Introduction
Collateral informants, usually parents, are routinely included in the diagnosis of mental health disorders in children and adolescents. Despite the recommendation to always collect a collateral report (Meyer et al., 2001), this practice is rarely seen in the diagnostic process for adults. This is not surprising, as reports from multiple informants usually differ, leaving researchers and clinicians with the task of interpreting discrepant results. However, decades of research, mostly in the field of child and adolescent psychopathology, have shown that informant discrepancies indicate more than just measurement error (De Los Reyes et al., 2013) and may contain relevant incremental information (De Los Reyes et al., 2022). Discrepancies might arise from different perspectives, including different roles or knowledge of behavior, and different personal characteristics among informants (Achenbach et al., 2005). Kraemer et al. (2003) suggested a satellite model, meaning that the combination of multiple sources of information would yield the most accurate picture of an individual’s trait of interest. This implies that gaining a comprehensive understanding of an individual’s psychopathology requires collateral reports even beyond their 18th birthday. The importance of collateral reports in the diagnosis of mental disorders is emphasized by studies showing that diagnoses solely based on self-reports diverge from those based on reports from multiple informants (Kuhn et al., 2017; Meyer et al., 2001; van Dulmen & Egeland, 2011; Youngstrom et al., 2003).
Self- and collateral reports can be collected in different forms, e.g., questionnaires or interviews. As interviews are usually time-consuming to administer and interpret and often require trained professionals, questionnaires are more economical and therefore widely used to gather self- and collateral reports on symptoms of psychopathology. Certain questionnaires provide parallel forms for self- and collateral reports, such as the CBCL/YSR questionnaires for children and adolescents (Achenbach, 1991a, b) and the equivalent ASEBA questionnaires for adults (Achenbach & Rescorla, 2003). The use of parallel questionnaires significantly improves the comparability of different informants’ reports (De Los Reyes et al., 2015). Most research on informant discrepancies, including the findings presented next, has therefore been conducted using parallel questionnaires.
In clinical care, reports on psychopathology are used to assess needs and set goals for treatment; thus, this process can be improved by taking reports from multiple informants into account (De Los Reyes et al., 2022). Clinical research forms the basis of our understanding of mental health disorders and their treatment, which requires a thorough assessment of psychopathology. As outlined above, the best practice to achieve this would also be to use reports from multiple informants. A special case, but very relevant in clinical research, are longitudinal studies on psychopathology that begin in childhood. As children are usually too young to report on their mental health at the onset of these studies, their parents provide collateral ratings. As children get older, however, they become able to provide self-reports. As reports from different informants are not interchangeable (De Los Reyes et al., 2015), it is difficult for researchers to determine who to collect reports from and how to interpret them. A better understanding of agreement between different informants’ reports and factors related to informant discrepancies in adulthood is therefore vital for both clinical care and research.
Among children and adolescents, agreement between self- and parent reports is moderate (Rescorla et al., 2013), and different factors have been shown to be associated with informant discrepancies in childhood and adolescence. Parental mental health problems (Affrunti & Woodruff-Borden, 2015; Ehrlich et al., 2011; Fjermestad et al., 2017) and lower socioeconomic status (Stone et al., 2013) are related to larger discrepancies, and a better parent‒child relationship is related to smaller discrepancies (Ehrlich et al., 2011; Lohaus, Rueth & Vierhaus, 2020). Most research on the associations of informant discrepancies regarding child psychopathology has investigated differences in agreement about internalizing and externalizing problems (e.g., Rescorla et al., 2013). Overall cross-informant agreement including different informant pairs (e.g., parent‒child, parent-teacher, child-teacher) seems to be higher for externalizing than internalizing problems, indicating that agreement might be higher for problems that are easily observable by collateral informants (De Los Reyes et al., 2015). However, no significant differences in the level of agreement were found between internalizing and externalizing problems when only taking adolescents’ self-reports and collateral parent reports into account (Rescorla et al., 2013). Research on associations of agreement with the severity of mental health problems in children is still lacking. Radicke et al. (2021), however, report higher agreement when children experience more psychological difficulties. Associations with child gender have been widely examined, but the results remain inconclusive (De Los Reyes & Kazdin, 2005).
Research on informant discrepancies in adulthood is still scarce and has thus far focused on the level of agreement between self- and collateral reports without investigating factors associated with informant discrepancies or the longitudinal development of informant agreement. Achenbach et al. (2005) and Rescorla et al. (2016) reported moderate agreement between adults’ self-reports and different collateral informants’ reports of adult psychopathology and no differences in internalizing and externalizing problems. Factors that moderate the levels of correspondence between self- and collateral reports in adulthood still need to be identified (De Los Reyes et al., 2020). Moreover, most of our insight on informant discrepancies in adulthood is based on studies conducted on individuals aged 18 to 59 years that did not differentiate between different types of collateral informants, such as spouses, partners, family members or friends (e.g., Achenbach et al., 2005; Rescorla et al., 2016). This is (1) a very wide age range covering diverse phases of life and (2) a heterogeneous sample of informants that usually differ in their knowledge of the index person. It is not yet clear whether agreement varies among different age groups or different types of collateral informants. More research on cross-informant agreement is therefore needed to determine whether informant discrepancies change throughout the life course or among different collateral informants.
