Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Oct 1.
Published in final edited form as: J Psychopathol Clin Sci. 2022 Jun 16;131(7):741–753. doi: 10.1037/abn0000769

A Multi-Method, Multi-Informant Study of Early Childhood Temperament and Depression and Anxiety Symptoms in Adolescence

Thomas M Olino 1, Megan C Finsaas 2, Margaret W Dyson 3, Gabrielle A Carlson 4, Daniel N Klein 5
PMCID: PMC9560959  NIHMSID: NIHMS1822434  PMID: 35708930

Abstract

There is an extensive literature documenting associations between dimensions of temperament and depressive and anxiety disorders. However, much of the research in this area relies on cross-sectional and monomethod designs, uses samples of older youth and adults, and examines only the main effects of temperament. We examined longitudinal associations between temperamental positive emotionality (PE) and negative emotionality (NE) in early childhood and depressive and anxiety symptoms in adolescence. We assessed temperament using laboratory observations and parent reports, and symptoms using youth and parent report. We also examined potential moderators of these associations, including maternal and paternal history of depressive and anxiety disorders, youth experience of stressful life events, and parenting styles. Early childhood NE was associated with both adolescent depression and anxiety. Additionally, life events and psychological controlling and firm parenting moderated the associations between NE and depression symptoms, and maternal depression, paternal anxiety, and psychological controlling and firm parenting moderated the associations between NE and anxiety symptoms. Interaction effects were largely consistent with diathesis-stress interpretations. These findings show that temperamental NE, but not PE, is prospectively associated with risk for depressive and anxiety symptoms in adolescence, and that multiple additional factors potentiate these associations.

Keywords: temperament, negative emotionality, positive emotionality, depression, anxiety

General Scientific Summary:

Dimensions of temperament are frequently associated with depression and anxiety. However, those studies infrequently examine those associations across long periods of time. Here, we find that temperamental negative emotionality, but not positive emotionality, is associated with later depression and anxiety, and that the magnitude of these associations are influenced by family history, experience of stress, and experience of parenting.

A prominent line of research on the development of depression and anxiety focuses on temperament and personality as markers of vulnerability (Klein et al., 2011; Tackett, 2006).1 Although there is a substantial literature on this topic, this research is frequently limited by samples that are well into the risk period for onset of emotional disorders and often relies on the same method and data source to assess both temperament and psychopathology (Klein et al., 2011). Moreover, to understand the relationship of temperament with emotional disorders, it may be necessary to go beyond simple main effect models and incorporate moderators of these associations, such as family history of depression and anxiety, stressful life events, and early caregiving experiences (see Klein et al., 2012 for a review of potential moderators).

In one of the most influential models of temperament and psychopathology, Clark and Watson (1991) proposed that neuroticism/negative emotionality (NE) is a general vulnerability for depression and anxiety, and low extraversion/positive emotionality (PE) is a specific vulnerability marker for depression.2 In a meta-analysis of the voluminous cross-sectional literature on traits and psychopathology, Kotov et al. (2010) reported that NE was higher in individuals with depressive and anxiety disorders than those without these disorders, and PE was lower in individuals with depressive disorder than non-depressed participants. Surprisingly, however, individuals with anxiety disorders also exhibited lower levels of PE than those without anxiety disorders.

There have also been a number of longitudinal studies examining prospective relationships of traits with depression (Hakulinen et al., 2015; Jeronimus et al., 2016; Klein et al., 2011), and a smaller number examining anxiety (Goldstein et al., 2018; Kendall et al., 2015; Zinbarg et al., 2016). These studies have consistently supported NE predicting depression and anxiety, but mixed findings for PE predicting depression. Most prospective studies have used adolescent and adult samples; however, interpreting their findings is complicated because depression and anxiety symptoms are already common by adolescence. Given that temperament is believed to emerge very early in development, prospective longitudinal studies beginning in early childhood and extending at least into adolescence are critical. Unfortunately, most prospective studies of children have been limited to short follow-up periods, often concluding before adolescence (Klein et al., 2012), when risk for onset of depressive and many anxiety disorders dramatically increases. Additionally, these studies have generally relied on questionnaire measures of both sets of constructs, typically administered to the same informant (parents, in studies of children; self-report in studies of adolescents and adults). This may inflate associations through shared method and informant variance.

Only a small number of studies have used stronger prospective designs to elucidate temperamental risk for depression and anxiety, assessing young children and following them into adolescence or adulthood, using multiple methods/informants. Caspi et al. (1996) examined associations between behavioral observations during assessments of cognitive and motor function at age 3 and diagnoses of mental disorders at age 21. An inhibited behavior profile characterized by low approach, fearfulness, and distress, was associated with greater likelihood of depressive, but not anxiety, disorders. Although this study assessed children prior to the risk period for most depressive and anxiety disorders, followed the sample into early adulthood, and used observational measures of temperament, their construct of inhibited3 temperament combined aspects of both NE and low PE. Thus, their results do not map clearly onto current models of temperament and psychopathology. Bould et al. (2014) examined parents’ ratings of their 6-year old children on a standard temperament questionnaire, and found that emotionality (similar to NE) predicted current youth self-reported depression diagnoses by age 18. However, they did not examine anxiety disorders.

The effect sizes for the association between child temperament and psychopathology in adolescence and early adulthood are quite small (Hakulinen et al., 2015; Jeronimus et al., 2016). This is likely due to the other, non-temperamental, pathways to the development of internalizing disorders (i.e., equifinality). In addition, the lengthy interval between early childhood and adolescence provides opportunities for other factors to intervene to augment or reduce risk, moderating the association between temperament and psychopathology (Klein et al., 2012).

Three of the most plausible moderators are parental psychopathology, life stress, and parenting styles and practices (see Klein et al., 2012). Parental depressive and anxiety disorders are among the best-established predictors of depression and anxiety in offspring, with evidence for both homotypic and cross-transmission (Goodman, 2020; Lawrence et al., 2019). However, many offspring of parents with depression and anxiety do not develop emotional disorders (Klein et al., 2005). High NE and/or low PE may mark a subgroup of offspring at particularly high risk (Olino et al., 2010). Alternatively, children with high NE or low PE may pose particular challenges for parents struggling with their own depression and anxiety, setting in motion transactional processes that increase child symptoms (e.g., Kopala-Sibley, Jelinek, et al., 2017).

Surprisingly few studies have examined parental psychopathology as a moderator of the relationship between child temperament and later psychopathology. Jessee, Mangelsdorf, Shigeto, and Wong (2012) reported that when children (mean age 4.90 years) were high on NE or low on PE, maternal depression was associated with more total child behavior problems. In contrast, when children were low on NE or high on PE, maternal depression was not associated with child symptoms.

Studies, mainly with adolescents and adults and focusing on NE and depression, have shown that life stress moderates the relationship between traits and psychopathology. Most of these effects have been consistent with a diathesis-stress model, where the joint presence of trait vulnerability and high stress predicts the greatest symptoms (Brown & Rosellini, 2011; Kendler et al., 2004; Ormel et al., 2001; van Os & Jones, 1999). However, in some studies interactions have taken other forms, such as dual vulnerability, where trait vulnerability and stress are each associated with greater symptoms, but their joint effect is no greater than either individual effect (i.e., no additive effect; Gulley et al., 2016; Kopala-Sibley, Klein, et al., 2017; Leve et al., 2005). Finally, some studies failed to find trait by stress interactions (e.g., Mineka et al., 2020).

