Parenting During Early Adolescence and Adolescent-Onset Major Depression

Abstract

In this study, we investigated the prospective relationship between maternal behaviors observed during mother-adolescent interactions and the onset of major depressive disorder (MDD) between early and late adolescence (ages 12–18). Maternal expressions of emotion and maternal responses to their child’s expressions of emotion were both examined. Results demonstrated that higher rates of maternal aggressive behavior and lower rates of maternal positive behavior prospectively predicted MDD onset across adolescence. In addition, negative (i.e., aggressive and dysphoric) maternal responses to adolescents’ aggressive and positive behaviors predicted MDD onset. Maternal dysphoric behavior and the way mothers respond to adolescents’ dysphoria were not related to MDD onset. These results extend previous findings on the relationship between parenting behaviors and depression onset in early to midadolescence and suggest that maternal emotion socialization behaviors in early adolescence prospectively predict MDD onset across the entire course of adolescence.

Keywords

adolescent development depression social interaction emotional development

Depression in adolescence is common, with 12% to 28% of young people experiencing an onset of major depressive disorder (MDD) by age 18 or 19 (Lewinsohn, Rohde, & Seeley, 1998; Merikangas et al., 2010). Myriad family factors have been investigated in relation to adolescent depression (Restifo & Bögels, 2009; Sheeber, Hops, & Davis, 2001; Weich, Patterson, Shaw, & Stewart-Brown, 2009), including parental socialization of emotion (PSE), which has been identified as an important and potentially modifiable component of risk (Eisenberg, Cumberland, & Spinrad, 1998; Morris, Silk, Steinberg, Myers, & Robinson, 2007). PSE is defined as parental behaviors that shape the development of children’s understanding, experience, expression, and regulation of emotion (Denham, Bassett, & Wyatt, 2007; Eisenberg et al., 1998; Morris et al., 2007). Two specific PSE behaviors are considered here: (a) maternal expression of emotion, which is thought to shape the emotional climate of the family and influence the emotional development of children through processes such as modeling and social referencing (Halberstadt, Crisp, & Eaton, 1999; Morris et al., 2007); and (b) maternal response to their child’s expression of emotion, which is a more explicit and active process by which mothers shape their child’s emotional development (Denham et al., 2007; Sheeber, Hops, Andrews, Alpert, & Davis, 1998).

Observational research into parental expressions of emotion has revealed that high rates of negative emotion expression (i.e., aggression, dysphoria) and low rates of positive emotion expression are associated with adolescent depressive symptoms, both cross-sectionally and longitudinally (Burge & Hammen, 1991; Messer & Gross, 1995; Schwartz et al., 2012; Sheeber, Davis, Leve, Hops, & Tildesley, 2007; Sheeber & Sorensen, 1998). Observational research into parental responses to youth emotional behaviors has generally been conducted by comparing the family interactions of young people diagnosed with depression or elevated depressive symptoms with the family interactions of nondepressed control participants. Such studies have shown that parental reinforcement of youth sadness or dysphoria (Dadds, Sanders, Morrison, & Rebgetz, 1992; Schwartz et al., 2011; Sheeber, Allen, Davis, & Sorensen, 2000; Sheeber et al., 1998), parental reciprocity of anger (Cook, Asarnow, Goldstein, Marshall, & Weber, 1990; Schwartz et al., 2011), and a lack of parental reinforcement of positive behavior (Cole & Rehm, 1986; Messer & Gross, 1995; Yap, Allen, & Ladouceur, 2008) have all been associated with youth depression. To the best of our knowledge, however, the only published longitudinal research examining family behaviors as predictors of case-level MDD onset is that reported in a previous analysis of the current cohort. Schwartz et al. (2011) reported that the onset of MDD up to midadolescence (i.e., approximately age 15) was prospectively predicted by maternal aggression and dysphoria in response to adolescent aggression, as well as by the suppression of maternal aggression and dysphoria in response to adolescent dysphoria.

The current study fills four important gaps in the existing knowledge base. First, no published research exists on the prospective relationship between maternal expression of emotion and the onset of case-level depressive disorder (as opposed to elevated symptoms) during adolescence. Thus, we do not know whether higher rates of negative maternal behavior and lower rates of positive behavior are present prior to the onset of disorder or whether these behaviors emerge as concomitant or consequent features of adolescent depression. Second, in our previous work, we identified parental response behaviors that prospectively predict depression onset up to age 15 (Schwartz et al., 2011). This work is fundamentally limited by its inability to account for depression onset in the late adolescent years from age 15 to 19, which are the peak years for adolescent-onset depression (Hankin et al., 1998; Kessler, McGonagle, Swartz, Blazer, & Nelson, 1993; Seeley & Lewinsohn, 2009). In the current study, we considered depression onset across the entire course of adolescence, from age 12 to age 18 or 19, thereby providing a comprehensive assessment of the prospective relationship between affective parenting behaviors and adolescent-onset depression. Third, in the current study, we significantly extend previous work by examining both maternal expressions of emotion and maternal responses to their child’s emotions within one study, thereby providing the most thorough examination to date of both individual and interactional maternal socialization behaviors and their prospective relationship with adolescent-onset depression. Fourth, in most observational studies, researchers have collected parenting behavior data only in a single context (usually a family problem-solving discussion designed to elicit conflict) and, thus, cannot address the question whether it is the overall rates of behaviors across contexts or the rates of behaviors within specific types of interactions that are most strongly predictive. In this study, we used observational assessments during interaction tasks designed to differentially elicit positive and negative affect, thereby enabling an explicit examination of the effect of the interactional context, a factor that has been increasingly emphasized as important to the study of affective processes (Aldao, 2013). Our previous work using positive and negative interaction tasks suggests that negative parental behaviors exhibited during a positive task are particularly salient in regard to the prediction of adolescent depression, possibly because they are not elicited by task demands (Schwartz et al., 2011). Although we have not found positive parental behaviors during a negative task to be similarly predictive (Schwartz et al., 2012), there is evidence from the marital literature that the use of positive affect during conflictual discussions is a predictor of both marital stability and marital satisfaction 6 years after the wedding (Gottman, Coan, Swanson, & Carrere, 1998), which suggests that a similar interpersonal process may also corrode parent-child relationships and increase risk for adolescent depression.

In addition to addressing these gaps, this study extends the existing literature by controlling for maternal depressive symptoms in examining the relations between parenting behavior and onset of adolescent MDD. Given that maternal depression has been identified as a risk factor for emergence of depressive disorder and has been shown to affect parenting behaviors (Goodman & Gotlib, 1999; Lovejoy, Graczyk, O’Hare, & Neuman, 2000), we controlled for maternal depressive symptoms and thereby enabled a clearer delineation of the associations between parenting behavior and outcome by removing the variance in maternal parenting behaviors attributable to maternal depressive symptoms per se.

The aim of the current study was to identify maternal socialization behaviors that prospectively predict the onset of case-level MDD across the entire course of adolescence. On the basis of previous findings, we hypothesized that MDD onset would be predicted by higher rates of maternal aggression, lower rates of maternal positivity, negative maternal responses to adolescent aggression and positivity, and the suppression of maternal negativity in response to adolescent dysphoria. In addition, we hypothesized that both negative maternal behaviors during the positive interaction task and positive maternal behaviors during the negative interaction task would predict adolescent depression.

