319–336. 10.1037/a0024917 [DOI] [PubMed] [Google Scholar]
  31. Hopwood CJ, Pincus AL, DeMoor RM, & Koonce EA (2008). Psychometric characteristics of the Inventory of Interpersonal Problems–Short Circumplex (IIP–SC) with college students. Journal of Personality Assessment, 90(6), 615–618. [DOI] [PubMed] [Google Scholar]
  32. Horowitz LM, Alden LE, Wiggins JS, & Pincus AL (1979). Inventory of interpersonal problems. American Psychological Association. [Google Scholar]
  33. Hu L, & Bentler PM (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. [Database] 10.1080/10705519909540118 [DOI] [Google Scholar]
  34. Hunter LR, & Schmidt NB (2010). Anxiety psychopathology in African American adults: Literature review and development of an empirically informed sociocultural model. Psychological Bulletin, 136(2), 211–235. 10.1037/a0018133 [DOI] [PubMed] [Google Scholar]
  35. IBM Corp. (2016). IBM SPSS statistics for windows, Version 24.0. IBM Corp. [Google Scholar]
  36. Kaufman EA, & Crowell SE (2018). Biological and behavioral mechanisms of identity pathology development: An integrative review. Review of General Psychology, 22(3), 245–263. 10.1037/gpr0000138 [DOI] [Google Scholar]
  37. Kaurin A, Egloff B, Stringaris A, & Wessa M (2016). Only complementary voices tell the truth: A reevaluation of validity in multi-informant approaches of child and adolescent clinical assessments. Journal of Neural Transmission, 123(8), 981–990. [DOI] [PubMed] [Google Scholar]
  38. Kenny DA (1991). A general model of consensus and accuracy in interpersonal perception. Psychological Review, 98(2), 155–163. 10.1037/0033-295x.98.2.155 [DOI] [PubMed] [Google Scholar]
  39. Lapsley DK (1993). Toward an integrated theory of adolescent ego development: The “new look” at adolescent egocentrism. The American Journal of Orthopsychiatry, 63(4), 562–571. 10.1037/h0079470 [DOI] [PubMed] [Google Scholar]
  40. Little RJ, & Rubin DB (2019). Statistical analysis with missing data (Vol. 793). John Wiley & Sons. [Google Scholar]
  41. Lohaus A, Rueth J-E, & Vierhaus M (2020). Cross-informant discrepancies and their association with maternal depression, maternal parenting stress, and mother-child relationship. Journal of Child and Family Studies, 29(3), 867–879. 10.1007/s10826-019-01625-z [DOI] [Google Scholar]
  42. Lyons-Ruth K, Brumariu LE, Bureau J-F, Hennighausen K, & Holmes B (2015). Role confusion and disorientation in young adult-parent interaction among individuals with borderline symptomatology. Journal of Personality Disorders, 29(5), 641–662. [DOI] [PubMed] [Google Scholar]
  43. Markon KE, Quilty LC, Bagby RM, & Krueger RF (2013). The development and psychometric properties of an informant-report form of the personality inventory for DSM-5 (PID-5). Assessment, 20(3), 370–383. [DOI] [PubMed] [Google Scholar]
  44. Martel MM, Eng AG, Bansal PS, Smith TE, Elkins AR, & Goh PK (2021). Multiple informant average integration of ADHD symptom ratings predictive of concurrent and longitudinal impairment. Psychological Assessment, 33(5), 443–451. 10.1037/pas0000994 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Martel MM, Markon K, & Smith GT (2017). Research review: Multi-informant integration in child and adolescent psychopathology diagnosis. Journal of Child Psychology and Psychiatry, 58(2), 116–128. 10.1111/jcpp.12611 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. McGilloway A, Hall RE, Lee T, & Bhui KS (2010). A systematic review of personality disorder, race and ethnicity: Prevalence, aetiology and treatment. BMC Psychiatry, 10(1), 33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Miller JD, Pilkonis PA, & Clifton A (2005). Self- and other-reports of traits from the five-factor model: Relations to personality disorder. Journal of Personality Disorders, 19(4), 400–419. 10.1521/pedi.2005.19.4.400 [DOI] [PubMed] [Google Scholar]