A particularly interesting age range to investigate cross-informant agreement is so-called emerging adulthood. Emerging adulthood describes the age range from 18 years old to the mid-/late twenties and is now considered a separate and important period of the life course, meaning that individuals in this age group are neither adolescents nor adults (Arnett, 2007). Development in this period is very heterogeneous and can be very challenging, as it is characterized by identity exploration, important decisions, and experiences, e.g., regarding romantic relationships and work (Arnett, 2007). The relevance of this period is emphasized by the prevalence of mental health problems in this age group: more than 50% of this population meet the criteria for at least one lifetime mental disorder (Kessler & Wang, 2008), and the 12-month prevalence rate is higher than that in any other age group (Jacobi & Groß, 2014). Emerging adulthood therefore is an important period, which makes a thorough psychopathological assessment to obtain valid diagnoses crucial for clinical interventions as well as research. An important factor for successful development and, thus, mental health in emerging adults (EAs) is the parent‒child relationship (Reed-Fitzke et al., 2021; Steele & McKinney, 2019). Although the parent‒child relationship transitions throughout adolescence, it remains important for EAs’ wellbeing as they develop autonomy and individuation (Parra, Oliva & del Carmen Reina, 2015; Tsai et al., 2013). The age of moving out of the parents’ household has risen significantly during the last decades (Seiffge-Krenke, 2013), and many EAs have not yet committed to long term relationships (Shulman & Connolly, 2013), making it difficult to routinely assess collateral reports from partners. In childhood and adolescents, teachers are often asked to provide collateral reports. However, due to the diverse educational and professional development in emerging adulthood, teachers can no longer be considered as standard informants. Therefore, parents remain key informants regarding EAs’ mental health problems, as they still play an important role in EAs’ lives. Due to the unique developmental challenges, the evolving parent‒child relationship, and the high prevalence of mental disorders, emerging adulthood is a critical period of life. Informant discrepancies, especially between self- and parent reports, must be investigated thoroughly to both obtain accurate research results and provide optimal clinical care.
Despite the crucial importance of emerging adulthood, thus far, only one study has examined parent‒child agreement on child psychopathology in this age group: Szkody et al. (2022) investigated a sample of N = 129 undergraduate psychology students aged 18 to 25 years (M = 18.4 years, SD = 0.9) from a southern United States university and N = 98 of their parents (25% male) using the ASEBA questionnaires (Achenbach & Rescorla, 2003). They found moderate to high associations between the parents’ and EAs’ reports of psychopathology (r =.32 to r =.50). Discrepancies in self- and parent-reported depression problems in EAs were significantly and negatively associated with affection in the parent-EA relationship. Higher levels of parent-reported parental depressive and antisocial symptoms were significantly linked to larger discrepancies regarding antisocial problems in EAs. Szkody et al. (2022) also reported that affection quality increased when parents’ reports of depressive symptoms in their EA children was higher than the EAs’ self-reports and, therefore, concluded that the direction of the discrepancy might be relevant. They emphasized that in addition to psychopathology among EAs and their parents, the parent-EA relationship is also associated with informant discrepancies. As this is the only study thus far, more research that examines discrepancies between parent- and self-reports on psychopathology and their associations in emerging adulthood is needed. Specifically, (1) more diverse samples are needed (not only university students), (2) factors that have been shown to be associated with informant discrepancies in childhood and adolescence should be examined, and (3) the direction of informant discrepancies needs to be investigated more closely.
Aims
The study at hand aimed to advance the existing research literature by investigating discrepancies between maternal- and self-reported symptoms of psychopathology in EA in a sample from a longitudinal prevention study. We aimed (1) to analyze agreement between self- and maternal reports about internalizing, externalizing, and total mental health problems in EAs and (2) to examine how closely self- and maternal reports agree with clinicians’ ratings of mental disorders in EAs. We further aimed (3) to identify factors that are related to discrepancies between maternal and self-reports of internalizing, externalizing, and total mental health problems in emerging adulthood, considering the direction of discrepancies. Based on findings in childhood and adolescence, we examined maternal psychopathology, current psychological distress in the emerging adults themselves and their mothers, the mother-child relationship (relationship quality and frequency of contact), SES, and child gender.
Methods
Procedure
Data were collected as part of the DFG-funded project (German Research Foundation) Future Family IV (FF-IV, funding Code: JO 1632/1-1), which is the 18-year catamnesis of the longitudinal prevention study Future Family (FF; Heinrichs et al., 2005; Heinrichs & Jensen-Doss, 2010; Supke et al., 2021). The FF studies investigated the effectiveness of a preventive parenting training and the development of different child and family characteristics. In 2001, N = 280 families were recruited from different kindergartens in Braunschweig, Germany (FF-I). As families with lower SES were underrepresented in the FF-I sample, in 2003, an additional N = 197 families were recruited from kindergartens in socially disadvantaged areas of Braunschweig (FF-II; see Heinrichs & Jensen-Doss, 2010, for further information on study design and recruitment). For the 18-year catamnesis (FF-IV), families that had previously participated in the FF project were contacted by post and informed about the new assessment. A total of N = 316 families from the FF-I and FF-II agreed to participate in the FF-IV, which equals a retention rate of 67%. Six families were excluded because they did not meet the inclusion criteria. The assessment took place from January 2020 to January 2022 and included a 1.5-hour interview on the current life situation, a structured clinical interview, and a set of questionnaires. EAs and their parents were assessed separately. Due to the COVID-19 pandemic, most interviews were conducted via telephone. Questionnaires were completed online or as paper-pencil questionnaires. The interview started with questions about the participants’ current life situations, followed by the clinical interview. Participants gave their written informed consent, and each received a 50€ financial incentive for completing both the interview and questionnaires (30€ if only the questionnaires were answered). The ethics committee of the University of Braunschweig reviewed and approved all procedures (D-2020-09).
Participants
A total of N = 294 emerging adults and N = 292 parents (n = 261 mothers, n = 18 fathers, n = 13 mother-father dyads) participated in the 18-year catamnesis. For the following analyses, we excluded fathers as well as families in which only one member (parent or child) participated, leading to a total sample size of N = 256 mother-child dyads.
