Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Apr 16.
Published in final edited form as: J Fam Psychol. 2020 Apr 23;34(7):836–845. doi: 10.1037/fam0000653

Reciprocal Effects of Maternal and Child Internalizing Symptoms Before and After a Natural Disaster

Estee M Hausman 1, Sarah R Black 2, Evelyn Bromet 3, Gabrielle Carlson 3, Allison Danzig 4, Roman Kotov 3, Daniel N Klein 1
PMCID: PMC9013012  NIHMSID: NIHMS1793130  PMID: 32324018

Abstract

Objective.

After natural disasters, mothers and children are vulnerable to internalizing symptoms, such as depression and anxiety, and levels of mothers’ and children’s symptoms are significantly associated. However, the disaster literature has rarely examined reciprocal effects within families. The present study capitalizes on the occurrence of Hurricane Sandy during the course of an ongoing longitudinal study to address this gap.

Method.

Three hundred and forty-seven children (54.2% male, 84.7% Caucasian) and their mothers completed measures of internalizing symptoms when the children were 9 years old. Hurricane Sandy occurred an average of 1 year later. Eight weeks after the hurricane, mothers and children completed the same measures again. Mothers also reported on their family’s stress exposure from Hurricane Sandy.

Results.

After controlling for pre-disaster symptoms, longitudinal actor-partner interdependence models indicated that mother’s and children’s internalizing symptoms were linked. Mothers’ pre-hurricane depression symptoms also predicted increases in children’s depression symptoms over time independent of hurricane-related stress. Children’s pre-hurricane anxiety symptoms predicted increases in mothers’ depression symptoms only at low levels of hurricane-related stress.

Conclusions.

Rather than the emergence of reciprocal effects, mother’s depression symptoms and children’s internalizing symptoms changed in tandem after Hurricane Sandy. High levels of Hurricane Sandy stress did not produce symptom spillover effects, but rather may have interrupted the unfolding of normative developmental parent-child reciprocal symptom processes.

Keywords: natural disasters, disasters, reciprocal effects, bidirectional effects, parents, children

Natural disasters are relatively common, sudden events that impact families and entire communities (Bonanno, Brewin, Kaniasty, & La Greca, 2010). Although most individuals show resilience, adults and children exposed to natural disasters can experience a host of psychological symptoms, including posttraumatic stress, depression, and anxiety (Bonanno et al., 2010; Kopala-Sibley et al., 2016a,b). Several demographic variables (younger age, female gender, racial/ethnic minority status), resource factors (lower socioeconomic status, lower social support), and disaster-related factors (higher degree of exposure, greater proximity to the disaster, higher number of deaths) have been shown to predict post-disaster outcomes in children, adults, and families (Bonanno et al., 2010; Pfefferbaum, Houston, North, & Regens, 2008).

Research has also demonstrated that parents, especially mothers, and children are particularly vulnerable to adverse outcomes after natural disasters. Indeed, parents exhibit higher levels of psychological symptoms than non-parents following natural disasters, which may be due to increased caregiving burden and concerns for their children’s health and well-being (Bonanno et al., 2010; Norris et al., 2002a; Pfefferbaum & North, 2008). In addition, children often experience higher levels of psychological symptoms and more severe impairment than adults, particularly in the presence of parental distress and psychopathology (Bonanno et al., 2010; Masten & Narayan, 2010; Norris et al., 2002a), although some exceptions have been noted (e.g., Bromet et al., 2000). Post-disaster parental distress is directly distressing to children, who rely on their parents for emotional security and support (Banks & Weems, 2014). Moreover, parental distress negatively impacts parenting behaviors and the parent-child relationship, which also influences children’s post-disaster functioning, (Bonanno et al., 2010; Costa, Weems, & Pina, 2009; Masten & Narayan, 2010; Norris et al., 2002a; Pfefferbaum & North, 2008).

Consistent with social ecological and family systems models (Bronfenbrenner & Morris, 2006; Minuchin, 1985; Weems & Overstreet, 2008), there may be bidirectional effects of parent and child psychological symptoms on each other’s post-disaster outcomes. However, little is known about how children’s symptoms influence their parents after natural disasters. Investigation of the reciprocal relations of parent and child psychological symptoms following natural disasters may enhance our understanding of heterogeneity in responses to natural disasters in parent-child dyads as well as distress contagion processes within families following natural disasters.