We previously examined this question in a more limited form in the current sample. We found that mothers’ reports of the stress associated with a natural disaster at a mean age of 10 years moderated laboratory observations of NE at age 3 to predict mother’s reports of children’s anxiety and depression symptoms immediately following the disaster. Elevated levels of NE predicted increased symptoms at higher levels of stress (Kopala-Sibley et al., 2016).

Finally, in keeping with the view that the “fit” between children’s temperament and their environments play a critical role in development (Chess & Thomas, 2013), parenting styles and practices have been examined as moderators of the link between child temperament and psychopathology. Chen, Deater-Deckard, and Bell (2014) reported that child NE predicted greater total symptoms when mothers exhibited higher levels of negative parenting, but was unrelated to symptoms when mothers showed lower levels of negative parenting. Similarly, Oldehinkel et al (2006) found that some facets of NE, such as fearfulness and frustration, interacted with parental overprotection, low warmth, and rejection in predicting depressive symptoms, such that associations were stronger in children with higher levels of these facets.

Parenting may also be protective for children with problematic temperament styles. Gallitto (2015) found that positive parenting mitigated risk for internalizing symptoms among children with low temperamental adaptability. However, null and counterintuitive findings have also been reported. For example, Davis, Votruba-Drzal, and Silk (2015) found that parental warmth predicted higher levels of internalizing symptoms in children with higher NE.

The present study examines associations between temperament assessed at age 3 and depression and anxiety symptoms through mid-adolescence. We used multimethod assessments of temperament and psychopathology constructs to reduce inflation of associations based on shared method variance. Our work also goes beyond studies of main effect associations by examining theoretically motivated moderators of the associations between temperament and internalizing problems. We hypothesized that we would find significant positive associations between temperamental NE and both depression and anxiety. We also hypothesized that we would find a negative association between temperamental PE and depression. However, we did not have a hypothesis regarding PE and anxiety, as empirical findings have not always been consistent with theory. Finally, we also examined whether parental history of depressive and anxiety disorders, stressful life events, and parenting styles moderated the associations between early NE and PE and subsequent depression and anxiety symptoms.

Methods

Participants

Data come from the Stony Brook Temperament Study (Klein & Finsaas, 2017; Olino et al., 2010). Participants were recruited through commercial mailing lists, and children with any significant medical conditions or developmental disabilities were excluded. The sample included 559 children at baseline assessment (54% Male; Mage=42.2 months [SD=3.10]) and their biological parents. On average, mothers and fathers were 36.0 (SD=4.5) and 38.3 (SD=5.4) years of age, respectively. Most participants (86.9%) were White and middle class, as measured by Hollingshead’s Four Factor Index of Social Status (M=45.1 [SD=10.9]). Approximately half the mothers (54.7%) and fathers (45.7%) had at least a 4-year college degree. Most children (95.0%) lived with both biological parents. Participants were assessed again at age 12 (Mage=12.66 years, SD=0.46) and age 15 (Mage=15.25 years, SD=0.41). Of the 559 original participants, 437 and 458 children completed diagnostic interviews at the age 12 and age 15 follow-ups, respectively. We compared youth with follow-up data at either age 12 or age 15 to those without follow-up data on youth temperament at age 3, child sex, race/ethnicity, and parental education, as well as the moderators that we examined. Across those comparisons, only one significant difference was found. Father report of negative affect on the CBQ was significantly lower for the youth with than without follow-up data (completers M=3.77, SD=0.54; non-completers M=3.92, SD=0.46; Welch t(78.94)= 2.10, p=.038). Thus, there is modest systematic attrition in the sample.

Temperament Measures

Laboratory Temperament Assessment Battery (Lab-TAB).

At age 3, children were assessed in an observational laboratory setting with sessions lasting approximately 2 hours. The assessment included a set of 11 episodes (risk room, tower of patience, arc of toys, stranger, car go, transparent box, pop-up snakes, impossibly perfect green circles, popping bubbles, and box empty) adopted from the Lab-TAB (Gagne, van Hulle, Aksan, Essex, & Goldsmith, 2011), and one additional episode (exploring new objects) that was designed specifically for this study (Dyson et al., 2012). Children were given a short play break between episodes, providing them time to return to a neutral affective state. Details on the specific laboratory episodes are described in Dyson et al. (2012) and in Supplementary Table 1.

Temperament is conceptualized as general patterns of behavior that are reasonably consistent across contexts. Our observational coding system implemented this by making global ratings of the same dimensions of child behavior across all episodes. Undergraduate research assistants, study staff, and graduate students served as coders following extensive training. Each coder was assigned to specific episodes. Coders had to reach at least 80% agreement on all specific codes within the episode with a “master” rater before coding independently.

Each time-stamp recorded instance of facial, bodily, and vocal affect during all 12 laboratory episodes was rated on a three-point intensity scale (low, moderate, high) for positive affect, anger, sadness, and fear. Intensity ratings within each laboratory episode were summed within each channel of affect (facial, bodily, vocal). These ratings were then averaged across channels, resulting in scores for each of the 12 episodes for each of the four affective traits. These variables were then standardized, and the standardized scores were averaged across episodes for each affect trait. Interest/engagement was rated on a single 4-point scale (none, low, moderate, and high) for each episode, and these ratings were summed across the 12 episodes. PE consisted of the sum of the standardized total positive affect and interest/engagement variables. NE was the sum of the standardized total sadness, fear, and anger variables.

Internal consistency (N = 559) for PE and NE were α = .82 and .74, respectively. Videotapes of 35 children were re-coded by an independent coder to examine interrater reliability; intraclass correlations (ICC) for PE and NE were .89 and .82, respectively.

Tester impressions.

The research assistant conducting the laboratory visit completed a set of global ratings about the child adapted from the post-Lab-TAB rating scale (Gagne et al., 2011). The experimenter was with the child from the moment the family arrived on campus until the moment they departed, including the unstructured play breaks between each of the Lab-TAB episodes. Thus, experimenter ratings are both more global and based on a larger sample of child behavior. Experimenters rated the participant on 24 scales, 10 of which were used here to derive measures of NE and PE. Each variable was rated on a 5-point Likert scale (1 = rarely, 2 = subtle or ambiguous signs, 3 = mild, 4 = moderate, 5 = extreme). NE was the average of overall negative affect, fearfulness, frustration with tasks, anger or irritability, and sadness (α = .79). PE was the average of overall positive affect, interest in test materials and stimuli, enthusiasm toward tasks, initiative, and anticipatory positive affect (α = .89). Interrater reliabilities are not available as there was only one experimenter per Lab-TAB assessment. However, Gagne et al. (2011) has reported excellent interrater reliability for the items in this measure.

Child Behavior Questionnaire (CBQ).

Parent-reported child temperament was assessed at age 3 using the CBQ, which assesses 15 primary scales in 3-7 year olds. It consists of 191 items rated on a 7-point Likert scale ranging from 1 (extremely untrue of your child) to 7 (extremely true of your child). The present study used the broadband negative affectivity scale (consisting of 5 primary scales) and a composite of the approach anticipation and smiling/laughter scales to index NE and PE, respectively. The mean of the internal consistencies of mothers’ and fathers’ ratings on the primary scales used in this study were: anger/frustration (α = .78), discomfort (α = 68), soothability (α = .71), sadness (α = .65), fear (α = .70), approach anticipation (α = .71), and smiling/laughter (α = .75).