The EAs were aged 19 to 26 years (M = 22.3 years; SD = 1.2), 50% (n = 129) were female, and 32% (n = 83) still lived with their parents. Mothers had a M age of 53.6 years (SD = 4.7), 98% (n = 252) were their child’s biological mother, and 2% (n = 4) were the stepmother, foster mother, or adoptive mother. Most families had a high socioeconomic status (54%, n = 137), only 2% (n = 4) had a low SES, and 16% (n = 42) of the sample had a migration background.
Measures
Self- and Parent Reported Symptoms of Psychopathology in Emerging Adults
Two parallel questionnaires, the Adult Self Report (ASR) and Adult Behavior Checklist (ABCL; Achenbach & Rescorla, 2003), were used to measure symptoms of psychopathology in the EAs that occurred in the previous six months. Each questionnaire consists of 126 items that are rated on a three-point Likert scale (0 = not true, 1 = somewhat or sometimes true, 2 = very true or often true). Items can be summed to form the Internalizing, Externalizing, and Total Problems scales. The Total Problems score is based on the items of the Internalizing and Externalizing scales, as well as additional problem items. All three scales were used in the following analyses to assess a broad spectrum of symptoms of psychopathology. Higher scores indicated higher levels of mental health problems. Both questionnaires showed high internal consistency in the current sample (ASR: Total Problems α = .96, 95% CI [.96, .97], Internalizing α = .94, 95% CI [.92, .95], Externalizing α = .90, 95% CI [.88, .92]; ABCL: Total Problems α = .97, 95% CI [.96, .97], Internalizing α = .91, 95% CI [.90, .93], Externalizing α = .92, 95% CI [.90, .93]), and high test-retest reliability has been demonstrated for both questionnaires (ASR: Total Problems r = .94, Internalizing r = .89, Externalizing r = .91; ABCL: Total Problems r = .92, Internalizing r = .80, Externalizing r = .92; Achenbach & Rescorla, 2003).
Clinicians’ Ratings of Mental Disorders
The emerging adults and their mothers completed a half-structured clinical interview (Diagnostic Interview for Mental Disorders, DIPS-OA; Margraf et al., 2017). The DIPS-OA includes the most common mental disorders (e.g., anxiety disorders, depressive disorders, trauma and stress disorders, substance use disorders) and enables differential diagnoses according to the DSM-V and ICD-10. Multiple studies on large samples have proven good validity as well as high interrater-reliability and retest-reliability (Margraf, Cwik, Pflug & SchneideMargraf et al., 2017a, b). The interview was conducted by five licensed interviewers. Current diagnoses were dichotomized (none vs. one or more diagnoses) for the areas of all mental health disorders and internalizing disorders. Internalizing disorders were operationalized according to the ABCL/ASR Internalizing scale (including depressive disorders, anxiety disorders except specific phobias and agoraphobia, and somatization disorder). Externalizing disorders (operationalized according to the ABCL/ASR Externalizing scale) could not be analyzed separately due to the low number of cases in the current sample (n = 2).
Current Psychological Distress
Current psychological distress in the EAs and their mothers was assessed by the Depression module of the Patient Health Questionnaire (PHQ-9; Kroenke et al., 2001) and the Generalized Anxiety Disorder Questionnaire (GAD-7; Löwe et al., 2008). The PHQ-9 consists of nine items measuring depressive symptoms during the last two weeks on a four-point Likert scale (0 = not at all– 3 = almost every day). The seven items of the GAD-7 questionnaire assess the frequency of worry and symptoms of anxiety during the last two weeks on the same rating scale as the PHQ-9. For the following analyses, the scores of the PHQ-9 and GAD-7 were summed as an indicator of general psychological distress, with higher values indicating higher levels of distress. Both questionnaires showed high internal consistency in the current sample (PHQ-9: α = .87, 95% CI [.85, .89] in EAs and α = .85, 95% CI [.82, .88] in their mothers; GAD-7: α = .87, 95% CI [.84, .89] in EAs and their mothers; total score: α = .93, 95% CI [.91, .94] in EAs and α = .92, 95% CI [.90, .93] in their mothers).
Parent-EA Relationship
EAs reported on the current quality of the parent-EA relationship and the frequency of contact using two individual items (“How would you describe your relationship with your mother right now?” and “How often are you currently in contact with your mother?”) that were developed within the FF-IV project and rated on a six-point scale (1 = very bad, 2 = bad, 3 = rather bad, 4 = rather good, 5 = good, 6 = very good for relationship quality; 0 = no contact at all, 1 = less than once a month, 2 = 1–3 times a month, 3 = once a week, 4 = multiple times a week, 5 = daily contact for frequency).
Socioeconomic Status (SES)
Mothers provided information about their education, occupation, and household income over the last month. The answers were scored and combined into an index value, with cutoff values for low, medium, and high SES (based on the procedure suggested by Lampert et al., 2018).
Data Analysis
Statistical analyses were conducted using IBM SPSS Statistics version 28, RStudio 2022.07.2, and psych package version 2.2.9 for R (Revelle, 2022). Agreement between self- and parent reports of Total Problems, Internalizing and Externalizing was analyzed by Pearson correlations. Agreement with the clinicians’ ratings of (1) all mental disorders and (2) internalizing disorders was analyzed by point biserial correlations. According to Cohen (1988), correlations of ǀrǀ = .10 were regarded as small, ǀrǀ = .30 were regarded as medium, and ǀrǀ = .50 were regarded as large. Factors related to discrepancies between self- and parent reports were examined using relative weight analyses, which allow to determine the relevance of different predictors (Lorenzo-Seva et al., 2010). In relative weight analyses, predictors are first transformed into orthogonal variables, the criterion is then regressed onto these orthogonal variables. Next, the orthogonal variables are related back to the original predictor variables, and relative weights are generated (for more information see Tonidandel et al., 2009). Relative weight analyses do not include information on statistical significance, however, statistical significance can be determined by creating confidence intervals for standardized regression weights (Lorenzo-Seva et al., 2010; Tonidandel et al., 2009). As recommended by De Los Reyes and Kazdin (2004), standardized difference scores were used as a measure of discrepancy between parent and self-reports. ASR and ABCL scores were z-standardized, and then the mother’s ABCL z score was subtracted from the EA’s ASR z score. The clinicians’ ratings of the mothers’ mental health, the current psychological distress of the EAs and their mothers, the quality of the mother-EA relationship, the frequency of contact, SES, and EA gender were used as predictors, the standardized difference scores as criterion. As this procedure did not allow us to analyze interaction effects, analyses were carried out separately for positive and negative discrepancies. Based on Tonidandel et al. (2009), we report zero-order correlations, standardized regression weights, and relative weights in the results. Boxplots were used to identify outliers. Assumptions of normality, linearity, and homoscedasticity were analyzed graphically using histograms and zresid vs. zpred plots.