Method

The Orygen Adolescent Development Study (ADS) is a large longitudinal study conducted in Melbourne, Australia.¹ The first wave of intensive data collection (T1) included mother-adolescent interactions, a diagnostic interview to exclude adolescent participants with a history of MDD, and adolescent- and parent-rated questionnaires. Waves 2, 3, and 4 (T2–T4) were conducted approximately 2.5, 4, and 6 years after T1, respectively, and included a diagnostic interview and adolescent- and parent-rated questionnaires.

Recruitment and screening of participants

The recruitment and screening of participants has been reported in detail previously (e.g., Yap, Allen, & Ladouceur, 2008). Screening was conducted to identify a community sample of 10- to 12-year-old primary school students representing the full spectrum of risk for psychopathology as a function of temperament, as measured by the Early Adolescent Temperament Questionnaire–Revised (Ellis & Rothbart, 2001). To ensure equal representation of adolescents at high, medium, and low risk, we selected equal numbers of students with scores on the Negative Affectivity and Effortful Control dimensions of temperament that were 0 to 1, 1 to 2, 2 to 2.5, and more than 2.5 standard deviations above and below the mean. These dimensions were chosen because they have been associated with depressed mood in adolescents (Ellis & Rothbart, 2001). A total of 2,453 students (53.5% of the total sampling population; 52% female, 48% male; mean age = 11.62 years) participated in the screening, from which a sample comprising 415 students was selected on the basis of temperament scores. Of these selected students, 245 participated in the T1 intensive data collection.

Family-interaction data were available for 197 participants. Two adolescents were diagnosed with MDD at T1 and were excluded to enable the study to be prospective in relation to MDD onset, leaving a sample of 195 participants (95 females and 100 males). Of the 195 families, 160 were mother-adolescent dyads and 35 were father-adolescent dyads. Given the low number of participating fathers, only the mother-adolescent dyads were included in the current analyses, leaving a final sample comprising 160 participants. There were no differences between these 160 participants and the remaining 255 of the selected 415 participants with regard to gender, χ²(1, N = 415) = 0.15, n.s., d = 0.04, age, t(413) = 0.19, n.s., d = 0.02, socioeconomic status, t(413) = 0.41, n.s., d = 0.04, or temperament scores on Negative Affectivity, t(413) = 0.16, n.s., d = 0.02, and Effortful Control, t(413) = 0.001, n.s., d < 0.01.

Procedure

Psychosocial assessments at T1 to T4

At each of the four waves of the ADS, adolescents completed a semistructured diagnostic interview and were reimbursed for their time. The T1 interview assessed for current and lifetime episodes of MDD, and the T2 through T4 diagnostic interviews assessed for current MDD and any new episodes of MDD since the date of the last assessment. The rate of participation in each wave is outlined in Figure 1. Diagnostic interview data from all of the time points were integrated by two reviewers (O. S. Schwartz and M. L. Byrne) who made independent ratings and then constructed a consensus variable that indicated whether participants had experienced an onset of MDD at any time between the T1 and T4 time points. This variable was calculated for the 120 participants for whom diagnostic data were available at each time point. Between-group analyses indicated that there were no differences between these 120 participants and the 40 participants with missing data (for whom these data were imputed; see our discussion later) with regard to gender, χ²(1, N = 160) = 0.00, n.s., d < 0.01, socioeconomic status, t(156) = 1.16, n.s., d = 0.22, and T1 scores on Negative Affectivity, t(157) = −0.58, n.s., d = 0.11, Effortful Control, t(158) = 0.58, n.s., d = 0.11, and depression symptoms (as measured by the Centre for Epidemiological Symptoms−Depression scale, CES-D; Radloff, 1977), t(130) = 1.11, n.s., d = 0.22. Participants with missing T1 through T4 MDD-onset data were found to be younger than participants with data, t(158) = 2.40, p = .02, d = 0.44; however, because the average difference in age was approximately 2 months, this difference was not considered developmentally meaningful.

Fig. 1.

Flowchart of participation across Waves 1 through 4 (T1–T4, respectively).

Family-interaction assessment at T1

Adolescents and mothers completed the lab-based interaction assessment approximately 2 months (range = 0.03–8.04 months) after the psychosocial assessment at T1. Mother-adolescent dyads completed two 20-min interaction tasks that were video recorded for subsequent coding. An event-planning interaction (EPI) was completed first, followed by a problem-solving interaction (PSI). The EPI and PSI tasks were intended to differentially elicit positive and negative behavior, respectively. The ordering of tasks was fixed because of concern that negative affective states elicited by the PSI had the potential to persist into the positive EPI if conducted second (Gilboa & Revelle, 1994).

For the EPI, mothers and adolescents were instructed to plan one or more pleasant activities to do together, with up to five activities chosen on the basis of items that both the mother and adolescent rated as being “very pleasant” on the Pleasant Events Schedule (MacPhillamy & Lewinsohn, 1976). For the PSI, mother-adolescent dyads were instructed to try to resolve one or more issues of disagreement, with up to five issues selected that the mother and adolescent endorsed as occurring the most frequently and generating the highest intensity of anger on the Issues Checklist (Prinz, Foster, Kent, & O’Leary, 1979). One family did not complete the PSI. Mothers and adolescents were reimbursed for their time and travel expenses.

Measures

Living in Family Environments (LIFE) coding system

The LIFE (Hops, Biglan, Tolman, Arthur, & Longoria, 1995) is an observational, microsocial coding system that allows for a detailed analysis of individual family members’ behaviors and interactive family behaviors. The LIFE system consists of 10 nonverbal affect codes (e.g., anger, dysphoria, happy) and 27 verbal content codes (e.g., validation, complaint, provoke). To code the video-recorded interactions, we used an event-based protocol in which new codes were entered each time the affect or content of one of the interactants changed. The affect and content codes were used to develop composite behavior constructs. In this study, the constructs were aggressive, dysphoric, and positive behaviors. The aggressive construct included all behaviors with contemptuous, angry, or belligerent affect, as well as cruel, provocative, annoying/disruptive, or argumentative verbal statements made with neutral affect. The dysphoric construct consisted of all behaviors with dysphoric, anxious, or whining affect, as well as complaints and self-derogatory verbal comments made with neutral affect. The positive construct included all behaviors with happy or caring affect as well as approving, validating, affectionate, or humorous comments made with neutral affect.

We used the LIFE data to construct two variables representing maternal socialization behaviors. First, the rate-per-minute variables were used to measure maternal expression of emotion. These variables indicate the average number of times a mother expressed each behavior type (i.e., aggressive, dysphoric, and positive) per minute and were calculated separately for the EPI and PSI. Second, the conditional probabilities of negative maternal behaviors occuring in response to adolescent behaviors were measured using behavior sequences in which adolescents’ aggressive, dysphoric, and positive behaviors were followed by maternal aggressive and dysphoric behaviors. These behavior sequences were selected for analysis on the basis of previous research (Schwartz et al., 2011, 2012; Yap, Allen, & Ladouceur, 2008). The conditional probability of a particular maternal behavior occurring immediately after a particular adolescent behavior was calculated by dividing the number of times the particular adolescent behavior was followed by the particular maternal behavior by the base rate (i.e., total frequency) of the particular adolescent behavior (Gottman & Roy, 1990). We then used these conditional probabilities to calculate Allison-Liker z scores (Allison & Liker, 1982), using a binomial test that also takes into account the base rate of the maternal behaviors. The z scores thus reflect the extent to which a specified sequence of behavior occurred more or less frequently than would be expected as a function of the base rate of each behavior based on a normal distribution.