Moderators

Parent Psychopathology.

At age 3, children’s biological parents were interviewed using the Structured Clinical Interview for DSM-IV, Non-patient version (SCID-NP; First et al., 1996). Interviews were conducted by telephone, which yields similar results as face-to-face interviews (Rohde et al., 1997), by two Masters-level raters with no knowledge of the temperament ratings. SCIDs were obtained from 535 (99.8%) mothers and 443 (82.6%) fathers. When parents were unavailable, family history interviews were conducted with the co-parent. Diagnoses based on family history data were obtained for an additional one (0.2%) mother and 83 (15.5%) fathers. Based on audiotapes of 30 SCID interviews, kappas for inter-rater reliability of lifetime diagnoses were .93 for depressive disorder and .91 for anxiety disorder.

Life Stress.

At age 9, each child and a parent were administered an interview for stressful life events based on Costello, Angold, March, and Fairbanks (1998). The interview covered the occurrence of 16 stressful life events commonly assessed in studies of depression and anxiety (e.g., death of a loved one, parental relationship dissolution, loss of a best friend, attending school in an unsafe area, noticeable reduction in the family’s standard of living) during the period since the age 6 assessment. We view this measure as an indicator of the level of stress experienced during childhood, rather than a proximal precipitant of internalizing symptoms in adolescence. For the present analyses, we summed the number of events experienced by the child. Inter-rater reliability was good (N = 72; ICC = .79).

Parenting.

When children were age 9, mothers and fathers completed the parent-report version of the Children's Report of Parent Behavior Inventory (CRPBI; Margolies & Weintraub, 1977; Schwarz et al., 1985). The CRPBI includes three 10-item scales: Psychological Control vs. Psychological Autonomy, Acceptance vs. Rejection, and Firm vs. Lax Control. The mean αs for mothers’ and fathers’ ratings on the three scales were .74, .74, and .68, respectively. Mother and father ratings were averaged to yield single indices. Higher scores indicate higher levels of Control, Acceptance, and Firm Discipline.

Outcome: Youth Psychopathology.

At the age 12 and 15 waves, one parent (85.6% mothers at age 12 and 91.2% mothers at age 15) and the youth were interviewed using the Kiddie Schedule for the Affective Disorders and Schizophrenia (K-SADS-PL; Kauffman et al., 1997). Doctoral students in clinical psychology and a master’s-level clinician, supervised by a child psychiatrist and clinical psychologist, administered the K-SADS to the parent and then to the youth. Parent and youth reports were combined into summary ratings for each symptom based on interviewer judgement. Rates of diagnoses of depressive disorders were 1.6% and 6.3%, and rates of anxiety disorders were 19.5% and 17.2% for the age 12 and age 15 assessments. Dimensional scores summed relevant symptom ratings in the interval before the interview at each wave. Dimensional scores followed skip-out rules consistent with typical KSADS administration. ICCs for interrater reliability (N = 25) were .97 for depression, and .93 for anxiety.

At the age 12 and 15 waves, children and both parents completed the 41-item youth self-report and parent-report versions, respectively, of the Screen for Childhood Anxiety Related Disorders (SCARED; Birmaher et al., 1997, 1999). Respondents rated the presence of anxiety symptoms in the child over the past 3 months on a three-point scale (0 = Not true or hardly ever true; 1 = Somewhat true or sometimes true; 2 = Very true or often true). Across time and informants, αs for SCARED total scores ranged from .89-.93.

Similarly, at the age 12 and 15 waves, children and both parents completed the youth self-report and parent report versions, respectively, of the Children’s Depression Inventory (CDI; Kovacs, 1992). The youth version consists of 28 items; the parent version includes 17 items. Across time and informants, αs ranged from .79-.83.

We averaged K-SADS, SCARED, and CDI scores within informants across the age 12 and age 15 waves to use as indicators in our models. In addition to limiting the number of analyses, this reduces the likelihood that assessments are capturing transient subthreshold symptoms, which are common in community samples. Aggregating across assessments may better capture consistent and meaningful variance in depression and anxiety.

Data analysis

All modeling was completed using Mplus 8.6 (Muthén & Muthén, 1998) and facilitated using the MplusAutomation (Hallquist & Wiley, 2018) package in R (R Core Team, 2018). Initial models tested the measurement models for age 3 temperament and adolescent internalizing psychopathology separately. We then estimated a model examining the structural associations between temperament and psychopathology. In building the measurement models, we were aware that the literature shows modest cross-method and cross-informant associations for the same trait. We also expected that there would be stronger associations across traits, but within informants. Thus, we were open to model modifications that correlated residuals within methods and informants. After identifying a measurement model that had at least an adequate fit to the data, we included additional, observed variables as moderators of the associations between temperament and depression and anxiety.

All models were estimated using robust maximum likelihood estimation (MLR). The interaction effects were estimated between latent and observed variables using the XWITH function in Mplus. As this required the use of random effects, absolute fit information was not available for these models. Post-hoc simple slopes were estimated at the 20th, 50th, and 80th percentile values of the moderator. We examined regions of significance between the 20th and 80th percentiles of the independent variable (in shaded regions of the plots). We evaluated models on 3 goodness of fit indices: the comparative fit index (CFI; Bentler, 1990), Root Mean Square Error of Approximation (RMSEA; Steiger, 1990), and standardized root mean square residual. Although cut-offs are somewhat arbitrary (Marsh et al., 2004), current conventions suggest that excellent model fit is indicated by CFI values ≥.95 (Hu & Bentler, 1999) and RMSEA values ≤.05 (MacCallum et al., 2006); adequate fit is indicated by CFI > .90, RMSEA between .05 and .08, and SRMR < .08.

This study was not preregistered. Data are available from the corresponding author upon request.

Results

Supplementary Table 2 shows intercorrelations among study variables, including those included within the measurement models and moderator variables. As expected, among indices of temperament, like-constructs from similar assessment methods showed stronger associations than like-constructs from different methods.

Measurement model: Age 3 temperament

The age 3 temperament model was specified with the PE factor reflected by Lab-TAB coded PE, tester impression PE, and maternal and paternal reports on the CBQ smiling/laughter and approach anticipation scales. The NE factor was reflected by Lab-TAB coded NE, tester impression NE, and maternal and paternal reported CBQ NA. Fit of the model was poor (Table 1). A second model that included a priori covariances between indicators sharing the same informant or instrument (e.g., Lab-TAB, tester impression, and CBQ) showed a good fit to the data. The inclusion of the residual covariance paths minimally impacted the loadings of the indicators on the factors, but reduced the correlation between the NE and PE latent factors from r = −.29 to r = −.20. All indicators loaded significantly on their latent factors.

Table 1.