A power analysis was conducted using G*Power version 3.1.9.7 (Faul et al., 2007) and revealed sufficient power for all analyses (.91 to .99).
Results
Agreement Between Parent Reports and EA Self-Reports
Correlations between EA symptoms of psychopathology reported by the EAs (ASR) and their mothers (ABCL) were moderate to high, with an r =.53 (p <.001) for Internalizing, r =.48 (p <.001) for Externalizing, and r =.53 (p <.001) for Total Problems. Post hoc analysis revealed no significant difference between Internalizing and Externalizing (p =.33).
Agreement with Clinical Ratings
Among EAs, n = 247 (96.5%) participated in the clinical interview. A total of 35.6% (n = 88) met the criteria for at least one mental disorder, and 21.5% (n = 53) met the criteria for at least one internalizing disorder. We found significant associations between the clinicians’ ratings of EA’ mental disorders and the EAs’ self-reports (ASR) and their mothers’ reports (ABCL) of EA psychopathology. The clinicians’ ratings of all mental disorders correlated significantly with ASR Total Problems (r =.58, p <.001) and ABCL Total Problems (r =.43, p <.001). The clinicians’ rating of internalizing disorders was significantly correlated with ASR Internalizing (r =.58, p <.001) and ABCL Internalizing (r =.37, p <.001). Post hoc analyses revealed that the self-reports had significantly higher correlations with the clinicians’ ratings than the mothers’ reports (Total Problems: p <.003; Internalizing: p <.001). The correlations of the clinicians’ ratings of all mental disorders with ABCL Total Problems, and of the clinicians’ ratings of internalizing disorders with ABCL Internalizing were not significantly different (p =.22). The correlations of the clinicians’ ratings with ASR Total Problems and ASR Internalizing (r =.58 each) were evidently not different.
Factors Related to Discrepancies
The number of outliers in the different regression criteria ranged from five to seven. Analyses of assumptions showed a normal distribution of residuals for positive discrepancies but a negatively skewed distribution of residuals for negative discrepancies. Zresid vs. zpred plots showed no violations of linearity but heteroscedasticity was shown for both positive and negative discrepancies. Tolerance and VIF identified no multicollinearity. These patterns were consistent in Total problems, Internalizing, and Externalizing. Therefore, bootstrapping (1000 samples) was performed to calculate model parameters and confidence intervals through robust significance tests.
Table 1 shows the distribution of the mother-EA dyads with positive and negative discrepancies for Total Problems, Internalizing, and Externalizing. N = 76 (30.8%) dyads showed positive discrepancies and n = 92 (37.2%) dyads showed negative discrepancies on all three scales. Descriptive statistics of all measures can be found in the electronic supplement.
Table 1
Distribution of mother-child dyads with positive and negative standardized difference scores
n positive discrepancy (ASR > ABCL) | n negative discrepancy (ASR < ABCL) | |
|---|---|---|
Total Problems | 120 (48.6%) | 127 (51.4%) |
Internalizing | 121 (49.0%) | 126 (51.0%) |
Externalizing | 112 (45.3%) | 135 (54.7%) |
Total Problems
N = 120 (48.6%) mother-EA dyads had a positive discrepancy, meaning that the child’s standardized self-report (ASR) score was higher than the standardized maternal report (ABCL) score. The model yielded a significant result (R² = .423), but only the EA’s and mother’s level of current psychological distress (PHQ-9 + GAD-7 scores) significantly predicted a positive discrepancy (see Table 2). When considering the relative weight, however, mothers’ psychological distress only explained 4% of the model R², indicating poor relevance of this predictor. EAs’ psychological distress, on the other hand, explained 86% of the variance. Higher levels of current psychological distress in EAs thus were associated with larger positive discrepancies between self- and maternal reports. EAs with higher levels of current psychological distress therefore seemed to rate their symptoms of psychopathology within the last six months as more severe than their mothers. N = 127 (51.4%) mother-EA dyads had a negative discrepancy, meaning that the standardized maternal report score was higher than the EA’s standardized self-report score. This model also turned out to be significant (R² = .143); however, this time, only the mother’s level of current psychological distress (PHQ-9 + GAD-7 scores) significantly predicted a negative discrepancy (see Table 2), meaning that higher levels of maternal psychological distress were associated with larger negative discrepancies between self- and maternal reports. This predictor explained 61% of the predictable variance in mother-child discrepancies, making it highly relevant. Mothers with higher levels of current psychological distress therefore rated the EAs’ symptoms of psychopathology as more severe than the EAs themselves.