Coders were extensively trained and blind to the clinical and demographic characteristics of the participants. Approximately 20% of the interactions were coded by a second observer to provide an estimate of observer agreement. Random pairs of observers were assigned to the interactions to minimize “drift” between any two observers. Interobserver agreement was assessed using kappa, a conservative index that controls for chance agreement (Fleiss, 1981). Kappa coefficients for the aggressive, dysphoric, and positive behavior constructs were .70, .60, and .86, respectively, across interactants. The validity of the LIFE system as a measure of family processes has been established in numerous studies (e.g., Katz & Hunter, 2007; Sheeber et al., 2007).

Schedule for Affective Disorders and Schizophrenia for School-Age Children, Present and Lifetime version (K-SADS-PL)

The K-SADS-PL (Kaufman et al., 1997) is a semistructured diagnostic interview assessing current and lifetime symptoms and diagnoses of Axis I disorders in youths aged 6 to 18 years. The K-SADS-PL has demonstrated good interrater agreement, test-retest reliability, and diagnostic validity (Kaufman et al., 1997). Recordings of 20% to 25% of all K-SADS-PL interviews at each of the four ADS waves of data collection were rerated by a second coder. The kappa coefficient for symptom-level agreement was 0.80 at T1, 0.78 at T2, 0.86 at T3, and 0.81 at T4, which suggests excellent interrater reliability across diagnoses at all time points (Kaufman et al., 1997).

CES-D

The CES-D (Radloff, 1977) is a 20-item self-report questionnaire designed to measure depression symptoms in the general population. Higher scores indicate greater frequency and number of depression symptoms. The CES-D was used to quantify levels of depressive symptoms among mothers who took part in the study.

Results

Descriptive analyses

Table 1 shows the rate of first-onset MDD at each of the three follow-up time points and the total number of participants who experienced first-onset MDD following the T1 assessment. Overall, significantly more females than males experienced MDD onset during the follow-up period, χ²(1, N = 120) = 6.14, p = .013.

Table 1.

Rates of Major Depressive Disorder Onset Between Time 2 and Time 4

Time	Female	Male	Total
T2 (female: n = 70; male: n = 73)	7	5	12
T3 (female: n = 60; male: n = 59)	6	1	7
T4 (female: n = 54; male: n = 59 )	9	4	13
T2–T4 (female: n = 60; male: n = 60)	22 (36.67%)	10 (16.67%)	32** (26.67%)

Significantly more females than males, p < .05.

Design of path analyses

Four path analyses were used to test whether maternal PSE behaviors prospectively predicted MDD onset across adolescence. Two models assessed the association between the rate per minute of the three maternal behavior constructs and MDD onset, with separate models specified using data from the EPI and PSI, and two models assessed the association between negative maternal responses (i.e., aggressive and dysphoric) to adolescent aggressive, dysphoric, and positive behavior and MDD onset, with separate models specified for data taken from the two tasks. In all models, we controlled for maternal depression in the prediction of the observed maternal behavior variables and for the duration of time between the T1 psychosocial and family-interaction assessments on the MDD-onset outcome variable (see Figs. 2 –5; for ease of presentation, the figures do not show the paths representing the regression of MDD onset on the duration of time between the T1 assessments or the covariances between the observed parenting variables). Path analyses were run using Mplus software (version 7.11) and weighted least squares mean- and variance-adjusted estimation. Missing data were estimated using full-information maximum likelihood in Mplus. Full-information maximum likelihood yields comparatively good standard errors and is efficient in evaluating parameter significance with missing categorical data (Graham, 2009; Wothke, 2000).

Fig. 2.

The influence of rates of maternal behaviors during the EPI on MDD onset across adolescence. Thick black arrows represent significant regression paths for girls and boys. Thin grey arrows represent nonsignificant regression paths for girls and boys. Data for girls are provided on the left side of the forward slash. RPM = rate per minute; EPI = event-planning interaction; MDD = major depressive disorder; CFI = comparative fit index; RMSEA = root-mean-square error of approximation; WRMR = weighted root-mean-square residual; T2–T4 = Waves 2–4.

Fig. 3.

The influence of rates of maternal behaviors during the PSI on MDD onset across adolescence. Thick black arrows represent significant regression paths for girls and boys. Thin grey arrows represent nonsignificant regression paths for girls and boys. Data for girls are provided on the left side of the forward slash. RPM = rate per minute; PSI = problem-solving interaction; MDD = major depressive disorder; CFI = comparative fit index; RMSEA = root-mean-square error of approximation; WRMR = weighted root-mean-square residual; T2–T4 = Waves 2–4.

Fig. 4.

The influence of maternal response behaviors during the EPI on MDD onset across adolescence. Thick black arrows represent significant regression paths for girls and boys. Thick dashed arrows represent regression paths that are significant for boys but not girls. Thin grey arrows represent nonsignificant regression paths for girls and boys. Data for girls are provided on the left side of the forward slash. Mat = maternal; Agg = aggressive; Dys = dysphoric; Pos = positive; | = given; Adol = adolescent; EPI = event-planning interaction; MDD = major depressive disorder; CFI = comparative fit index; RMSEA = root-mean-square error of approximation; WRMR = weighted root-mean-square residual; T2–T4 = Waves 2–4.

Fig. 5.

The influence of maternal response behaviors during the PSI on MDD onset across adolescence. Thick black arrows represent significant regression paths for girls and boys. Thick dotted arrows represent regression paths that are significant for girls but not boys. Thin grey arrows represent nonsignificant regression paths for girls and boys. Data for girls are provided on the left side of the forward slash. Mat = maternal; Agg = aggressive; Dys = dysphoric; Pos = positive; | = given; Adol = adolescent; PSI = problem-solving interaction; MDD = major depressive disorder; CFI = comparative fit index; RMSEA = root-mean-square error of approximation; WRMR = weighted root-mean-square residual; T2–T4 = Waves 2–4.

Analysis of adolescent gender

A multiple-group analysis was employed in all models to test the invariance of regression parameters across adolescent gender (Muthén & Muthén, 1998–2012). All parameter estimates corresponding to regression coefficients initially were constrained to be equal across groups. Following the procedure recommended by Muthén and Muthén (1998–2012), we sequentially released constrained parameters with a modification index of approximately 2 or more in both groups and reran the model. A chi-square difference test determined whether releasing the identified parameter resulted in a significant improvement in model fit. No modifications were made to the two models analyzing the rate-per-minute variables, which suggested no significant gender moderation in these models. In both models analyzing the conditional probability variables, the gender equality constraint was released for the regression of MDD onset on the conditional probability of maternal aggressive behavior given adolescent aggressive behavior—EPI: χ²_diff(1, N = 160) = 13.27, p < .001; PSI: χ²_diff(1, N = 160) = 9.62, p = .002. And in the EPI, the regression of the conditional probability of maternal dysphoria given adolescent positivity on maternal depressive symptoms was also released, χ²_diff(1, N = 160) = 6.69, p = .01, which indicated gender moderation of these paths.