Fit of Measurement Models

Model χ2 df CFI RMSEA SRMR
NE PE Model 489.34*** 34 0.538 0.155 (0.143-0.167) 0.116
NE PE Model Covariances 57.429*** 23 0.965 0.052 (0.035-0.069) 0.060
Depression & Anxiety Model 248.10*** 19 0.739 0.161 (0.144-0.180) 0.078
Depression & Anxiety Model Covariances 16.395 10 0.993 0.037 (0.000-0.068) 0.031
Structural Model 226.18*** 109 0.947 0.044 (0.036-0.052) 0.066
Structural Model w/ Sex 276.143*** 123 0.933 0.047 (0.040-0.055) 0.067
Open in a new tab
***

p<.001

CFI = Comparative Fit Index; RMSEA = Root Mean Square Error of Approximation; SRMR = Standardized Root Mean Square Residual

Measurement model: Adolescent depression and anxiety

The depression and anxiety model was specified with the depression factor reflected by K-SADS current depressive symptoms and youth self-, maternal-, and paternal-report CDI scores. Likewise, the anxiety factor was reflected by K-SADS current anxiety symptoms and youth self-, maternal-, and paternal-report SCARED total scores. Fit for this model was poor (Table 1). A second model included a priori covariances between indicators of the same informant and instrument. The revised model fit well and all indicators loaded significantly on their latent factors. The inclusion of the residual covariance paths minimally impacted the loading of the indicators on the factors, but modestly reduced the correlation between the depression and anxiety latent factors (initial model: r =.72; revised model: r = .67).

Given the robust relationship between depression and anxiety, we tested whether a single factor (internalizing psychopathology) would reflect the manifest indicators at least as well as the two-factor model. The models significantly differed (χ2[1] = 32.64, p < .001); thus, we retained our models examining depression and anxiety as separate constructs.

Main effect model: Age 3 temperament and adolescent depression and anxiety

We relied on the revised measurement models which included residual covariance paths for the temperament and internalizing factors models. We regressed the depression and anxiety latent factors on the age 3 PE and NE latent factors. This model was a good fit to the data (Table 1; Figure 1). In this model, age 3 NE was significantly associated with depression (b = .15, SE = .07, t = 2.19, p < .05) and anxiety (b = .21, SE = .06, t = 3.39, p < .01). However, PE was not associated with either depression (b = −.01, SE = .07, t = −.13, p = .90) or anxiety (b = −.04, SE = .06, t = −0.82, p = .41).

Figure 1. Full model showing fully standardized associations between early childhood temperament and adolescent depression and anxiety.

Figure 1.

Open in a new tab

Note: All factor loadings to observed indicators are significant at p<.05. *p<.05; **p<.01; ***p<.001. All model paths are standardized. LT. = Laboratory-Temperament Assessment Battery Observed; TI = Tester Impression Rating; M = Mother Report; F = Father Report; Ch = Child Self-Report; S/L = Smiling Laughter; AA = Approach Anticipation; NA = Negative Affectivity; CDI = Child Depression Inventory; Dep = Depression Symptoms; SCARED = Screen for Anxiety and Related Disorders; Anx = Anxiety Symptoms.

As sex differences are often evident for PE and NE (Olino et al., 2013) and depression and anxiety (Seedat et al., 2009), we included sex as an additional covariate in the model (Table 1). To do so, we modeled the associations between sex as a binary variable and the PE and NE latent factors by regressing the latent PE and NE factors on sex. Similar to the initial model, this model was a good fit to the data. In this model, age 3 NE was significantly associated with depression (b = .15, SE = .07, t = 2.09, p < .05) and anxiety (b = .21, SE = .07, t = 3.14, p < .01). PE was associated with neither depression (b = −.02, SE = .07, t = −0.03, p = .76) nor anxiety (b = −.08, SE = .06, t = −1.36, p = .17). Female sex was associated with anxiety (b = .49, SE = .12, t = 4.29, p < .001), but not depression (b = .19, SE = .13, t = 1.52, p = .13).

We also explored youth race, ethnicity, and parental education as predictors/moderators of the associations between temperamental PE and NE and later depression and anxiety. Non-Hispanic White youth had lower levels of anxiety (b = −.26, SE = .12, t = −2.16, p < .05), and higher levels of parental education were associated with lower levels of depression (b = −.32, SE = .10, t = −3.08, p < .01). No other main effects or interactions were found for these variables.

Moderator models

The main effect models were extended to include moderators of the relationships between age 3 PE and NE and later depression and anxiety factors, while adjusting for youth sex4. We estimated individual models for each moderator. Complete model results are presented in Table 2. There were significant positive associations of maternal depression, maternal anxiety, and youth life events and negative associations of parental acceptance with adolescent depression. There were significant positive associations of female sex, maternal depression, maternal anxiety, and youth life events with adolescent anxiety. We found significant interactions of age 3 NE with maternal depression, paternal anxiety, parental control/autonomy, and parental firm/lax control predicting adolescent anxiety and of NE with life stress, parental control/autonomy, and parental firm/lax control predicting adolescent depression. No variables moderated the relationship between age 3 PE and either adolescent depression or anxiety.

Table 2.

Tests of Moderation of the Association Between NE and PE and Depression and Anxiety

Moderator Parameter Depression
b (SE)		 Anxiety
b (SE)
Child Sex Child Sex 0.193 (0.127) 0.503 (0.117)***
NE 0.153 (0.076)* 0.207 (0.07)**
Sex X NE 0.141 (0.15) −0.068 (0.149)
PE −0.027 (0.072) −0.082 (0.061)
Sex X PE 0.025 (0.146) −0.081 (0.12)
Maternal Depression M.Dep 0.357 (0.13)** 0.303 (0.117)*
NE 0.122 (0.086) 0.202 (0.081)*
M.Dep X NE 0.096 (0.226) 0.419 (0.212)*
PE −0.041 (0.08) −0.095 (0.066)
M.Dep X PE −0.264 (0.21) −0.082 (0.173)
Maternal Anxiety M.Anx 0.349 (0.14)* 0.522 (0.123)**
NE 0.128 (0.079) 0.175 (0.077)*
M.Anx X NE 0.031 (0.182) −0.062 (0.192)
PE −0.031 (0.078) −0.096 (0.066)
M.Anx X PE −0.148 (0.185) −0.138 (0.157)
Paternal Depression D. Dep 0.081 (0.161) 0.286 (0.149)
NE 0.137 (0.074) 0.197 (0.067)**
D. Dep X NE 0.05 (0.211) 0.027 (0.184)
PE −0.021 (0.073) −0.087 (0.06)
D. Dep X PE 0.138 (0.159) 0.23 (0.158)
Paternal Anxiety P.Anx −0.049 (0.125) 0.116 (0.129)
NE 0.139 (0.073) 0.202 (0.068)**
P.Anx X NE −0.024 (0.22) 0.38 (0.157)*
PE −0.016 (0.072) −0.099 (0.062)
P.Anx X PE 0.043 (0.146) −0.073 (0.136)
Life Events Life Events 0.146 (0.044)** 0.084 (0.04)*
NE 0.177 (0.085)* 0.246 (0.074)**
Life Events X NE 0.126 (0.052)* 0.029 (0.041)
PE −0.052 (0.088) −0.062 (0.069)
Life Events X PE −0.028 (0.074) 0.004 (0.052)
Psychological Control vs. Autonomy Control 0.03 (0.029) −0.03 (0.024)
NE 0.161 (0.081)* 0.249 (0.076)**
Control X NE −0.094 (0.038)* −0.06 (0.028)*
PE 0.018 (0.068) −0.018 (0.064)
Control X PE −0.021 (0.028) 0.014 (0.025)
Acceptance vs. Rejection Acceptance −0.121 (0.045)** −0.053 (0.029)
NE 0.155 (0.079) 0.242 (0.074)**
Acceptance X NE −0.016 (0.04) −0.006 (0.037)
PE 0.028 (0.06) −0.013 (0.062)
Acceptance X PE 0.001 (0.036) −0.016 (0.031)
Firm vs. Lax Control Firm −0.018 (0.039) −0.051 (0.031)
NE 0.098 (0.091) 0.20 (0.077)*
Firm X NE −0.191 (0.078)* −0.085 (0.039)*
PE 0.029 (0.067) −0.024 (0.06)
Firm X PE 0.003 (0.045) −0.062 (0.034)
Open in a new tab
*

p<.05

**

p<.01

***

p<.001.