Reciprocal effects within families have been well-documented in non-disaster community samples. Classic developmental models suggested that parents influence their children’s behavior and psychological functioning, and that children also influence their environments, including their parents’ behavior and psychological functioning (Bell, 1968; Patterson, 1982; Sameroff, 1975). Most longitudinal studies have focused on maternal depression symptoms, which robustly predict children’s later internalizing symptoms (anxiety and depression symptoms; Goodman & Gotlib, 1999), particularly in girls (Goodman et al., 2011). At the same time, youth’s internalizing problems also predict their mother’s later depression symptoms and disorders (Allmann, Kopala-Sibley, & Klein, 2016; Hughes & Gullone, 2010; Raposa et al., 2011; Sellers et al., 2016), with one study finding this association only in mother-daughter dyads (Sellers et al., 2016). A number of mechanisms have been identified for these reciprocal relationships, including shared genetics, family functioning, marital distress, parent-child relationship quality, and parenting behaviors, efficacy, and distress (Goodman & Gotlib, 1999; Hughes & Gullone, 2010). Further, intervention studies have demonstrated that parental treatment improves children’s symptoms and children’s treatment reduces maternal symptoms (Cuijpers, Weisz, Keyotaki, Garber, & Andersson, 2015; Goodman & Garber, 2017; Neill, Weems, & Sheeringa, 2018; Perloe, Esposito-Smythers, Curby, & Renshaw, 2014).

In disaster-exposed samples, most work has focused on the effects of parent psychological symptoms on poorer outcomes in children (see Bonanno et al., 2010; Masten & Narayan, 2010; Norris et al., 2002a for reviews), as parents are an important source of coping assistance and emotional security for children (Banks & Weems, 2014; Prinstein, La Greca, Vernberg, & Silverman, 1996). Parents’ distress and symptoms associated with experiencing a disaster may also be compounded by their children’s distress and symptoms. Indeed, a few studies of human disasters have suggested that children’s psychological symptoms are also distressing to their parents. For example, parent perceptions of their children’s distress level after the 9/11 terrorist attacks were associated with parents’ own distress immediately after the attacks and a year later (Phillips, Featherman, & Liu, 2004). In another study, after their children survived a shipping disaster, mothers whose children had met full criteria for Posttraumatic Stress Disorder had higher posttraumatic stress symptoms than mothers’ whose children did not (Mirzamani & Bolton, 2002). Similarly, Koplewicz and colleagues (2002) found that children’s posttraumatic stress symptoms predicted their parents’ posttraumatic stress symptoms three months as well as six years after children were exposed to the 1993 World Trade Center bombing. These studies have made important contributions to the literature, but have been limited by small sample sizes and/or cross-sectional designs, examining only the unidirectional effects of children on their parents, and parents and children differing from each other on physical proximity and exposure to the disaster.

Two existing studies have examined reciprocal relationships of parent-youth psychological symptoms following natural disasters, but reported conflicting findings (Juth, Silver, Seyle, Widyatmoko, & Tan, 2015; Shi et al., 2018). Using a longitudinal actor-partner interdependence (APIM) model (Cook & Kenny, 2005), Shi and colleagues (2018) found reciprocal effects of posttraumatic stress symptoms in mothers and adolescents assessed 12 and 18 months after the 2008 Wenchung earthquake. In contrast, reciprocal associations did not emerge in Juth et al.’s (2015) study of parent and child symptoms and distress levels at a single time-point 3 years after the 2006 Indonesian earthquake. Like many disaster studies, given their unpredictable nature (Masten & Narayan, 2010; Pfefferbaum et al., 2013), neither study assessed parent-youth dyads prior to the natural disaster. This is of critical importance for establishing the direction of effects because individuals exposed to natural disasters may have prior internalizing symptoms, which has been shown to robustly predict post-disaster outcomes (e.g., Bonanno et al., 2010; Norris et al., 2002a; Weems et al., 2007). In addition, reciprocal effects may be strengthened when accounting for prior symptoms. For example, exposure to a natural disaster may exacerbate the stress associated with parenting a child who is already experiencing internalizing problems (Raposa et al., 2011). In addition, given dose-response relationships of disaster-related stress exposure in children and adults (Masten & Narayan, 2010; Pfefferbaum et al., 2013), degree of exposure to the natural disaster may impact the presence or the magnitude of reciprocal effects, but this has not yet been explored empirically.