Path analysis results

Model fit indices showed that the final version of all models provided an acceptable fit to the data (see Figs. 2 –5).

Maternal depressive symptoms

Results showed that higher levels of maternal depressive symptoms predicted higher rates of maternal dysphoric behavior in the EPI—Girls: β = 0.17, 95% confidence interval (CI) = [0.04, 0.29]; Boys: β = 0.15, 95% CI = [0.03, 0.26]. Higher levels of maternal depressive symptoms also predicted lower rates of maternal positivity in the PSI—Girls: β = −0.26, 95% CI = [–0.42, –0.10]; Boys: β = −0.26, 95% CI = [–0.42, –0.11] (see Tables 2 and 3 for regression parameter estimates). In both interactions, maternal depressive symptoms were also associated with an increased likelihood of mothers’ reciprocating their adolescent’s dysphoric behavior—EPI Girls: β = 0.14, 95% CI = [0.01, 0.26]; EPI Boys: β = 0.15, 95% CI = [0.01, 0.29]; PSI Girls: β = 0.24, 95% CI = [0.09, 0.40]; PSI Boys: β = 0.22, 95% CI = [0.07, 0.36]. In the EPI, maternal depressive symptoms were associated with a lower likelihood of responding with dysphoria to boys’ positivity, β = −0.24, 95% CI = [–0.44, –0.04]. In the PSI, maternal depressive symptoms were associated with an increased likelihood of mothers’ both reciprocating aggression, Girls: β = 0.32, 95% CI = [0.12, 0.51]; Boys: β = 0.25, 95% CI = [0.11, 0.39], and responding with aggression to adolescent’s positivity—Girls: β = 0.22, 95% CI = [0.10, 0.33]; Boys: β = 0.20, 95% CI = [0.08, 0.31].

Table 2.

Regression Parameter Estimates for Rate-Per-Minute Data

Parameter	b (SE)	Unstandardized 95% CI		β
Parameter	b (SE)	Lower 2.5%	Upper 2.5%	Girl	Boy
Mat Agg EPI regressed on
Mat CES-D	0.01 (0.00)	−0.00	0.01	0.14	0.14
Mat Dys EPI regressed on
Mat CES-D	0.01 (0.00)	0.00	0.01	0.17	0.15
Mat Pos EPI regressed on
Mat CES-D	−0.01 (0.01)	−0.02	0.00	−0.15	−0.14
MDD onset regressed on
Mat Agg EPI	0.73 (0.30)	0.15	1.31	0.30	0.29
Mat Dys EPI	−0.09 (0.31)	−0.69	0.51	−0.03	−0.04
Mat Pos EPI	−0.30 (0.27)	−0.82	0.23	−0.15	−0.15
T1 Ax Lag	0.08 (0.07)	−0.06	0.21	0.09	0.12
Mat Agg PSI regressed on
Mat CES-D	0.01 (0.01)	−0.00	0.02	0.14	0.13
Mat Dys PSI regressed on
Mat CES-D	0.00 (0.00)	−0.00	0.01	0.09	0.08
Mat Pos PSI regressed on
Mat CES-D	−0.02 (0.01)	−0.03	−0.01	−0.26	−0.26
MDD onset regressed on
Mat Agg PSI	−0.02 (0.24)	−0.50	0.46	−0.01	−0.01
Mat Dys PSI	0.07 (0.44)	−0.79	0.92	0.02	0.02
Mat Pos PSI	−0.44 (0.18)	−0.79	−0.08	−0.31	−0.29
T1 Ax Lag	0.08 (0.07)	−0.06	0.22	0.10	0.13

Note: Statistically significant confidence intervals (CIs) are highlighted in bold font. Mat = maternal; Agg = aggressive; EPI = event-planning interaction; CES-D = Centre for Epidemiological Studies–Depression scale; Dys = dysphoric; Pos = positive; MDD = major depressive disorder; T1 Ax Lag = duration between Time 1 psychosocial and family-interaction assessments; PSI = problem-solving interaction.

Table 3.

Regression Parameter Estimates for Conditional Probability Variables

		Unstandardized 95% CI		β
Parameter	b (SE)	Lower 2.5%	Upper 2.5%	Girl	Boy
Mat Agg \| Adol Agg EPI regressed on
Mat CES-D	0.03 (0.01)	−0.00	0.05	0.16	0.16
Mat Dys \| Adol Agg EPI regressed on
Mat CES-D	0.00 (0.01)	−0.01	0.02	0.02	0.02
Mat Agg \| Adol Dys EPI regressed on
Mat CES-D	0.01 (0.01)	−0.00	0.03	0.10	0.12
Mat Dys \| Adol Dys EPI regressed on
Mat CES-D	0.02 (0.01)	0.00	0.04	0.14	0.15
Mat Agg \| Adol Pos EPI regressed on
Mat CES-D	0.01 (0.01)	−0.00	0.02	0.07	0.08
Mat Dys \| Adol Pos EPI regressed on
Mat CES-D	0.02/–0.02 (0.01/0.01)	−0.00/−0.04	0.04/−0.00	0.20	−0.24
MDD onset regressed on
Mat Agg \| Adol Agg EPI	0.21/−0.56 (0.09/0.18)	0.03/−0.92	0.38/−0.20	0.34	−0.84
Mat Dys \| Adol Agg EPI	0.48 (0.15)	0.19	0.76	0.38	0.40
Mat Agg \| Adol Dys EPI	0.18 (0.11)	−0.02	0.39	0.24	0.19
Mat Dys \| Adol Dys EPI	−0.03 (0.11)	−0.25	0.18	−0.05	−0.04
Mat Agg \| Adol Pos EPI	0.15 (0.09)	−0.03	0.34	0.20	0.19
Mat Dys \| Adol Pos EPI	0.43 (0.13)	0.17	0.69	0.40	0.38
T1 Ax Lag	0.08 (0.07)	−0.05	0.21	0.10	0.12
Mat Agg \| Adol Agg PSI regressed on
Mat CES-D	0.09 (0.03)	0.04	0.14	0.32	0.25
Mat Dys \| Adol Agg PSI regressed on
Mat CES-D	−0.00 (0.01)	−0.02	0.01	−0.04	−0.04
Mat Agg \| Adol Dys PSI regressed on
Mat CES-D	−0.01 (0.01)	−0.03	0.02	−0.05	−0.04
Mat Dys \| Adol Dys PSI regressed on
Mat CES-D	0.03 (0.01)	0.01	0.05	0.24	0.22
Mat Agg \| Adol Pos PSI regressed on
Mat CES-D	0.03 (0.01)	0.02	0.05	0.22	0.20
Mat Dys \| Adol Pos PSI regressed on
Mat CES-D	−0.00 (0.01)	−0.01	0.01	−0.01	−0.01
MDD onset regressed on
Mat Agg \| Adol Agg PSI	0.29/−0.04 (0.07/0.07)	0.16/−0.18	0.41/0.10	0.81	−0.15
Mat Dys \| Adol Agg PSI	0.13 (0.12)	−0.12	0.37	0.13	0.11
Mat Agg \| Adol Dys PSI	−0.07 (0.08)	−0.22	0.08	−0.10	−0.12
Mat Dys \| Adol Dys PSI	−0.01 (0.11)	−0.22	0.20	−0.02	−0.02
Mat Agg \| Adol Pos PSI	−0.08 (0.07)	−0.21	0.05	−0.13	−0.14
Mat Dys \| Adol Pos PSI	0.25 (0.11)	0.03	0.47	0.25	0.25
T1 Ax Lag	0.08 (0.07)	−0.06	0.21	0.10	0.13