All coefficients are on their raw metrics. NE = Negative Emotionality; PE = Positive Emotionality; M.Dep = Maternal Depressive Disorder; M.Anx = Maternal Anxiety Disorder; P.Dep = Paternal Depressive Disorder; P.Anx = Paternal Anxiety Disorder.

The association between age 3 NE and adolescent anxiety was significant in children of mothers with (b = .49, SE = .20, t = 2.39, p < .05), but not without (b = .07, SE = .07, t = 1.00, p = .31), a history of depression (Figure 2, Top Left). The association between age 3 NE and adolescent anxiety was also significant in children of fathers with (b = .46, SE = .11, t = 4.07, p < .001), but not without (b = .08, SE = .09, t = 0.85, p = .40), a history of anxiety (Figure 2, Top Right).

Figure 2. Conditional effects of temperamental NE on anxiety by maternal depression (Top Left) and paternal anxiety (Top Right) and on depression by stressful life events (Bottom).

Figure 2.

Open in a new tab

Shaded regions around slopes show 95% Confidnce Intervals. Shaded background regions show range of x for which there are associations between the moderator and y. The y-axis is the latent factor score for the specific outcome and the x-axis is the latent NE factor in standardized units.

The association between NE and depression was significant for high (b = .36, SE = .13, t = 2.68, p < .01) and moderate (b = .18, SE = .08, t = 2.08, p < .05), but not low (b = −.01, SE = .09, t = −0.09, p = .92), levels of stressful life events (Figure 2 Bottom).

The associations between NE and both forms of internalizing psychopathology were significant at lower (anxiety: b = .33, SE = .10, t = 3.11, p < .01; depression: b = .38, SE = .10, t = 3.72, p < .001), but not higher (anxiety b = .05, SE = .09, t = 0.49, p = .62; depression b = −.25, SE = . 02, t = −1.21, p = .22), levels of firm/lax control (Figure 3 Top, Left and Right, respectively). At moderate levels of firm/lax control, NE was significantly associated with anxiety (b = .20, SE = .08, t = 2.57, p < .05), but not depression (b = .09, SE = .09, t = 0.99, p = .32).

Figure 3. Conditional effects of NE on depression (Top Left) and anxiety (Top Right) for parental firm vs. lax control and conditional effects of NE on depression (Bottom Left) and anxiety (Bottom Right) for parental psychological control vs. autonomy.

Figure 3.

Open in a new tab

Shaded regions around slopes show 95% Confidnce Intervals. Shaded background regions show range of x for which there are associations between the moderator and y. The y-axis is the latent factor score for the specific outcome and the x-axis is the latent NE factor in standardized units.

The associations between age 3 NE and both forms of adolescent internalizing psychopathology were significant at lower (anxiety: b = .35, SE = .10, t = 3.44, p < .01; depression: b = .32, SE = .09, t = 3.47, p < .01) and moderate (anxiety: b = .25, SE = .08, t = 3.28, p < .01; depression: b = .16, SE = .08, t = 1.99, p < .05), but not higher (anxiety: b = .14, SE = .08, t = 1. 87, p = .07; depression: b = −.01, SE = .12, t = −0.12, p = .90), levels of psychological control/autonomy (Figure 3 Bottom, Left and Right, respectively).

Discussion

Previous studies of associations of temperament with depression and anxiety have largely relied on cross-sectional or short-term longitudinal designs and single informants and methods in older children, adolescents, and adults. These studies have provided consistent support for associations between NE and both depression and anxiety and mixed support for associations between PE and both forms of internalizing psychopathology. We extended these studies by examining longer-term associations starting in preschool and integrating multiple ratings of both temperament and psychopathology. Beyond examining main effect prediction of internalizing problems from temperament, we explored parental history of depression and anxiety, childhood life stress, and parenting styles as moderators of the relationships of NE and PE with depression and anxiety.

Overall, we found that NE, but not PE, at age 3 predicted both depression and anxiety symptoms in adolescence. We also found that multiple variables influenced the magnitude of associations of NE with depression and anxiety, with these interactions showing a fair degree of specificity. However, there were no significant interactions for child biological sex or other demographic characteristics. Thus, the associations identified were consistent across levels of these constructs.

Our finding that NE at age 3 significantly predicted depression and anxiety in mid-adolescence is consistent with Clark and Watson’s (Clark, 2005; Clark & Watson, 1999) model positing that NE is a general risk factor for both depression and anxiety. In contrast to their model, however, we did not find significant main effects of PE on depression or anxiety. It is possible that NE may be more central to the emergence of symptoms than PE, whereas PE may play a role in the course of disorder after onset (Klein & Small, 2014). Alternatively, our assessments of depression at ages 12 and 15 is still early in the risk period (Hankin et al., 1998). Thus, many cases have not yet developed elevated depressive symptoms, possibly limiting power to detect associations between PE and depression.

With the one exception discussed below, the pattern of interactions for associations of NE with depression and anxiety were consistent with a diathesis-stress interpretation. Associations between NE and anxiety were stronger in offspring of depressed mothers and anxious fathers. The former finding is consistent with evidence of heterotypic transmission between depression and anxiety (Lawrence et al., 2019). Both findings extend the literature showing main effects of parental internalizing disorders on youth anxiety by suggesting that the confluence of youth temperament and parental psychopathology may reflect a subgroup of offspring who are at particularly elevated risk for future anxiety. However, a caveat is necessary in interpreting the NE by paternal anxiety disorder effect on offspring anxiety symptoms. Although the simple slope was significant for offspring of fathers with, but not without, a history of anxiety disorders, there was no point on the distribution of youth NE scores within the range examined that significantly distinguished the paternal anxiety disorder groups. Thus, interpretation of this significant interaction warrants some caution.

Higher levels of stress also strengthened the magnitude of the association between NE and depression. The literature on temperament by stress interactions on depression is inconsistent, but our findings are in line with a number of prior studies of adults (Brown & Rosellini, 2011; Kendler et al., 2004; Ormel et al., 2001; van Os & Jones, 1999), as well as an earlier report from our study when children were younger and the stressor, occurring at approximately age 10, was a single event – a hurricane (Kopala-Sibley et al., 2016). However, most of these studies examined proximal stressors, whereas we used more distal stressors that occurred in a 3-year period from 3-9 years before the outcome assessments.