The present study examined reciprocal effects of maternal depression and child internalizing (anxiety and depression) symptoms using longitudinal data obtained an average of 1 year prior to, and 8 weeks after, Hurricane Sandy. We focus on the mother-child relationship within families because mother’s symptoms have been shown to have a pronounced effect on family-level post-disaster distress (e.g., McFarlane, 1987a) and children’s functioning (e.g., Shi et al., 2018). Hurricane Sandy hit the tri-state area (New York, New Jersey, and Connecticut) in October 2012; it was the most destructive hurricane in the region in many years and one of the costliest in United States history, destroying over 100,000 homes (Neria & Shultz, 2012). Consistent with Shi et al. (2018), we expected that reciprocal (bidirectional) effects of maternal depression and child internalizing symptoms would emerge, after accounting for pre-hurricane symptom severity. Because the number and severity of disaster-related exposures are associated with greater psychological symptoms in children and adults (Masten & Narayan, 2010; Pfefferbaum et al., 2013), we conducted exploratory analyses examining whether reciprocal effects of maternal depression and child internalizing symptoms were intensified in families experiencing higher levels of hurricane-related stress. We also conducted exploratory analyses to examine whether reciprocal effects varied by child gender, given that girls are more influenced by their mothers’ depression symptoms than boys in community samples (Goodman et al., 2011), and that women and girls often experience higher levels of post-disaster psychopathology than men and boys (Bonanno et al., 2010), although the literature is somewhat mixed (Pfefferbaum et al., 2008). All analyses adjusted for both mother and child internalizing symptoms prior to Hurricane Sandy to account for any effects of pre-Sandy symptoms on post-Sandy symptoms.

Method

Participants

The sample included 347 children (M age = 9.14 years, SD = 0.32 at the assessment prior to the hurricane; M age = 10.24 years, SD = 0.77 at the time of the hurricane) and their mothers (M age = 41.68 years, SD = 4.46 at the pre-hurricane assessment) derived from a larger longitudinal study examining early antecedents of depressive and anxiety disorders (see Klein & Finsaas, 2017 for details). Children and their caregivers were initially recruited from the community using commercial mailing lists when children were 3 years old. Children were eligible for the study if they did not have a significant medical condition or developmental disability and were living with at least one English-speaking biological parent.

Of the larger sample, 446 families had completed age 9 assessments prior to Hurricane Sandy; these families were invited to participate in the post-Sandy assessment. Three-hundred sixty-two families (81.2%) agreed to participate, but 15 families were excluded who were not in the region when Hurricane Sandy occurred, leaving a final sample of 347 children and their mothers. All families in the current study lived in areas declared disaster zones by the Federal Emergency Management Agency (FEMA). Children in the present sample were primarily Caucasian (84.7%) and 54.2% were male. Families had an average annual income of approximately $100,000 prior to the hurricane. Most mothers were married (>85%) at both the pre- and post-Sandy assessments, and 59.2% had completed a 4-year college degree.

Children and mothers who participated in the current study did not differ from the larger sample of 446 families on the measures included in this report (p’s > .10), except for pre-Sandy child anxiety scores. Specifically, children who participated in the current study had lower pre-Sandy anxiety symptoms (M = 18.82, SD = 10.84) than did those who did not participate (M = 22.59, SD = 12.48), t(423) = 2.69, p < .01, d = 0.32.

Procedure

During the age 9 assessment, children completed self-report measures of their depression and anxiety symptoms and mothers completed a self-report measure of their own depression symptoms. Six weeks after Hurricane Sandy, mothers were contacted and asked to complete a measure of hurricane-related stress exposure and their own depression symptoms. Children again completed measures of depression and anxious symptoms. Surveys were completed an average of 8.44 weeks (SD = 1.51) after the hurricane and 62.26 weeks (SD = 31.93) after the age 9 assessment. This study was approved by the Stony Brook Institutional Review Board. Written informed consent and assent were obtained from all youth and their guardians prior to any study procedures.

Measures

Hurricane Sandy Stress.