Note: When estimates are equal for girls and boys, only a single estimate is reported. When estimates differ for girls and boys, the estimate for girls is given on the left side of the forward slash. Statistically significant confidence intervals (CIs) are highlighted in bold font. Mat = maternal; Agg = aggressive; | = given; Adol = adolescent; EPI = event-planning interaction; CES-D = Centre for Epidemiological Studies–Depression scale; Dys = dysphoric; Pos = positive; MDD = major depressive disorder; PSI = problem-solving interaction; T1 Ax Lag = duration between Time 1 psychosocial and family-interaction assessments.

Maternal socialization behaviors and MDD onset

After accounting for the influence of maternal depressive symptoms on the maternal socialization behaviors, we found that a number of these behaviors prospectively predicted MDD onset in adolescents. Table 2 shows that for both girls and boys, MDD onset during adolescence was prospectively predicted by higher rates of maternal aggressive behavior during the EPI, Girls: β = 0.30, 95% CI = [0.06, 0.55]; Boys: β = 0.29, 95% CI = [0.06, 0.51], and by lower rates of maternal positive behavior during the PSI—Girls: β = −0.31, 95% CI = [–0.56, –0.05]; Boys: β = −0.29, 95% CI = [–0.52, –0.05]. Rates of maternal dysphoric behavior did not predict MDD onset in either model.

Table 3 shows that maternal responses to both adolescent aggression and positivity prospectively predicted MDD onset in adolescence. Specifically, adolescent girls were at increased risk for MDD onset when their mothers were more likely to reciprocate their expressions of aggression in both the EPI, β = 0.34, 95% CI = [0.05, 0.62], and the PSI, β = 0.81, 95% CI = [0.53, 1.08]. In contrast, adolescent boys were at increased risk for MDD onset when their expression of aggression suppressed their mothers’ aggression during the EPI, β = −0.84, 95% CI = [–1.36, –0.32], but this behavior sequence during the PSI was not associated with MDD onset in boys, β = −0.15, 95% CI = [–0.62, 0.32]. Both girls and boys were at increased risk of MDD onset when their mothers responded with dysphoric behavior to the adolescents’ expression of aggression during the EPI but not during the PSI—EPI Girls: β = 0.38, 95% CI = [0.14, 0.62]; EPI Boys: β = 0.40, 95% CI = [0.15, 0.64]; PSI Girls: β = 0.13, 95% CI = [–0.12, 0.37]; PSI Boys: β = 0.11, 95% CI = [–0.10, 0.32]. Finally, adolescents whose mothers were more likely to respond with dysphoric behavior to the adolescents’ expression of positivity in both interaction tasks were at increased risk of MDD onset—EPI Girls: β = 0.40, 95% CI = [0.16, 0.65]; EPI Boys: β = 0.38, 95% CI = [0.14, 0.62]; PSI Girls: β = 0.25, 95% CI = [0.03, 0.47]; PSI Boys: β = 0.25, 95% CI = [0.02, 0.48]. Maternal responses to adolescent dysphoric behavior and maternal aggressive responses to adolescent positivity were not prospectively associated with MDD onset.

Discussion

The results showed that higher rates of maternal aggressive behavior and lower rates of maternal positive behavior prospectively predicted the onset of MDD across adolescence. Consistent with previous findings, our findings showed that there was no evidence that maternal expressions of dysphoria were associated with MDD onset. In addition, we found that maternal aggressive responses to adolescent aggression predicted adolescent-onset MDD in girls, whereas the suppression of maternal aggression in response to adolescent aggression was associated with risk for MDD onset for boys. In both boys and girls, maternal dysphoric responses to adolescent aggression and positivity prospectively predicted MDD onset across adolescence. However, there was no evidence that maternal responses to adolescent dysphoria were associated with MDD onset. Overall, these results indicate that maternal expression of emotion and the way mothers respond to their child’s expression of emotion during early adolescence are prospectively associated with MDD onset across the entire course of adolescence. It is important to note that we controlled for the influence of maternal depressive symptoms on maternal behaviors and, thus, discounted the possibility that the observed relationships were purely attributable to the effects of maternal depression on parenting (which were again demonstrated in this study).

The finding that MDD onset across adolescence was prospectively predicted by both aggressive and positive maternal behaviors is consistent with findings from previous research that have respectively associated negative and positive parental behaviors with higher and lower levels of depressive symptoms over time (Burge & Hammen, 1991; Schwartz et al., 2012; Sheeber, Hops, Alpert, Davis, & Andrews, 1997). The current findings demonstrate that such parenting behaviors also prospectively predict the onset of case-level depressive disorder across adolescence.

As predicted, the affective context of family interactions appeared to play an important role in determining the predictive salience of some maternal socialization behaviors. Specifically, rates of maternal aggressive behavior during the positive EPI task significantly predicted MDD onset (even though rates of maternal aggression were higher during the PSI), whereas rates of maternal positive behavior during the negative PSI task were associated with risk for MDD onset. Thus, the predictive maternal behaviors occurred during tasks that were not designed to elicit these behaviors. It is possible that the predictive importance of maternal behaviors is contextually determined and not based simply on the rates of these behaviors. Specifically, adolescents whose mothers behave aggressively when asked to discuss something pleasant with their child, or who display low levels of positivity when attempting to resolve a conflictual topic, may be at particular risk for depression onset across the course of their adolescence.

However, the interactional context was less differentiating in relation to the results based on mothers’ responses to adolescent emotional behavior. Rather, the results across both tasks were similar in suggesting that maternal responses to adolescent aggression and positivity are particularly relevant in relation to MDD onset, whereas maternal responses to dysphoria are not.

Links between aggressive or angry behavior and depression have been well documented (e.g., Painuly, Sharan, & Mattoo, 2005; Safer, 2009). We speculate that our findings might suggest that the way young people are socialized in relation to their anger and, thus, how they learn to regulate their anger may be particularly important for the development of depression across adolescence (Jackson, Kuppens, Sheeber, & Allen, 2011). Our findings regarding the reciprocation of aggressive behavior were different for girls and boys, which suggests that different family-interaction processes may be operating (Cassano, Perry-Parrish, & Zeman, 2007; Eisenberg et al., 1998). For girls, it may be that maternal reciprocity of aggression is particularly stressful, or this finding might reflect a more conflictual mother-daughter relationship for girls who go on to develop depression. For boys, the reverse may be true, in that maternal suppression of aggression in response to their sons’ aggression may function as a negative reinforcement of this behavior and could possibly increase vulnerability to depression through an externalizing pathway (Kendler, Gardner, & Prescott, 2006).