Table 2
Zero-order correlations, standardized regression weights, and relative weights for ASR/ABCL total problems
95% CI | |||||
|---|---|---|---|---|---|
r | β | Lower | Upper | Relative weight (%) | |
Total Problems– positive discrepancy, N = 120, R² = .423 | |||||
PHQ-9 + GAD-7 EA | .62 | .69* | .52 | .83 | 85.6 |
PHQ-9 + GAD-7 mother | −.10 | −.22* | −.48 | −.01 | 4.4 |
Mental disorder mother | .06 | .10 | −.07 | .25 | 1.0 |
Quality of mother-EA relationship | −.10 | −.04 | −.22 | .18 | 4.2 |
Frequency of contact | .05 | −.02 | −.17 | .11 | 1.6 |
SES | .07 | .05 | −.11 | .20 | 1.4 |
EA gender | −.05 | −.11 | −.28 | .10 | 1.9 |
Total Problems– negative discrepancy, N = 127, R² =.143 | |||||
PHQ-9 + GAD-7 EA | −.09 | −.04 | −.27 | .23 | 6.0 |
PHQ-9 + GAD-7 mother | −.31 | −.32* | −.54 | −.12 | 61.0 |
Mental disorder mother | −.15 | −.05 | −.33 | .17 | 21.2 |
Quality of mother-EA relationship | .09 | .10 | −.11 | .32 | 5.7 |
Frequency of contact | −.06 | −.11 | −.25 | .05 | 2.5 |
SES | −.06 | −.11 | −.26 | .07 | 3.1 |
EA gender | .02 | .02 | −.16 | .20 | 0.5 |
Internalizing
N = 121 (49.0%) mother-child dyads showed a positive discrepancy on the Internalizing scale (ASR/self-report > ABCL/mother report). The model was significant (R² = .280), but only the EA’s level of current psychological distress (PHQ-9 + GAD-7 scores) significantly predicted a positive discrepancy (see Table 3), meaning that higher levels of psychological distress in the EAs were associated with larger positive discrepancies between self- and maternal reports. EA’s level of current psychological distress explained 88% of the predictable variance, indicating very high relevance of this predictor. EAs with higher levels of current psychological distress therefore also rated their symptoms of internalizing problems during the last six months as more severe than their mothers. N = 126 (51.0%) mother-EA dyads had a negative discrepancy (ASR/self-report < ABCL/maternal report). This model also yielded a significant result (R² = .210). The mother’s current psychological distress (PHQ-9 + GAD-7 scores), the frequency of contact, and family SES significantly predicted a negative discrepancy (see Table 3). Higher levels of maternal mental stress, lower relationship quality, a higher frequency of contact, and higher SES were thus associated with larger negative discrepancies between self- and maternal reports, meaning that mothers reported higher levels of internalizing problems in EAs than the EAs did. Inspection of the relative weights indicated that mothers’ psychological distress was the most relevant predictor, explaining 46% of the predictable variance, followed by frequency of contact (21%) and SES (11%).
Table 3
Zero-order correlations, standardized regression weights, and relative weights for ASR/ABCL INTERNALIZING
95% CI | |||||
|---|---|---|---|---|---|
r | β | Lower | Upper | Relative weight (%) | |
Internalizing Problems– positive discrepancy, N = 121, R² = .280 | |||||
PHQ-9 + GAD-7 EA | .51 | .53* | .31 | .73 | 87.5 |
PHQ-9 + GAD-7 mother | .05 | −.00 | −.03 | .02 | 2.5 |
Mental disorder mother | .05 | .04 | −.26 | .32 | 1.0 |
Quality of mother-EA relationship | −.11 | −.03 | −.18 | .15 | 5.6 |
Frequency of contact | −.06 | −.06 | −.19 | .04 | 1.7 |
SES | −.04 | −.05 | −.26 | .15 | 0.6 |
EA gender | −.02 | −.04 | −.26 | .19 | 1.1 |
Internalizing Problems– negative discrepancy, N = 126, R² = .210 | |||||
PHQ-9 + GAD-7 EA | −.10 | −.04 | −.25 | .18 | 5.8 |
PHQ-9 + GAD-7 mother | −.33 | −.37* | −.56 | −.10 | 45.4 |
Mental disorder mother | −.10 | .03 | −.20 | .24 | 9.0 |
Quality of mother-EA relationship | .13 | .16 | .00 | .36 | 8.3 |
Frequency of contact | −.21 | −.28* | −.43 | −.12 | 20.8 |
SES | −.15 | −.21* | −.34 | −.02 | 10.6 |
EA gender | −.01 | .00 | −.17 | .17 | 0.1 |
Externalizing
N = 112 (45.3%) mother-EA dyads showed a positive discrepancy on the Externalizing scale (ASR/self-report > ABCL/mother report). The model yielded a significant result (R² = .291). The EA’s current psychological distress (PHQ-9 + GAD-7 scores) and the quality of the mother-EA relationship significantly predicted a positive discrepancy (see Table 4), meaning that higher levels of psychological distress in EAs and lower relationship quality were associated with larger positive discrepancies between self- and maternal reports. EAs thus rated their externalizing problems as more severe than their mothers did among the dyads in which the EA reported higher levels of current psychological distress and a lower-quality mother-EA relationship. The relative weights showed that EAs’ psychological distress explained 55% of the predictable variance, while the relationship quality explained 30%. The standardized regression coefficient of mothers’ current psychological distress (PHQ-9 + GAD-7 scores) was also significant, however, the relative weight (5%) indicated low relevance of this predictor. N = 135 (54.7%) mother-child dyads had a negative discrepancy (ASR/self-report < ABCL/mother report). This model was also significant (R² = .152). The quality of the mother-EA relationship and frequency of contact significantly predicted a negative discrepancy (see Table 4), meaning that a lower relationship quality and higher frequency of contact were associated with larger negative discrepancies between self- and maternal reports, i.e., mothers reported higher levels of externalizing problems in their children than the EAs did. Inspection of relative weights showed that relationship quality explained 44% of the predictable variance in mother-child discrepancy, while frequency of contact explained 11%. Mothers’ psychological distress yielded a comparatively high relative weight (25%), but the standardized regression coefficient did not reach statistical significance.