Mothers were asked to complete the Hurricane Sandy Stress Exposure Inventory (HSSEI; Kopala-Sibley et al., 2016a,b), a 13-item questionnaire concerning the effects of Hurricane Sandy on their family (see Table 1 for details). Items were drawn from previous questionnaires developed for Hurricane Ike and Hurricane Katrina, which have demonstrated adequate internal consistency and significant associations with measures of PTSD symptoms (Galea et al., 2007; Norris, Sherrieb, & Galea, 2010). Items 1–8 were rated on a 5-point scale ranging from 1 (not at all affected) to 5 (extremely affected), Items 9–10 were rated based on duration on a 5-point scale ranging from 1 (0 days) to 5 (2 weeks or more), and Items 11–13 were rated as present or absent. To create an overall exposure severity score, nonbinary items were rescaled such that 0 = absent and 1 = present. We selected cutoffs for each item using a combination of statistical and clinical considerations that distinguished a subgroup of participants with a high and clinically significant level of stress exposure on that item. This involved examining the distribution of responses on each item and considering the nature of the stressor and the response options (see Kopala-Sibley at el., 2016a,b for details). For most items, exposure was defined as ratings of 4 (affected quite a bit) or 5 (extremely affected). The single exception was the item on financial hardships. Responses to this item were extremely skewed, which may, in part, have been due to the phrasing of the item (“hardships”), suggesting a high threshold for endorsement. Hence, we selected a cutoff of 3 (moderately) for this item. Total scores ranged from 0–13. This scale showed adequate internal consistency (a = .72) in the present sample.

Table 1.

Frequency of Family Exposure to Hurricane Sandy Stressorsa

Item N (%) of participants endorsing items
 1. Damage to home or possessions 40 (11.6%)
 2. Own or family’s safety threatened 44 (12.8%)
 3. Financial hardships 43 (12.5%)
 4. Children fear for their own or their family’s safety 79 (22.9%)
 5. Life disrupted by Hurricane Sandy 82 (23.8%)
 6. Difficulty finding gasoline 56 (16.3%)
 7. Difficulty getting food, water, or warmth 69 (20.0%)
 8. Children quarreling or complaining more than usual 54 (15.7%)
 9. Length of time without power 124 (35.9%)
 10. Length of time children’s school closed 169 (49.1%)
 11. Self, friends, family, pets or co-workers robbed, injured or got lost 11 (3.2%)
 12. Apply to FEMA, government aid, or Red Cross or other aid agency 14 (4.0%)
 13. Evacuated home 16 (4.6%)
Average number of stressors reported 2.32
Open in a new tab

Note: FEMA = Federal Emergency Management Agency.

a

Item scores were dichotomized as present at a significant level or not.

Child Internalizing Symptoms.

The Children’s Depression Inventory (CDI; Kovacs, 1985) was used to assess children’s depression symptoms occurring in the past two weeks. The CDI is a widely used 27-item self-report measure of depression symptoms in children ages 7–18. Items are rated on a 3-point scale ranging from 0–2. Coefficient alphas for the CDI were 0.74 in the full sample at pre-Sandy and 0.87 in the post-Sandy sample.

The Screen for Child Anxiety-Related Emotional Disorders (SCARED-SR; Birmaher, Chiappetta, Bridge, Monga, & Baugher, 1999) was used to assess children’s anxiety symptoms occurring over the past three months. It is a 41-item self-report measure of anxiety symptom severity in youth ages 8–18. Items are rated on a 3-point scale ranging from 0–2. Coefficient alpha for the SCARED was 0.89 in the full sample at pre-Sandy. In the post-Sandy assessment, children were asked to rate symptoms in the period since the hurricane to prevent assessing anxiety symptoms prior to hurricane exposure; coefficient alpha was 0.93 in the post-Sandy sample.

Mother Depression Symptoms.

Mothers’ depression symptoms were assessed with the Diagnostic Inventory for Depression (DID; Zimmerman et al., 2004), a 22-item measure designed to assess Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV) criterion symptoms of major depressive disorder (MDD) occurring in the past week. Items are rated on a 5-point scale ranging from 0–4. The DID correlates well with diagnoses of MDD based on structured diagnostic interviews, discriminates MDD from other disorders, and is sensitive to change (Zimmerman et al., 2004). In the current study, coefficient alphas for the DID were 0.90 in the full sample at pre-Sandy and 0.91 in the post-Sandy sample.