Consistent with the findings in our earlier work with this sample, in which we investigated predictors of MDD onset in midadolescence, our current results showed that maternal dysphoric responses to adolescent aggression prospectively predicted MDD onset into late adolescence, which speculatively could indicate a maladaptive form of conflict resolution that prevents adolescents from learning optimal aggression-regulation strategies. Overall, our findings suggest that the way mothers respond to adolescent aggressive behavior, or the way that conflict is managed within the family, may be particularly important predictors of future depression.

Our findings also highlight that maternal dysphoric responses to adolescent positivity are prospectively associated with MDD. A study based on the first wave of the ADS data demonstrated a cross-sectional association between this behavior sequence (referred to as “dampening”) and depressive symptoms in early adolescence (Yap, Allen, & Ladouceur, 2008), and the current findings are the first to suggest that such behaviors are also prospectively associated with case-level MDD onset across adolescence. In particular, maternal dampening of adolescent positivity, or the failure to appropriately reinforce positivity, may be associated with difficulty upregulating positive emotion, which may create vulnerability to depression, particularly in relation to the anhedonic features of this disorder (Feng et al., 2009; Silk, Shaw, Forbes, Lane, & Kovacs, 2006).

We previously found that the suppression of maternal aggression in response to adolescent dysphoria predicted MDD onset in early to midadolescence (Schwartz et al., 2011); however, this finding did not replicate in the current analyses, in which depression onset was examined into late adolescence, and thereby suggests that this association may be more proximal and may wash out over longer time frames.

The absence of significant findings relating to rates of maternal dysphoria and maternal responses to adolescent dysphoria is puzzling given the critical role of dysphoric affect in MDD. A possible explanation is that although parental socialization processes are important in the development of aggressive and positive emotion regulation abilities across the entire duration of adolescence, these family processes do not significantly shape the development of the regulation of dysphoric emotions during this time.

The primary strength of the current study is its prospective, longitudinal design, which enabled the examination of potential risk factors for MDD onset across the entire course of adolescence. In addition, the observational measurement of maternal socialization behaviors allowed for a relatively objective assessment of maternal behavior, and the observation of behavior across two affective contexts permitted the identification of context-specific family processes that predict MDD onset. Finally, our use of the LIFE coding system enabled the differentiation of aggressive and dysphoric behaviors, which are often combined and considered together as negative behaviors. Our work suggests important differences in relation to these affective behaviors and highlights the importance of considering them independently.

A limitation of this study is the inability to validly examine fathers’ behavior because of the low number of participating fathers. Fathers play a significant role in the socialization of emotion in their children, and the findings relating to mothers’ socialization behaviors may not generalize to fathers (Cassano, Adrian, Veits, & Zeman, 2006; Jacob & Johnson, 2001; Sheeber et al., 2007). Furthermore, maternal and paternal socialization behaviors may interact in important ways such that they may be additive or may buffer the effects of the other parent on child outcomes.

In addition, as a result of the low power associated with examining the onset of other disorders with lower incidence rates, only MDD outcomes have been considered in the current analyses. Hence, it is unknown whether the maternal socialization behaviors identified as prospective predictors are specific to risk for depression or whether they represent a more general vulnerability to psychopathology across adolescence.

Finally, although these findings are prospective and longitudinal, they are correlational. Prospective longitudinal designs are considered more powerful than are cross-sectional designs because they can explicitly test the directionality of effects and discount the possibility that observed relationships are solely due to state-dependent factors (Cowan & Cowan, 2002; Sheeber et al., 2001). However, the current findings could be attributed to one of three possible scenarios: (a) Parenting behaviors may influence the onset of adolescent depression, (b) adolescent characteristics at baseline may elicit parenting behaviors, or (c) an unmeasured factor may account for both parenting behaviors and the development of depression (Garber, Robinson, & Valentiner, 1997). Researchers have argued that intervention studies showing that an improvement in parenting is related to an improvement in symptoms, or the prevention of depression onset (e.g., Compas et al., 2009; Compas et al., 2010; Compas et al., 2011), offer the most convincing evidence about the directionality of influence (Cowan & Cowan, 2002). However, prospective studies are required to identify mechanisms to target in intervention studies. Future research could help to elucidate the question of causality by applying the findings of the current study to a clinical prevention trial.

In addition, the question of directionality of effects could be more comprehensively investigated by conducting multiple family-interaction assessments over time. Such assessments would also help to determine whether parenting during early adolescence continues to be associated with depression onset in late adolescence after controlling for more proximal assessments of parenting.

The importance of considering context when observing parenting behaviors has been highlighted in the current study, and in future studies, researchers should consider including an interaction task explicitly designed to elicit depressive affect. The inclusion of such a task may be particularly important in light of the current study’s failure to identify significant predictors in relation to dysphoric behavior.

Finally, the findings of this study also clearly indicate the need for more research into the potentially protective behaviors mothers could use when responding to their child’s aggression, dysphoria, and positivity because the current study has primarily identified potential risk rather than protective factors. For example, the exploration of emotion-coaching techniques may provide an adaptive alternative response that could be offered to parents in intervention research (Gottman, Katz, & Hooven, 1997; Shortt, Sheeber, Low, & Katz, 2010; Stocker, Richmond, Rhoades, & Kiang, 2007).

Footnotes

Declaration of Conflicting Interests

The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article.

Funding

This research was supported by grants from the Colonial Foundation, the National Health and Medical Research Council Australia (Program Grant 350241), and the Australian Research Council (Discovery Grant DP0878136).

Notes

References

Aldao

(2013). The future of emotion regulation research: Capturing context. Perspectives on Psychological Science, 8, 155–172.

Allison

P. D.

Liker

J. K.

(1982). Analyzing sequential categorical data on dyadic interaction: A comment on Gottman. Psychological Bulletin, 91, 393–403.

Burge

Hammen

(1991). Maternal communication: Predictors of outcome at follow-up in a sample of children at high and low risk for depression. Journal of Abnormal Psychology, 100, 174–180.

Cassano

Adrian

Veits

Zeman

(2006). The inclusion of fathers in the empirical investigation of child psychopathology: An update. Journal of Clinical Child & Adolescent Psychology, 35, 583–589.

Cassano

Perry-Parrish

Zeman

(2007). Influence of gender on parental socialization of children’s sadness regulation. Social Development, 16, 210–231.

Cole

D. A.

Rehm

L. P.

(1986). Family interaction patterns and childhood depression. Journal of Abnormal Child Psychology, 14, 297–314.

Compas

B. E.

Champion

J. E.

Forehand

Cole

D. A.

Reeslund

Fear

. . . Roberts

(2010). Coping and parenting: Mediators of 12-month outcomes of a family group cognitive-behavioral preventive intervention with families of depressed parents. Journal of Consulting and Clinical Psychology, 78, 623–634.

Compas

B. E.

Forehand

Keller

Champion

J. E.

Rakow

Reeslund

K. L.

. . . Cole

D. A.

(2009). Randomized controlled trial of a family cognitive-behavioral preventive intervention for children of depressed parents. Journal of Consulting and Clinical Psychology, 77, 1007–1020.