Similarly, lower parental firm (vs. lax) control was associated with higher levels of youth anxiety and depression. Thus, the combination of higher temperamental NE with a parenting style that provides less structure and consistency appears to place youth at greater risk for both depression and anxiety symptoms.

Finally, youth with lower levels of NE exhibited lower levels of depression when parents reported using lower levels of psychological control (i.e., providing greater autonomy). However, youth with higher levels of NE experienced greater depressive symptoms regardless of the degree of parental psychological control. This pattern is consistent with a dual vulnerability interpretation wherein NE and parental high psychological control are statistically non-additive (and conceptually equifinal) risk factors for depression. This suggests that youth with lower levels of temperamental NE may do well with greater autonomy granting from parents, whereas youth with higher levels of temperamental NE may have similar levels of depression regardless of levels of parental autonomy/control.

In contrast, youth with higher levels of NE exhibited higher levels of anxiety symptoms when parents reported higher levels of psychological control (i.e., restricting autonomy). This pattern is consistent with a diathesis-stress interpretation, although there was no point on the distribution of NE scores within the range examined that significantly distinguished youth who experienced higher from lower levels of psychological control on anxiety symptoms. Hence, the nature of this significant interaction must be interpreted cautiously.

There was non-specificity in the main effect of early childhood NE on adolescent anxiety and depressive symptoms. However, there was both specificity and non-specificity in the effects of the moderators. Paternal anxiety, maternal depression, and low parental firm control interacted with youth NE to predict both youth anxiety and depression symptoms. However, the life events by youth NE interaction was evident only for youth depression symptoms (despite a significant main effect for life events on youth anxiety symptoms), and the two significant parental psychological control vs autonomy by youth NE interactions exhibited different patterns in predicting youth depression and anxiety symptoms.

This study benefited from a rigorous longitudinal design with fairly modest attrition across time. We used multiple informants to assess early childhood temperament, as well as adolescent depression and anxiety. Although the effects are relatively small, the findings are nonetheless impressive given the 9-12 year interval between the first and last assessments reported here. However, the work should be considered in light of several limitations. First, life events were assessed from age 6-9 – a long recall period, and quite distal from symptoms at age 12 and 15. Although this reduces temporal confounding between the moderator and dependent variables, it means that our measure of life events reflects relatively distal childhood stress exposure, rather than proximal precipitants of symptoms. Second, we tested a large number of interactions without correcting for multiple comparisons. Thus, these effects should be regarded with caution until they have been replicated. Third, although the sample is representative of the geographic region from which it was drawn, there was modest racial and ethnic diversity in the sample. Fourth, we examined depression and anxiety symptoms, rather than onsets of diagnosable disorders. This afforded greater power, particularly given that age 15 is still relatively early in the risk period for onsets of internalizing psychopathology and the number of diagnosable cases were low. However, extension of these findings to onsets of disorders is needed as the sample enters adulthood. Finally, we assessed moderators on only 1 occasion. It is possible that results of these analyses would be different if we had assessed the moderators at a different time point.

This study examined the longitudinal associations of early childhood PE and NE with adolescent depression and anxiety and moderators of those relationships. Overall, we found that NE was significantly associated with both forms of internalizing symptoms across a 9-12 year follow-up interval. We also found that parental depression and anxiety, parenting styles, and childhood life stress moderated the associations between NE and depressive and anxiety symptoms. Consistent with results of previous studies of offspring of parents with psychopathology (e.g., Cuijpers et al., 2015; Kennedy et al., 2009), these results suggest particular subgroups of children and some processes that might be usefully targeted in preventive interventions.

Supplementary Material

Supplemental Material 1
Supplemental Material 2

Acknowledgments

This work was supported by National Institute of Mental Health Grants R01 MH069942 (Dr. Klein), R01 MH107495 (Dr. Olino), and T32 MH013043 (Dr. Finsaas). The work was approved by the Stony Brook University Committee on Research Involving Human Subjects (protocol 88933 Temperamental Low PE in Preschoolers and Depression Risk). Data are available from the corresponding author upon request. This study was not preregistered.

Footnotes

1

Temperament and personality traditionally refer to individual differences in young children, and older youth and adults, respectively. However, this is an arbitrary distinction as there are strong continuities between core traits (e.g., the “Big Three”) across development (Clark & Watson, 1999; Klein et al., 2012). As this paper focuses on young children, we use the term “temperament’ to refer to traits irrespective of age.

2

NE and PE are viewed as the emotional cores of neuroticism and extraversion, respectively (Clark & Watson, 1999; Klein et al., 2012). For brevity, we use the terms NE and PE to refer to these constructs.

3

There is also a literature on the temperament trait of behavioral inhibition (BI) and risk for anxiety disorders. However, as we are focusing on the broad traits of NE and PE, and BI is much narrower - largely a facet of NE, with some aspects of low PE− we do not consider this literature here (for reviews, see Klein & Mumper, 2018).

4

We also explored the interaction between the latent PE and NE variables. The interaction term was not significantly associated with depression (b = .06, SE = .07, t = .85, p = .39) or anxiety (b = −.03, SE = .07, t = −.43, p = .67).