  48. Muthén LK, & Muthén BO (1998). MPlus user’s guide (8th ed.). Muthén & Muthén. [Google Scholar]
  49. Newhill CE, Eack SM, & Conner KO (2009). Racial differences between African and White Americans in the presentation of borderline personality disorder. Race and Social Problems, 1(2), 87–96. 10.1007/s12552-009-9006-2 [DOI] [Google Scholar]
  50. Olino TM, & Klein DN (2015). Psychometric comparison of self- and informant-reports of personality. Assessment, 22(6), 655–664. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Olino TM, Michelini G, Mennies RJ, Kotov R, & Klein DN (2021). Does maternal psychopathology bias reports of offspring symptoms? A study using moderated non-linear factor analysis. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 62(10), 1195–1201. [DOI] [PubMed] [Google Scholar]
  52. Pilkonis PA, Lawrence SM, Johnston KL, & Dodds NE (2019). Screening for personality disorders: A three-item screener from the Inventory of Interpersonal Problems (IIP-3). Journal of Personality Disorders, 33(6), 832–845. [DOI] [PubMed] [Google Scholar]
  53. Schwartz PD, Maynard AM, & Uzelac SM (2008). Adolescent egocentrism: A contemporary view. Adolescence, 43(171), 441–448. [PubMed] [Google Scholar]
  54. Selby EA, & Joiner TE (2008). Ethnic variations in the structure of borderline personality disorder symptomatology. Journal of Psychiatric Research, 43(2), 115–123. [DOI] [PubMed] [Google Scholar]
  55. Selman RL (2003). The promotion of social awareness: Powerful lessons from the partnership of developmental theory and classroom practice. Russell Sage Foundation. https://eric.ed.gov/?id=ED524046 [Google Scholar]
  56. Shaffer D, Fisher P, Lucas CP, Dulcan MK, & Schwab-Stone ME (2000). NIMH Diagnostic Interview Schedule for Children Version IV (NIMH DISC-IV): Description, differences from previous versions, and reliability of some common diagnoses. Journal of the American Academy of Child and Adolescent Psychiatry, 39(1), 28–38. 10.1097/00004583-200001000-00014 [DOI] [PubMed] [Google Scholar]
  57. Sharp C (2020). Adolescent personality pathology and the alternative model for personality disorders: Self development as nexus. Psychopathology, 53(3-4), 198–204. 10.1159/000507588 [DOI] [PubMed] [Google Scholar]
  58. Sharp C, Ha C, Michonski J, Venta A, & Carbone C (2012). Borderline personality disorder in adolescents: Evidence in support of the Childhood Interview for DSM-IV Borderline Personality Disorder in a sample of adolescent inpatients. Comprehensive Psychiatry, 53(6), 765–774. 10.1016/j.comppsych.2011.12.003 [DOI] [PubMed] [Google Scholar]
  59. Sharp C, Mosko O, Chang B, & Ha C (2011). The cross-informant concordance and concurrent validity of the Borderline Personality Features Scale for Children in a community sample of boys. Clinical Child Psychology and Psychiatry, 16(3), 335–349. 10.1177/1359104510366279 [DOI] [PubMed] [Google Scholar]
  60. Sharp C, Steinberg L, Temple J, & Newlin E (2014). An 11-item measure to assess borderline traits in adolescents: Refinement of the BPFSC using IRT. Personality Disorders: Theory, Research, and Treatment, 5(1), 70–78. 10.1037/per0000057 [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Sharp C, Venta A, Vanwoerden S, Schramm A, Ha C, Newlin E, Reddy R, & Fonagy P (2016). First empirical evaluation of the link between attachment, social cognition and borderline features in adolescents. Comprehensive Psychiatry, 64, 4–11. [DOI] [PubMed] [Google Scholar]