Table 4
Zero-order correlations, standardized regression weights, and relative weights for ASR/ABCL externalizing
95% CI | |||||
|---|---|---|---|---|---|
r | β | Lower | Upper | Relative weight (%) | |
Externalizing Problems– positive discrepancy, N = 112, R² = .291 | |||||
PHQ-9 + GAD-7 EA | .41 | .41* | .17 | .63 | 55.3 |
PHQ-9 + GAD-7 mother | −.11 | −.20* | −.38 | −.02 | 5.2 |
Mental disorder mother | .05 | .08 | −.17 | .28 | 1.1 |
Quality of mother-EA relationship | −.29 | −.28* | −.45 | −.09 | 29.5 |
Frequency of contact | .08 | .08 | −.11 | .22 | 2.9 |
SES | .07 | .07 | −.09 | .21 | 1.7 |
EA gender | −.11 | −.12 | −.30 | .06 | 4.3 |
Externalizing Problems– negative discrepancy, N = 135, R² = .152 | |||||
PHQ-9 + GAD-7 EA | −.10 | −.01 | −.29 | .28 | 7.6 |
PHQ-9 + GAD-7 mother | −.20 | −.21 | −.46 | .03 | 24.7 |
Mental disorder mother | −.11 | −.05 | −.31 | .18 | 9.7 |
Quality of mother-EA relationship | .27 | .30* | .07 | .57 | 43.7 |
Frequency of contact | −.13 | −.18* | −.33 | −.03 | 11.3 |
SES | −.01 | −.06 | −.22 | .10 | 0.7 |
EA gender | −.06 | −.05 | −.24 | .14 | 2.3 |
Discussion
The present study examined agreement and discrepancies of self- and maternal-reported symptoms of psychopathology in emerging adults. Cross-informant agreement on adult psychopathology has not yet been researched thoroughly, and to our knowledge, this is only the second study to investigate agreement between emerging adult children and their parents. Agreement between self- and maternal reports was in the moderate to high range (r =.48 to r =.53) and did not differ significantly between Internalizing and Externalizing. These results are very similar to those of Rescorla et al. (2016), who reported cross-informant correlations for the ASR/ABCL questionnaires of r =.50 for Total Problems, r =.53 for Internalizing, and r =.50 for Externalizing. Rescorla et al. (2016) did not differentiate between the type of collateral report because they found only minor differences between the ratings of spouses/partners and other collateral informants, such as family members and friends. This might indicate that parent reports are of similar quality as other collateral informant ratings, e.g., partners, in emerging adulthood. For EAs who do not have a close relationship or do not wish to include their partner in the diagnostic process, their parents might be a suitable alternative and vice versa.
Our results are also comparable to those regarding the agreement between self- and parent reports among adolescents: Rescorla et al. (2013) reported cross-informant correlations for the YSR/CBCL questionnaires of r =.45 for Total Problems, r =.45 for Internalizing, and r =.46 for Externalizing. Researchers agree that collateral ratings are of incremental value in the diagnosis of mental health problems in childhood and adolescence (De Los Reyes et al., 2022). The current results imply that collateral informant ratings in emerging adulthood have similar information content and should therefore remain part of the diagnostic routine. For a more comprehensive understanding of collateral informant reports in emerging adulthood, however, longitudinal studies that investigate the development of parent‒child agreement from adolescence to emerging adulthood are needed. This would provide valuable information about whether the direction and the magnitude of discrepancies change or remain stable on the individual level.
Both the EAs’ self-reports and their mothers’ reports correlated significantly with the clinicians’ ratings of mental disorders in the EAs, and there was no difference between Total Problems and Internalizing. Mothers’ reports therefore seem to be an important and reliable source of information about mental health problems in their emerging adult children. However, agreement with the clinicians’ ratings was significantly higher for the EAs themselves. As the clinical interview was conducted with the EA as the only informant, it is not surprising that their self-report was more strongly associated with the result determined from the interview. Unfortunately, associations with the clinicians’ ratings of externalizing problems could not be analyzed due to the low number of individuals in the current sample who met the criteria for an externalizing disorder. Although we did not find any differences in mother-child agreement for Externalizing, Internalizing, and Total Problems, some previous studies reported higher agreement regarding externalizing problems in childhood and adolescence compared to internalizing problems (e.g., De Los Reyes et al., 2015). Agreement with the clinical rating of externalizing problems in emerging adulthood might differ from agreement about total and internalizing problems and should therefore be investigated in future research with larger samples or samples with a focus on externalizing behavior problems.
The moderate to high associations between EAs’ self- and mothers’ collateral reports, and of both reports with the clinical rating, emphasizes the importance of using multiple informants’ reports to make valid differential diagnoses: Self- and maternal collateral reports share some degree of information, while both also contain unique aspects of EAs’ psychopathology.
Relative weight analyses revealed important and consistent patterns of factors associated with mother-child agreement in emerging adulthood. Among the dyads in which the EA child reported higher levels of psychopathology than their mother (positive discrepancy), the EA’s current psychological distress was significantly associated with mother-EA discrepancies in terms of Total Problems, Internalizing, and Externalizing, meaning that higher levels of self-reported psychological distress were linked to larger discrepancies. EA’s current psychological distress was by far the most relevant predictor of a positive discrepancy, explaining between 55% and 88% of the predictable variance. Acute distress (during the past two weeks) might cause EAs to also rate their level of mental health problems during the last six months as higher, which would result in a larger difference compared to their mothers’ ratings. The higher rating might be due to actual higher levels of psychopathology, but it is also possible that the current distress level distorts the self-report, similar to cognitive biases described as part of depressive symptoms (Beck, 1972). However, the results are contrary to Szkody et al.’s (2022) consideration that higher levels of self-reported mental health problems in EAs, including both internalizing and externalizing problems, might increase the likelihood that their parents notice their symptoms and therefore reduce discrepancies.