Contributor Information

Thomas M. Olino, Temple University

Megan C. Finsaas, Columbia University

Margaret W. Dyson, Child Mind Institute

Gabrielle A. Carlson, Stony Brook University

Daniel N. Klein, Stony Brook University

References

  1. Bentler PM (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107(2), 238–246. [DOI] [PubMed] [Google Scholar]
  2. Birmaher B, Brent DA, Chiappetta L, Bridge J, Monga S, & Baugher M (1999). Psychometric properties of the Screen for Child Anxiety Related Emotional Disorders (SCARED): A replication study. Journal of the American Academy of Child & Adolescent Psychiatry, 38(10), 1230–1236. [DOI] [PubMed] [Google Scholar]
  3. Birmaher B, Khetarpal S, Brent D, Cully M, Balach L, Kaufman J, & Neer SM (1997). The Screen for Child Anxiety Related Emotional Disorders (SCARED): Scale construction and psychometric characteristics. Journal of the American Academy of Child & Adolescent Psychiatry, 36(4), 545–553. [DOI] [PubMed] [Google Scholar]
  4. Bould H, Araya R, Pearson RM, Stapinski L, Carnegie R, & Joinson C (2014). Association Between Early Temperament and Depression at 18 Years. Depression and Anxiety, 31(9), 729–736. 10.1002/da.22294 [DOI] [PubMed] [Google Scholar]
  5. Brown TA, & Rosellini AJ (2011). The direct and interactive effects of neuroticism and life stress on the severity and longitudinal course of depressive symptoms. Journal of Abnormal Psychology, 120(4), 844. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Caspi A, Moffitt TE, Newman DL, & Silva PA (1996). Behavioral observations at age 3 years predict adult psychiatric disorders: Longitudinal evidence from a birth cohort. Archives of General Psychiatry, 53(11), 1033–1039. [DOI] [PubMed] [Google Scholar]
  7. Chen N, Deater-Deckard K, & Bell M (2014). The Role of Temperament by Family Environment Interactions in Child Maladjustment. Journal of Abnormal Child Psychology, 42(8), 1251–1262. 10.1007/s10802-014-9872-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chess S, & Thomas A (2013). Goodness of fit: Clinical applications, from infancy through adult life. Routledge. [Google Scholar]
  9. Clark LA (2005). Temperament as a unifying basis for personality and psychopathology. Journal of Abnormal Psychology, 114(4), 505–521. 10.1037/0021-843x.114.4.505 [DOI] [PubMed] [Google Scholar]
  10. Clark LA, & Watson D (1991). Tripartite model of anxiety and depression: Psychometric evidence and taxonomic implications. Journal of Abnormal Psychology, 100(3), 316–336. [DOI] [PubMed] [Google Scholar]
  11. Clark LA, & Watson D (1999). Temperament: A new paradigm for trait psychology. In Pervin LA& John OP (Eds.), Handbook of personality: Theory and research (2nd ed.) (pp. 399–423). Guilford Press. [Google Scholar]
  12. Cuijpers P, Weitz E, Karyotaki E, Garber J, & Andersson G (2015). The effects of psychological treatment of maternal depression on children and parental functioning: A meta-analysis. European Child & Adolescent Psychiatry, 24(2), 237–245. 10.1007/s00787-014-0660-6 [DOI] [PubMed] [Google Scholar]
  13. Davis S, Votruba-Drzal E, & Silk JS (2015). Trajectories of Internalizing Symptoms From Early Childhood to Adolescence: AssociationsWith Temperament and Parenting. Social Development, 24(3), 501–520. 10.1111/sode.12105 [DOI] [Google Scholar]
  14. First MB, Spitzer RL, Gibbon M, & Williams JBW (1996). The Structured Clinical Interview for DSM-IV Axis I Disorders – Non-patient edition. Biometrics Research Department, New York State Psychiatric Institute. [Google Scholar]
  15. Gagne JR, Van Hulle CA, Aksan N, Essex MJ, & Goldsmith HH (2011). Deriving childhood temperament measures from emotion-eliciting behavioral episodes: Scale construction and initial validation. Psychological Assessment, 23(2), 337–353. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Gallitto E (2015). Temperament as a moderator of the effects of parenting on children’s behavior. Development and Psychopathology, 27(3), 757–773. 10.1017/S0954579414000753 [DOI] [PubMed] [Google Scholar]
  17. Goldstein BL, Kotov R, Perlman G, Watson D, & Klein DN (2018). Trait and facet-level predictors of first-onset depressive and anxiety disorders in a community sample of adolescent girls. Psychological Medicine, 48(8), 1282–1290. 10.1017/S0033291717002719 [DOI] [PubMed] [Google Scholar]
  18. Goodman SH (2020). Intergenerational Transmission of Depression. Annual Review of Clinical Psychology, 16(1), 213–238. 10.1146/annurev-clinpsy-071519-113915 [DOI] [PubMed] [Google Scholar]
  19. Gulley LD, Hankin BL, & Young JF (2016). Risk for depression and anxiety in youth: The interaction between negative affectivity, effortful control, and stressors. Journal of Abnormal Child Psychology, 44(2), 207–218. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hakulinen C, Elovainio M, Pulkki-Råback L, Virtanen M, Kivimäki M, & Jokela M (2015). Personality and Depressive Symptoms: Individual Participant Meta-Analysis of 10 Cohort Studies. Depression and Anxiety, 32(7), 461–470. 10.1002/da.22376 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hallquist MN, & Wiley JF (2018). MplusAutomation: An R Package for Facilitating Large-Scale Latent Variable Analyses in M plus. Structural Equation Modeling: A Multidisciplinary Journal, 25(4), 621–638. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hankin BL, Abramson LY, Moffitt TE, Silva PA, McGee R, & Angell KE (1998). Development of depression from preadolescence to young adulthood: Emerging gender differences in a 10-year longitudinal study. Journal of Abnormal Psychology, 107(1), 128–140. [DOI] [PubMed] [Google Scholar]
  23. Hu L, & Bentler PM (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–55. [Google Scholar]
  24. Jeronimus BF, Kotov R, Riese H, & Ormel J (2016). Neuroticism’s prospective association with mental disorders halves after adjustment for baseline symptoms and psychiatric history, but the adjusted association hardly decays with time: A meta-analysis on 59 longitudinal/prospective studies with 443 313 participants. Psychological Medicine, 46(14), 2883–2906. 10.1017/S0033291716001653 [DOI] [PubMed] [Google Scholar]
  25. Jessee A, Mangelsdorf SC, Shigeto A, & Wong MS (2012). Temperament as a Moderator of the Effects of Parental Depressive Symptoms on Child Behavior Problems. Social Development, 21(3), 610–627. 10.1111/j.1467-9507.2011.00639.x [DOI] [Google Scholar]
  26. Kendall AD, Zinbarg RE, Mineka S, Bobova L, Prenoveau JM, Revelle W, & Craske MG (2015). Prospective associations of low positive emotionality with first onsets of depressive and anxiety disorders: Results from a 10-wave latent trait-state modeling study. Journal of Abnormal Psychology, 124(4), 933–943. 10.1037/abn0000105 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kendler KS, Kuhn J, & Prescott CA (2004). The interrelationship of neuroticism, sex, and stressful life events in the prediction of episodes of major depression. American Journal of Psychiatry, 161(4), 631–636. [DOI] [PubMed] [Google Scholar]
  28. Kennedy SJ, Rapee RM, & Edwards SL (2009). A Selective Intervention Program for Inhibited Preschool-Aged Children of Parents With an Anxiety Disorder: Effects on Current Anxiety Disorders and Temperament. Journal of the American Academy of Child & Adolescent Psychiatry, 48(6), 602–609. 10.1097/CHI.0b013e31819f6fa9 [DOI] [PubMed] [Google Scholar]
  29. Klein DN, Dyson MW, Kujawa AJ, & Kotov R (2012). Temperament and internalizing disorders. In Handbook of temperament (pp. 541–561). The Guilford Press. [Google Scholar]