  62. Sharp C, & Wall K (2021). DSM-5 level of personality functioning: Refocusing personality disorder on what it means to be human. Annual Review of Clinical Psychology, 17(1), 313–337. 10.1146/annurev-clinpsy-081219-105402 [DOI] [PubMed] [Google Scholar]
  63. Sharp C, Wright AGC, Fowler JC, Frueh BC, Allen JG, Oldham J, & Clark LA (2015). The structure of personality pathology: Both general (‘g’) and specific (‘s’) factors? Journal of Abnormal Psychology, 124(2), 387–398. 10.1037/abn0000033 [DOI] [PubMed] [Google Scholar]
  64. Shmueli-Goetz Y, Target M, Fonagy P, & Datta A (2008). The child attachment interview: A psychometric study of reliability and validity. Developmental Psychopathology, 44(4), 939–956. [DOI] [PubMed] [Google Scholar]
  65. Soldz S, Budman S, Demby A, & Merry J (1995). A Short Form of the Inventory of interpersonal problems circumples scales. Assessment, 2(1), 53–63. 10.1177/1073191195002001006 [DOI] [Google Scholar]
  66. Swartz M, Blazer D, George L, & Winfield I (1990). Estimating the prevalence of borderline personality disorder in the community. Journal of Personality Disorders, 4(3), 257–272. 10.1521/pedi.1990.4.3.257 [DOI] [Google Scholar]
  67. Tomko RL, Trull TJ, Wood PK, & Sher KJ (2014). Characteristics of borderline personality disorder in a community sample: Comorbidity, treatment utilization, and general functioning. Journal of Personality Disorders, 28(5), 734–750. 10.1521/pedi_2012_26_093 [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Treutler CM, & Epkins CC (2003). Are discrepancies among child, mother, and father reports on children’s behavior related to parents’ psychological symptoms and aspects of parent-child relationships? Journal of Abnormal Child Psychology, 31(1), 13–27. [DOI] [PubMed] [Google Scholar]
  69. Vanwoerden S, Garey L, Ferguson T, Temple JR, & Sharp C (2019). Borderline Personality Features Scale for Children-11: Measurement invariance over time and across gender in a community sample of adolescents. Psychological Assessment, 31(1), 114–119. 10.1037/pas0000640 [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Venta A, Shmueli-Goetz Y, & Sharp C (2014). Assessing attachment in adolescence: A psychometric study of the Child Attachment Interview. Psychological Assessment, 26(1), 238–255. 10.1037/a0034712 [DOI] [PubMed] [Google Scholar]
  71. Vine V, Byrd AL, Mohr H, Scott LN, Beeney JE, & Stepp SD (2020). The structure of psychopathology in a sample of clinically referred, emotionally dysregulated early adolescents. Journal of Abnormal Child Psychology, 48(11), 1379–1393. 10.1007/s10802-020-00684-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Wall K, Ahmed Y, & Sharp C (2019). Parent-adolescent concordance in borderline pathology and why it matters. Journal of Abnormal Child Psychology, 47(3), 529–542. [DOI] [PubMed] [Google Scholar]
  73. Wall K, Sharp C, Ahmed Y, Goodman M, & Zanarini MC (2017). Parent–adolescent concordance on the Revised Diagnostic Interview for Borderlines (DIB-R) and the Childhood Interview for Borderline Personality Disorder (CI-BPD). Personality and Mental Health, 11(3), 179–188. [DOI] [PubMed] [Google Scholar]
  74. Weitkamp K, Daniels J, Rosenthal S, Romer G, & Wiegand-Grefe S (2013). Health-related quality of life: Cross-informant agreement of father, mother, and self-report for children and adolescents in outpatient psychotherapy treatment. Child and Adolescent Mental Health, 18(2), 88–94. [DOI] [PubMed] [Google Scholar]
  75. Zanarini MC (2003). Childhood Interview for DSM-IV borderline personality disorder (CI-BPD). McLean Hospital [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material
NIHMS1784158-supplement-Supplementary_Material.docx (30.5KB, docx)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author, SV, upon reasonable request.

RESOURCES