For positive discrepancies in the report of externalizing problems, the mother-EA relationship was also significant, meaning that a worse relationship quality was linked to larger positive discrepancies. Perhaps a lower relationship quality leads to less accurate reports of psychopathology (Szkody et al., 2022), which could explain why mothers report lower levels of externalizing problems. Disagreement between mothers and EAs about their behavior, resulting in larger discrepancies, might also lead to and indicate a worse relationship quality (Szkody et al., 2022).
Maternal psychological distress was also significant regarding positive discrepancies in the report of total and externalizing problems. The relative weights, however, were very small (4–5%), indicating very low relevance of this predictor. Our results therefore suggest that maternal psychological distress is a negligible factor when examining higher EA than mother reports.
When mothers reported higher levels of psychopathology in the EAs than the EAs did (negative discrepancy), factors associated with discrepancies were slightly more diverse. Maternal current psychological distress was significantly associated with negative discrepancies for Total Problems and Internalizing, but not Externalizing. Higher levels of maternal psychological distress were linked to larger negative discrepancies. These results indicate that not the presence or absence of a mental disorder in general but rather current psychological distress in mothers influences their perceptions of their child’s behavior and mental health. Maternal psychopathology, usually assessed through self-report questionnaires, has been consistently linked to mother-child discrepancies (e.g., Affrunti & Woodruff-Borden, 2015; Fjermestad et al., 2017; Radicke et al., 2021). Our findings indicate that psychopathology itself is not the relevant construct but that maternal mental health problems related to mother-EA disagreement must be investigated in a more differential way. Higher levels of distress might make mothers more sensitive to or cause them to overinterpret mental health issues in their EA children and therefore perceive them as more serious. Similar processes have been postulated in the Depression-Distortion Hypothesis: Perceptual biases in mothers with depression lead to more severe ratings of their children’s mental health problems (Richters, 1992). This has important implications for clinical research and practice, as it warrants a screening of collateral informants’ mental distress (rather than their mental health status) as a possibly confounding covariate. It is surprising, however, that maternal psychological distress was not relevant for externalizing symptoms, as Szkody et al. (2022) found maternal depressive symptoms to be linked to larger discrepancies on reports of antisocial problems in EAs. This difference might be due to the operationalization of psychological distress in the study at hand, which included both depressive and anxiety symptoms. The relative weight for externalizing problems, however, still reached 25% in the current study, indicating some relevance despite non-significance. The association between maternal psychological distress and agreement on externalizing problems therefore needs further investigation and should include differentiated analyses of symptoms of depression and anxiety in mothers. The quality of the mother-child relationship was significant for Externalizing only, with a worse relationship quality and higher frequency of contact between mothers and EAs being associated with larger negative discrepancies (self-reports < maternal reports). According to Szkody et al. (2022), lower affection in a relationship might be interpreted as indicating a higher level of mental health problems, which could explain the link between lower relationship quality and larger negative discrepancies. Of course, it might also be possible that the negative relationship quality is a consequence of divergent assessments by EAs and their mothers. Lohaus et al. (2020) found evidence for bidirectional influences between informant discrepancies and the quality of the mother-child relationship. This would mean that a low-quality mother-child relationship leads mothers to report higher levels of psychopathology in their EA children and that the different understanding of symptoms in EAs also leads to a worse relationship quality between mothers and their EA children.
Regarding contact frequency, we found that more contact between mothers and their EA children was associated with larger negative discrepancies (self-reports < maternal reports) in reports of internalizing and externalizing problems; thus, the EAs reported fewer problems than their mothers. In contrast, de Los Reyes and Kazdin (2005) suggested that mothers spending more time with their young children leads to more observations of relevant behavior and higher cross-informant agreement. In emerging adulthood, more opportunities to observe and speak to the child seem to be associated with higher levels of mental health problems in mothers’ reports compared to EA self-reports. Szkody et al. (2022) suggested that parents who perceive higher levels of psychopathology in their EA children might be more involved in their children’s lives, which could also lead to more conflict in the parent‒child relationship. This could explain our findings that negative discrepancies were associated with both a worse mother-EA relationship and a higher frequency of contact at the same time.
For Internalizing, SES was also significant, with a higher SES related to larger negative discrepancies (self-reports < maternal reports). This is inconclusive, as it has previously been suggested that in youth, a lower SES is associated with larger discrepancies (e.g., Stone et al., 2013) or shows no association at all (De Los Reyes & Kazdin, 2004). Because the present sample included very few families with a low SES, whereas most had a high SES, resulting in very little variance in the data, our results might not be generalizable. The role of SES in parent‒child discrepancies in emerging adulthood therefore needs further investigation. Similar to results in childhood and adolescence (Achenbach et al., 1987), child gender was not associated with any discrepancies between maternal and self-reports in emerging adulthood.
As in most previous studies, the present study used self- and collateral report questionnaires to assess informant discrepancies. It should however be noted, that participants’ reports might be influenced or even biased by variables that we did not examine, such as social desirability. As questionnaires were filled out in private, however, we consider this potential bias to be negligible. The ASR/ABCL questionnaires assess symptoms during the previous six months (Achenbach & Rescorla, 2003). It is possible therefore that we also captured some degree of symptom chronification. Another important factor that has been shown to be associated with mental health problems is use of (social) media (e.g., Davila et al., 2012). As mentioned above, emerging adulthood is a developmental period that is characterized by diverse challenges, which might also influence EAs’ and mothers’ reports. EAs’ self-perception as well as their observable behavior might vary between different contexts, which might influence self- and collateral reports of psychopathology. A better understanding of these developmental challenges as well as measures to capture these constructs are needed to determine associations with informant discrepancies. The current study only investigated a few relevant factors, however, further research is needed to gain a comprehensive understanding of variables associated with multiple informants’ reports. It is important to emphasize that the possible influence of different known and unknown factors on reports of psychopathology supports the demand to always collect multiple reports for a thorough assessment of symptoms.