  30. Klein DN, & Finsaas MC (2017). The Stony Brook Temperament Study: Early Antecedents and Pathways to Emotional Disorders. Child Development Perspectives, 11(4), 257–263. 10.1111/cdep.12242 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Klein DN, Kotov R, & Bufferd SJ (2011). Personality and Depression: Explanatory Models and Review of the Evidence. Annual Review of Clinical Psychology, 7, 269–295. 10.1146/annurev-clinpsy-032210-104540 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Klein DN, Lewinsohn PM, Rohde P, Seeley JR, & Olino TM (2005). Psychopathology in the adolescent and young adult offspring of a community sample of mothers and fathers with major depression. Psychological Medicine, 35(03), 353–365. [DOI] [PubMed] [Google Scholar]
  33. Klein DN, & Mumper EE (2018). Behavioral Inhibition as a Precursor to Psychopathology. In Pérez-Edgar K & Fox NA (Eds.), Behavioral Inhibition: Integrating Theory, Research, and Clinical Perspectives (pp. 283–307). Springer International Publishing. 10.1007/978-3-319-98077-5_13 [DOI] [Google Scholar]
  34. Klein DN, & Small AE (2014). Course of depression: Persistence and recurrence. In Gotlib IH & Hammen CL (Eds.), Handbook of Depression and Its Treatment (3rd edition, pp. 64–83). Guilford Press. [Google Scholar]
  35. Kopala-Sibley DC, Danzig AP, Kotov R, Bromet EJ, Carlson GA, Olino TM, Bhatia V, Black SR, & Klein DN (2016). Negative emotionality and its facets moderate the effects of exposure to Hurricane Sandy on children’s postdisaster depression and anxiety symptoms. Journal of Abnormal Psychology, 125(4), 471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Kopala-Sibley DC, Jelinek C, Kessel EM, Frost A, Allmann AES, & Klein DN (2017). Parental depressive history, parenting styles, and child psychopathology over 6 years: The contribution of each parent’s depressive history to the other’s parenting styles. Development and Psychopathology, 29(4), 1469–1482. 10.1017/S0954579417000396 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Kopala-Sibley DC, Klein DN, Perlman G, & Kotov R (2017). Self-Criticism and Dependency in Female Adolescents: Prediction of First Onsets and Disentangling the Relationships Between Personality, Stressful Life Events, and Internalizing Psychopathology. Journal of Abnormal Psychology, 126(8), 1029–1043. 10.1037/abn0000297 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Kotov R, Gamez W, Schmidt F, & Watson D (2010). Linking “big” personality traits to anxiety, depressive, and substance use disorders: A meta-analysis. Psychological Bulletin, 136(5), 768–821. 10.1037/a0020327 [DOI] [PubMed] [Google Scholar]
  39. Kovacs M (1992). Children’s depression inventory. Multi-Health System. [Google Scholar]
  40. Lawrence PJ, Murayama K, & Creswell C (2019). Systematic Review and Meta-Analysis: Anxiety and Depressive Disorders in Offspring of Parents With Anxiety Disorders. Journal of the American Academy of Child & Adolescent Psychiatry, 58(1), 46–60. 10.1016/j.jaac.2018.07.898 [DOI] [PubMed] [Google Scholar]
  41. Leve LD, Kim HK, & Pears KC (2005). Childhood temperament and family environment as predictors of internalizing and externalizing trajectories from ages 5 to 17. Journal of Abnormal Child Psychology, 33(5), 505–520. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. MacCallum RC, Browne MW, & Sugawara HM (2006). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1(2), 130–149. [DOI] [PubMed] [Google Scholar]
  43. Margolies PJ, & Weintraub S (1977). The Revised 56-Item Crpbi as a Research Instrument: Reliability and Factor Structure. Journal of Clinical Psychology, 33(2), 472–476. 10.1002/1097-4679(197704)33:2<472::AID-JCLP2270330230>3.0.CO;2-S [DOI] [Google Scholar]
  44. Marsh HW, Hau KT, & Wen Z (2004). In search of golden rules: Comment on hypothesis-testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler’s (1999) findings. Structural Equation Modeling, 11(3), 320–341. 10.1207/s15328007sem1103_2 [DOI] [Google Scholar]
  45. Mineka S, Williams AL, Wolitzky-Taylor K, Vrshek-Schallhorn S, Craske MG, Hammen C, & Zinbarg RE (2020). Five-year prospective neuroticism–stress effects on major depressive episodes: Primarily additive effects of the general neuroticism factor and stress. Journal of Abnormal Psychology, 129(6), 646–657. 10.1037/abn0000530 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Muthén LK, & Muthén BO (1998). Mplus User’s Guide. Eighth Edition. Muthén & Muthén. [Google Scholar]
  47. Oldehinkel AJ, Veenstra R, Ormel J, De Winter AF, & Verhulst FC (2006). Temperament, parenting, and depressive symptoms in a population sample of preadolescents. Journal of Child Psychology & Psychiatry, 47(7), 684–695. 10.1111/j.1469-7610.2005.01535.x [DOI] [PubMed] [Google Scholar]
  48. Olino TM, Durbin CE, Klein DN, Hayden EP, & Dyson MW (2013). Gender Differences in Young Children’s Temperament Traits: Comparisons Across Observational and Parent-Report Methods. Journal of Personality, 81, 119–129. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Olino TM, Klein DN, Dyson MW, Rose SA, & Durbin CE (2010). Temperamental emotionality in preschool-aged children and depressive disorders in parents: Associations in a large community sample. Journal of Abnormal Psychology, 119(3), 468–478. 10.1037/a0020112 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Ormel J, Oldehinkel AJ, & Brilman EI (2001). The interplay and etiological continuity of neuroticism, difficulties, and life events in the etiology of major and subsyndromal, first and recurrent depressive episodes in later life. American Journal of Psychiatry, 158(6), 885–891. [DOI] [PubMed] [Google Scholar]
  51. R Core Team. (2018). R: A language and environment for statistical computing. R Foundation for Statistical Computing,. [Google Scholar]
  52. Rohde P, Lewinsohn PM, & Seeley JR (1997). Comparability of telephone and face-to-face interviews in assessing axis I and II disorders. American Journal of Psychiatry, 154(11), 1593–1598. [DOI] [PubMed] [Google Scholar]
  53. Schwarz JC, Barton-Henry ML, & Pruzinsky T (1985). Assessing Child-Rearing Behaviors: A Comparison of Ratings Made by Mother, Father, Child, and Sibling on the CRPBI. Child Development, 56(2), 462–479. 10.2307/1129734 [DOI] [PubMed] [Google Scholar]
  54. Seedat S, Scott KM, Angermeyer MC, Berglund P, Bromet EJ, Brugha TS, Demyttenaere K, de Girolamo G, Haro JM, & Jin R (2009). Cross-national associations between gender and mental disorders in the World Health Organization World Mental Health Surveys. Archives of General Psychiatry, 66(7), 785–795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Spinhoven P, Roelofs K, Hovens JG, Elzinga BM, van Oppen P, Zitman FG, & Penninx BW (2011). Personality, life events and the course of anxiety and depression. European Journal of Personality, 25(6), 443–452. 10.1002/per.808 [DOI] [PubMed] [Google Scholar]
  56. Steiger JH (1990). Structural model evaluation and modification: An interval estimation approach. Multivariate Behavioral Research, 25(2), 173–180. [DOI] [PubMed] [Google Scholar]
  57. Tackett JL (2006). Evaluating models of the personality–psychopathology relationship in children and adolescents. Clinical Psychology Review, 26(5), 584–599. 10.1016/j.cpr.2006.04.003 [DOI] [PubMed] [Google Scholar]
  58. van Os J, & Jones PB (1999). Early risk factors and adult person–environment relationships in affective disorder. Psychological Medicine, 29(5), 1055–1067. 10.1017/S0033291799001026 [DOI] [PubMed] [Google Scholar]
  59. Zinbarg RE, Mineka S, Bobova L, Craske MG, Vrshek-Schallhorn S, Griffith JW, Wolitzky-Taylor K, Waters AM, Sumner JA, & Anand D (2016). Testing a hierarchical model of neuroticism and its cognitive facets: Latent structure and prospective prediction of first onsets of anxiety and unipolar mood disorders during 3 years in late adolescence. Clinical Psychological Science, 4(5), 805–824. [Google Scholar]

Associated Data

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

Supplementary Materials

Supplemental Material 1
NIHMS1822434-supplement-Supplemental_Material_1.pdf (2.4MB, pdf)
Supplemental Material 2
NIHMS1822434-supplement-Supplemental_Material_2.xlsx (20.3KB, xlsx)

RESOURCES