Compas

B. E.

Forehand

Thigpen

J. C.

Keller

Hardcastle

E. J.

Cole

D. A.

. . . Roberts

(2011). Family group cognitive-behavioral intervention for families of depressed parents: 18- and 24-month outcomes. Journal of Consulting and Clinical Psychology, 79, 488–499.

10.

Cook

W. L.

Asarnow

J. R.

Goldstein

M. J.

Marshall

V. G.

Weber

(1990). Mother-child dynamics in early-onset depression and childhood schizophrenia spectrum disorders. Development and Psychopathology, 2, 71–84.

11.

Cowan

P. A.

Cowan

C. P.

(2002). Interventions as tests of family systems theories: Marital and family relationships in children’s development and psychopathology. Development and Psychopathology, 14, 731–759.

12.

Dadds

M. R.

Sanders

M. R.

Morrison

Rebgetz

(1992). Childhood depression and conduct disorder: II. An analysis of family interaction patterns in the home. Journal of Abnormal Psychology, 101, 505–513.

13.

Davenport

Yap

M. B. H.

Simmons

J. G.

Sheeber

L. B.

Allen

N. B.

(2011). Maternal and adolescent temperament as predictors of maternal affective behavior during mother-adolescent interactions. Journal of Adolescence, 34, 829–839.

14.

Denham

S. A.

Bassett

H. H.

Wyatt

(2007). The socialization of emotional competence. In Grusec

J. E.

Hastings

P. D.

(Eds.), Handbook of socialization: Theory and research (pp. 614–637). New York, NY: Guilford Press.

15.

Eisenberg

Cumberland

Spinrad

T. L.

(1998). Parental socialization of emotion. Psychological Inquiry, 9, 241–273.

16.

Ellis

L. K.

Rothbart

M. K.

(2001, April). Revision of the Early Adolescent Temperament Questionnaire. Paper presented at the Biennial Meeting of the Society for Research in Child Development, Minneapolis, MN.

17.

Feng

Keenan

Hipwell

A. E.

Henneberger

A. K.

Rischall

M. S.

Butch

. . . Babinski

D. E.

(2009). Longitudinal associations between emotion regulation and depression in preadolescent girls: Moderation by the caregiving environment. Developmental Psychology, 45, 798–808.

18.

Fleiss

J. L.

(1981). Statistical methods for rates and proportions (2nd ed.). New York, NY: Wiley.

19.

Garber

Robinson

N. S.

Valentiner

(1997). The relation between parenting and adolescent depression: Self-worth as a mediator. Journal of Adolescent Research, 12, 12–33.

20.

Gaté

M. A.

Watkins

E. R.

Simmons

J. G.

Byrne

M. L.

Schwartz

O. S.

Whittle

. . . Allen

N. B.

(2013). Maternal parenting behaviors and adolescent depression: The mediating role of rumination. Journal of Clinical Child & Adolescent Psychology, 42, 348–357.

21.

Gilboa

Revelle

(1994). Personality and the structure of affective responses. In Van Goozen

S. H. M.

Van de Poll

N. E.

Sergeant

J. A.

(Eds.), Emotions: Essays on emotion theory (pp. 135–160). Hillsdale, NJ: Erlbaum.

22.

Goodman

S. H.

Gotlib

I. H.

(1999). Risk for psychopathology in the children of depressed mothers: A developmental model for understanding mechanisms of transmission. Psychological Review, 106, 458–490.

23.

Gottman

J. M.

Coan

Swanson

Carrere

(1998). Predicting marital happiness and stability from newlywed interactions. Journal of Marriage and Family, 60, 5–22.

24.

Gottman

J. M.

Katz

L. F.

Hooven

(1997). Meta-emotion: How families communicate emotionally. Mahwah, NJ: Erlbaum.

25.

Gottman

J. M.

Roy

A. K.

(1990). Sequential analysis: A guide for behavioral researchers. Cambridge, England: Cambridge University Press.

26.

Graham

J. W.

(2009). Missing data analysis: Making it work in the real world. Annual Review of Psychology, 60, 549–576.

27.

Halberstadt

A. G.

Crisp

V. W.

Eaton

K. L.

(1999). Family expressivity. In Philippot

Feldman

R. S.

Coats

E. J.

(Eds.), The social context of nonverbal behavior (pp. 109–155). New York, NY: Cambridge University Press.

28.

Hankin

B. L.

Abramson

L. Y.

Moffitt

T. E.

Silva

P. A.

McGee

Angell

K. E.

(1998). Development of depression from preadolescence to young adulthood: Emerging gender differences in a 10-year longitudinal study. Journal of Abnormal Psychology, 107, 128–140.

29.

Hops

Biglan

Tolman

Arthur

Longoria

(1995). LIFE: Living in Family Environments coding system manual. Eugene, Oregon: Oregon Research Institute.

30.

Jackson

Kuppens

Sheeber

L. B.

Allen

N. B.

(2011). Expression of anger in depressed adolescents: The role of the family environment. Journal of Abnormal Child Psychology, 39, 463–474.

31.

Jacob

Johnson

S. L.

(2001). Sequential interactions in the parent-child communications of depressed fathers and depressed mothers. Journal of Family Psychology, 15, 38–52.

32.

Katz

L. F.

Hunter

E. C.

(2007). Maternal meta-emotion philosophy and adolescent depressive symptomatology. Social Development, 16, 343–360.

33.

Kaufman

Birmaher

Brent

Rao

Flynn

Moreci

. . . Ryan

(1997). Schedule for Affective Disorders and Schizophrenia for School-Age Children–Present and Lifetime version (K-SADS-PL): Initial reliability and validity data. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 980–988.

34.

Kendler

K. S.

Gardner

C. O.

Prescott

C. A.

(2006). Toward a comprehensive developmental model for major depression in men. American Journal of Psychiatry, 163, 115–124.

35.

Kessler

R. C.

McGonagle

K. A.

Swartz

Blazer

D. G.

Nelson

C. B.

(1993). Sex and depression in the National Comorbidity Survey I: Lifetime prevalence, chronicity and recurrence. Journal of Affective Disorders, 29, 85–93.

36.

Lewinsohn

P. M.

Rohde

Seeley

J. R.

(1998). Major depressive disorder in older adolescents: Prevalence, risk factors and clinical implications. Clinical Psychology Review, 18, 765–794.

37.

Lovejoy

M. C.

Graczyk

P. A.

O’Hare

Neuman

(2000). Maternal depression and parenting behavior: A meta-analytic review. Clinical Psychology Review, 20, 561–592.

38.

MacPhillamy

D. J.

Lewinsohn

P. M.

(1976). Manual for the Pleasant Events Schedule. Eugene: University of Oregon.

39.

Merikangas

K. R.

J.-P.

Burstein

Swanson

S. A.

Avenevoli

Cui

. . . Swendsen

(2010). Lifetime prevalence of mental disorders in U.S. adolescents: Results from the National Comorbidity Survey Replication–Adolescent Supplement (NCS-A). Journal of the American Academy of Child and Adolescent Psychiatry, 49, 980–989.

40.

Messer

S. C.

Gross

A. M.

(1995). Childhood depression and family interaction: A naturalistic observation study. Journal of Clinical Child Psychology, 24(1), 77–88.