In summary, we found moderate to high agreement between self- and mother-reported psychopathology in emerging adulthood. Both ratings were significantly associated with the clinicians’ ratings of mental disorders in the EAs. When EAs reported higher levels of psychopathology than their mothers, the EAs’ current psychological distress level seemed to be the most important factor associated with the discrepancy. When mothers reported higher levels of psychopathology than the EAs themselves, the mother’s current psychological distress level, quality of the mother-EA relationship, and frequency of contact seemed to be relevant. These results are partly in line with previous research, but we also add some important new findings to the literature: (1) Mother-EA agreement about psychopathology in emerging adulthood is comparable to mother-child agreement in childhood and adolescence as well as to the agreement of self- and collateral informant reports in adulthood. The large agreement shows that mothers’ ratings contain significant information about the EAs’ mental health. (2) Distinct factors are related to positive vs. negative discrepancies between maternal and self-reports of psychopathology. (3) Maternal current psychological distress levels seem to be more relevant for mother-child discrepancies than meeting the criteria for a mental disorder. As the administration of screenings for psychological distress is much easier and more reliable than the assessment of mental disorders, this is an advantage for future research.
Strengths and Limitations
This is only the second study to investigate mother-child agreement about EAs’ symptoms of psychopathology. With 256 mother-EA dyads recruited from a longitudinal prevention study, we recruited an important sample for our analyses. The ASR/ABCL questionnaires are frequently used to assess psychopathology and cross-informant agreement in adults. De Los Reyes et al. (2022) pointed out, using established measures in cross-informant research is of vital importance. To our knowledge, this is also the first study to include a half-structured clinical interview in the investigation of cross-informant agreement. This enabled us to use valid diagnoses of mental disorders made by trained clinical psychologists to compare to maternal and self-reports and to investigate the association of maternal mental disorders with mother-EA agreement. In most previous studies, maternal psychopathology has been assessed through self-report questionnaires only (e.g., Affrunti & Woodruff-Borden, 2015; Fjermestad et al., 2017; Radicke et al., 2021).
Our study, however, also has some limitations. We only used data from mothers. It is possible that father-EA agreement about psychopathology and its associations differs from mother-EA agreement. Second, most families in our sample had a high SES. As psychopathology is associated with SES (Dohrenwend, 1990; Reiss, 2013), it is possible that associations with cross-informant discrepancies could not be detected in the current sample. Most of the data were collected during the COVID-19 pandemic, which might have increased participants’ psychological distress and influenced the quality of the parent‒child relationship as well as the frequency of contact and thus our results. In particular, EAs’ mental health has been shown to be particularly impacted by the pandemic (Généreux et al., 2021; Kwong et al., 2021). To date, no German standardized values (T scores) exist for the ASR/ABCL questionnaires. As ASR/ABCL scores differ significantly among cultures (Rescorla et al., 2016), we refrained from using the American standardization and used z standardization instead. Therefore, it was not possible to investigate differences regarding the clinical significance of emerging adults’ symptoms of psychopathology. A single item was used to assess relationship quality in the current study due to the study design. Multiple established measures exist to assess relationship quality (i.e., Network of Relationships Inventory by Furman & Buhrmester, 2009, as used by Szkody et al., 2022), and future investigations are needed to compare our results to those using more comprehensive measures of relationship quality. Last, we used data from a cross-sectional study and cannot make assumptions about causal relations. It is therefore not possible to say whether mother-child discrepancies are simply related to certain factors, if there are causal relationships, or if there are bidirectional influences.
Implications for Research and Clinical Practice
Against the background of the previous research literature, our results support the assumption that mothers’ collateral reports contain important information about psychopathology in emerging adulthood. The routine practice of collecting cross-informant reports in research and clinical practice should therefore not end as individuals reach their 18th birthday but should be extended to adulthood. The distinct factors associated with informant discrepancies warrant screening (a) for mental health problems not only in the targeted person but also in the collateral informant(s) providing reports and (b) of the quality of the interpersonal relationship.
Our results also support the assumption that maternal and self-reports are not interchangeable (De Los Reyes et al., 2020). A change in informant between assessments would therefore be problematic, which is especially relevant for longitudinal studies. Researchers conducting longitudinal studies often face the question of whether the parent reports on their child’s psychopathology should be exchanged for child self-reports once the child is old enough to provide information themselves. If an EA’s self-report is included, it should not replace the parent report. When using a parent report only, however, researchers must be aware that it will most likely differ from the child’s point of view but nevertheless contain important information about the child’s mental health problems.
To date, only a few studies have longitudinally investigated cross-informant agreement in childhood and adolescence (e.g., Mastrotheodoros, Van der Graaff, Dekovic, Meeus & Branje, 2019; Yang et al., 2021), with inconsistent results. The longitudinal development of parent‒child agreement beyond adolescence has yet to be investigated. The examination of the stability of cross-informant agreement and discrepancies and associated factors would improve our understanding of collateral informant reports not only in emerging adulthood but also in childhood and adolescence. The understanding of collateral informant reports impacts decisions in clinical research and care (De Los Reyes et al., 2022) and is therefore of great importance for both professionals and participants in mental health assessment.
Acknowledgements
We thank all families for their great commitment to the study and Prof. Dr. Kurt Hahlweg for proofreading this article!
Declarations
Ethical Approval
Informed consent was obtained from all individual participants included in the Future Family IV project. The project was conducted according to the principles stated in the Declaration of Helsinki (64th WMA General Assembly, Fortaleza, Brazil, 2013). The project received ethical approval by the independent ethics committee of the University of Braunschweig (identification number D-2020-09; Faculty of Life Sciences).
Competing Interests
On behalf of all authors, the corresponding author states that there are no conflicts of interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