Data Analyses

Analyses were conducted in Mplus version 7.0 (Muthen & Muthen, 2012). Longitudinal actor-partner interdependence models (APIM; Cook & Kenny, 2005) examined reciprocal effects of child internalizing symptoms and mother depression symptoms from pre- to post-Sandy. Longitudinal APIMs examine both actor and partner effects among members of a dyad using path analysis or structural equation modeling. Actor effects represent the stability of each individual’s symptoms, whereas partner effects are the effects of one individual’s symptoms on the other’s later symptoms. The cross-sectional correlations among individuals’ symptoms are also modeled to reflect the interdependence of symptoms within a dyad at a given time.

We tested child anxiety symptoms and child depression symptoms in separate models because we were concerned about multicollinearity among these variables based on the high co-occurrence of these symptoms in children (e.g., Cummings, Caporino, & Kendall, 2014). All models utilized full-information maximum likelihood, in order to take missing data into account, and the MLR estimator, because some predictor variables were highly skewed. Model fit indices, including chi-square test of model fit, root mean square error of approximation (RMSEA), the comparative fit index (CFI), and the Tucker-Lewis index (TLI), were examined. Following Hu and Bentler (1999), a RMSEA of less than .06 and CFI and TLI values greater than .95 were taken to indicate good model fit. All models adjusted for the child’s gender, weeks between the age 9 and post-Sandy assessments, and family income. Child age at the time of the age 9 assessment was not included as a covariate because it was so highly correlated with number of weeks between the age 9 and post-Sandy assessments (r = .89, p < .001), and the latter variable had a larger range and variance. For exploratory purposes, we also examined two-way interactions of pre-Sandy mother and child symptoms with Hurricane Sandy stress and child gender. Continuous predictors were centered prior to creating interaction terms (Aiken & West, 1991). Interactions were interpreted by comparing simple slopes at high and low levels (±1 SD) of the moderator variable as well as using the Johnson-Neyman (Johnson & Fay, 1950) technique to better examine the pattern of moderation. Non-significant interactions were dropped from the models.

Results

Descriptive Statistics and Bivariate Correlations

Rates of Hurricane Sandy-related stressors are presented in Table 1. The most highly endorsed stressors included extended school closure (49.1%), extended power outages (35.9%), life being disrupted by the hurricane (22.9%), children fearing for their own or their family’s safety (23.8%), and difficulty obtaining food, water, or warmth (20.0%). Families experienced a mean 2.32 (SD = 2.25) of the 13 stressors.

Bivariate correlations and descriptive statistics for hurricane stress exposure and all preSandy variables are presented in Table 21. At pre-Sandy, child depression and anxiety symptoms were positively correlated, but both were unrelated to pre-Sandy maternal depression symptoms. Female child gender was associated with higher pre-Sandy child anxiety symptoms, but not child depressive symptoms. Higher maternal depressive symptoms at pre-Sandy were also associated with female child gender and fewer weeks between the pre- and post-Sandy assessments. Finally, lower family income was associated with higher pre-Sandy maternal depression and child anxiety and depression symptoms. Importantly, pre-Sandy child and maternal internalizing symptoms were unrelated to later reports of Hurricane Sandy stress.

Table 2.

Descriptive Statistics and Bivariate Correlations of Hurricane Stress and Age 9 Symptoms and Demographic Characteristics

Variable 1. 2. 3. 4. 5. 6. 7.
1. Pre-Sandy Child Depression --
2. Pre-Sandy Child Anxiety .38*** --
3. Pre-Sandy Mother Depression .01 .01 --
4. Hurricane Sandy Stress −.03 −.04 .09 --
5. Child Gender (0 = Male; 1 = Female) −.05 .11* .11* −.04 --
6. Time between pre- and post-Sandy assessments −.03 −.11* −.05 −.07 −.05 --
7. Family Income −.22*** −.14* −.17** −.02 −.08 .02 --
M 4.62 18.82 3.88 2.32 0.46 62.26 5.80
SD 3.97 10.84 5.89 2.25 0.50 5.80 2.11
Open in a new tab
*

p < .05

**

p < .01

***

p < .001.

Actor-Partner Interdependence Models

Fit indices for the initial child depression and mother depression model, X2(25) = 294.44, p < .001, CFI = 1.00, TLI = 1.00, RMSEA < .001, and the initial child anxiety and mother depression model, X2(25) = 275.91, p < .001, CFI = 1.00, TLI = 1.00, RMSEA < .001, both suggested excellent fit. Two-way interactions of child gender with Hurricane Sandy Stress were non-significant and were therefore removed from the final models.