41.

Morris

A. S.

Silk

J. S.

Steinberg

Myers

S. S.

Robinson

L. R.

(2007). The role of family context in the development of emotion regulation. Social Development, 16, 361–388.

42.

Muthén

L. K.

Muthén

B. O.

(1998–2012). Mplus user’s guide (7th ed.). Los Angeles, CA: Muthén & Muthén.

43.

Painuly

Sharan

Mattoo

S. K.

(2005). Relationship of anger and anger attacks with depression: A brief review. European Archives of Psychiatry and Clinical Neuroscience, 255, 215–222.

44.

Prinz

R. J.

Foster

S. L.

Kent

R. N.

O’Leary

K. D.

(1979). Multivariate assessment of conflict in distressed and nondistressed mother-adolescent dyads. Journal of Applied Behavior Analysis, 12, 691–700.

45.

Radloff

L. S.

(1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401.

46.

Restifo

K. K.

Bögels

(2009). Family processes in the development of youth depression: Translating the evidence to treatment. Clinical Psychology Review, 29, 294–316.

47.

Safer

D. J.

(2009). Irritable mood and the Diagnostic and Statistical Manual of Mental Disorders. Child and Adolescent Psychiatry and Mental Health, 3, 35.

48.

Schwartz

O. S.

Dudgeon

Sheeber

L. B.

Yap

M. B. H.

Simmons

J. G.

Allen

N. B.

(2011). Observed maternal responses to adolescent behavior predict the onset of major depression. Behaviour Research and Therapy, 49, 331–338.

49.

Schwartz

O. S.

Dudgeon

Sheeber

L. B.

Yap

M. B. H.

Simmons

J. G.

Allen

N. B.

(2012). Parental behaviors during family interactions predict changes in depression and anxiety during adolescence. Journal of Abnormal Child Psychology, 40, 59–71.

50.

Seeley

J. R.

Lewinsohn

P. M.

(2009). Epidemiology of mood disorders during adolescence: Implications for lifetime risk. In Allen

N. B.

Sheeber

L. B.

(Eds.), Adolescent emotional development and the emergence of depressive disorders (pp. 33–55). Cambridge, England: Cambridge University Press.

51.

Sheeber

Allen

Davis

Sorensen

(2000). Regulation of negative affect during mother-child problem-solving interactions: Adolescent depressive status and family processes. Journal of Abnormal Child Psychology, 28, 467–479.

52.

Sheeber

Davis

Leve

Hops

Tildesley

(2007). Adolescents’ relationships with their mothers and fathers: Associations with depressive disorder and subdiagnostic symptomatology. Journal of Abnormal Psychology, 116, 114–154.

53.

Sheeber

Hops

Alpert

Davis

Andrews

(1997). Family support and conflict: Prospective relations to adolescent depression. Journal of Abnormal Child Psychology, 25, 333–344.

54.

Sheeber

Hops

Andrews

Alpert

Davis

(1998). Interactional processes in families with depressed and non-depressed adolescents: Reinforcement of depressive behavior. Behaviour Research and Therapy, 36, 417–427.

55.

Sheeber

Hops

Davis

(2001). Family processes in adolescent depression. Clinical Child and Family Psychology Review, 4, 19–35.

56.

Sheeber

Sorensen

(1998). Family relationships of depressed adolescents: A multimethod assessment. Journal of Clinical Child Psychology, 27, 268–277.

57.

Shortt

J. W.

Sheeber

Low

Katz

L. F.

(2010, November). Emotion coaching as a unique predictor of adolescent internalizing problems. Paper presented at the Annual Convention of the Association for Behavioral and Cognitive Therapies, San Fransisco, CA.

58.

Silk

J. S.

Shaw

D. S.

Forbes

E. E.

Lane

T. L.

Kovacs

(2006). Maternal depression and child internalizing: The moderating role of child emotion regulation. Journal of Clinical Child & Adolescent Psychology, 35, 116–126.

59.

Stocker

C. M.

Richmond

M. K.

Rhoades

G. K.

Kiang

(2007). Family emotional processes and adolescents’ adjustment. Social Development, 16, 310–325.

60.

Weich

Patterson

Shaw

Stewart-Brown

(2009). Family relationships in childhood and common psychiatric disorders in later life: Systematic review of prospective studies. British Journal of Psychiatry, 194, 392–398.

61.

Whittle

Yap

M. B. H.

Sheeber

Dudgeon

Yucel

Pantelis

. . . Allen

N. B.

(2011). Hippocampal volume and sensitivity to maternal aggressive behavior: A prospective study of adolescent depressive symptoms. Development and Psychopathology, 23, 115–129.

62.

Whittle

Yap

M. B. H.

Yucel

Fornito

Simmons

J. G.

Barrett

. . . Allen

N. B.

(2008). Prefrontal and amygdala volumes are related to adolesents’ affective behaviors during parent-adolescent interactions. Proceedings of the National Academy of Sciences, USA, 105, 3652–3657.

63.

Whittle

Yap

M. B. H.

Yucel

Sheeber

Simmons

J. G.

Pantelis

Allen

N. B.

(2009). Maternal socialisation of adolescent positive affect is associated wtih adolescent prefrontal and amygdala volumes. Social Cognitive and Affective Neuroscience, 4, 247–256.

64.

Wothke

(2000). Longitudinal and multi-group modeling with missing data. In Little

T. D.

Schnabel

K. U.

Baumert

(Eds.), Modeling longitudinal and multilevel data: Practical issues, applied approaches, and specific examples (pp. 197–216). Mahwah, NJ: Erlbaum.

65.

Yap

M. B. H.

Allen

N. B.

Ladouceur

(2008). Maternal socialization of positive affect: The impact of invalidation on adolescent emotion regulation and depressive symptomatology. Child Development, 79, 1415–1431.

66.

Yap

M. B. H.

Allen

N. B.

Leve

Fainsilber-Katz

(2008). Maternal meta-emotion philosophy and socialization of adolescent affect: The moderating role of adolescent temperament. Journal of Family Psychology, 22, 688–700.

67.

Yap

M. B. H.

Allen

N. B.

O’Shea

Di Parsia

Simmons

J. G.

Sheeber

(2011). Early adolescents’ temperament, emotion regulation during mother-child interactions and depressive symptomatology. Development and Psychopathology, 23, 267–282.

68.

Yap

M. B. H.

Allen

N. B.

Sheeber

(2007). Using an emotion regulation framework to understand the role of temperament and family processes in risk for adolescent depressive disorders. Clinical Child and Family Psychology Review, 10, 180–196.

69.

Yap

M. B. H.

Schwartz

O. S.

Byrne

M. L.

Simmons

J. G.

Allen

N. B.

(2010). Maternal positive and negative interaction behaviors and early adolescents’ depressive symptoms: Adolescent emotion regulation as a mediator. Journal of Research on Adolescence, 20, 1014–1043.

70.

Yap

M. B. H.

Whittle

Yucel

Sheeber

Pantelis

Simmons

J. G.

Allen

N. B.

(2008). Interaction of parenting experiences and brain structure in the prediction of depressive symptoms in adolescents. Archives of General Psychiatry, 65, 1377–1385.