The final actor-partner interdependence model for child and mother depression, without gender interactions, again demonstrated excellent fit, X2(17) = 281.80, p < .001, CFI = 1.00, TLI = 1.00, RMSEA < .001, and was preferred to the initial child depression model for its greater parsimony. Standardized parameter estimates for the final model are presented in Figure 1 and Table 3. Lower family income was associated with higher pre-Sandy mother (r = −.18, p < .01) and child depression symptoms (r = −.19, p < .01). Having a daughter was also associated with higher pre-hurricane maternal depression symptoms (r = .11, p < .05). Pre-Sandy demographic characteristics were unrelated to post-Sandy mother and child symptoms. Both actor effects were significant and positive, indicating that child and mother depression symptoms were relatively stable pre- to post-Sandy. Of the partner effects, the effect of pre-hurricane maternal depression on post-hurricane child depression was significant and positive. The effect of pre-Sandy child depression on post-Sandy mother depression was nonsignificant, but this was qualified by a significant interaction of pre-hurricane child depression symptoms and Hurricane Sandy stress on post-Sandy mother depression. Inspection of this interaction indicated that the partner effect of pre-Sandy child depression on post-Sandy mother depression was not significant at low (β = .07, p = .10) or high levels of Hurricane Sandy stress (β = −.08, p =.10; see Figure 2, top). There were no significant regions in the Johnson-Neyman test. The non-significant simple slopes in the presence of the significant interaction suggests that the direction of the relationship of pre-Sandy child depression to post-Sandy mother depression differed significantly at high and low levels of Hurricane Sandy stress. Although mother and child depression symptoms were uncorrelated before Sandy, the residual variances of mother and child depression symptoms were positively correlated after the hurricane, as shown in the Figure 1. There were no main effects of Hurricane Sandy stress on post-Sandy mother or child depression symptoms. The full model explained 17.4% and 53.6% of the variance in post-Sandy child and mother depression symptoms, respectively.

Figure 1.

Figure 1.

Open in a new tab

Longitudinal actor-partner interdependence model of pre- and post-Sandy child and mother depression symptoms. Standardized parameter estimates are provided. Pre-Sandy covariates (child sex, time between the age 9 and post-Sandy assessments, and family income) are not shown in this figure, but were included in the analysis.

*p < .05, p** < .01, p*** < .001.

Table 3.

Actor-Partner Interdependence Model Predicting Post-Sandy Child and Mother Depression Symptoms

Child Depression Symptoms Maternal Depression Symptoms
Predictor β SE P β SE P
Intercept 1.03 0.18 1.03 0.16
Child Gender (0 = Male; 1 = Female) −0.06 0.05 .26 −0.03 0.04 .50
Time Between Pre- and Post-Sandy Assessments 0.02 0.05 .70 0.01 0.04 .88
Family Income −0.07 0.06 .21 −0.05 0.04 .15
Hurricane Stress 0.04 0.06 .54 0.06 0.04 .15
Pre-Sandy Child Depression Symptoms 0.37 a 0.06 <.0001 −0.01b 0.04 .84
Pre-Sandy Maternal Depression Symptoms 0.13 b 0.05 0.006 .71 a 0.05 <.0001
Interaction of Hurricane Stress and Pre-Sandy Child Anxiety Symptoms 0.06 0.07 .38 −.08 0.03 .009
Interaction of Hurricane Stress and Pre-Sandy Maternal Depression Symptoms −0.04 0.07 .51 0.01 0.07 .87
Open in a new tab

Note. ...

a

actor effect

b

partner effect

*

p < .05

**

p < .01

***

p < .001.

Figure 2.

Figure 2.

Open in a new tab

Relationship of pre-Sandy (age 9) child internalizing symptoms and post-Sandy mother depression symptoms at high and low levels of Hurricane Sandy stress. Results are shown for the effects of pre-Sandy child depression (top) and anxiety symptoms (bottom). Only the slope at low Hurricane Sandy stress in the child anxiety symptoms model is significant.

The final actor-partner interdependence model for child anxiety and mother depression symptoms, without gender interactions, also demonstrated excellent fit, X2(17) = 275.01, p < .001, CFI = 1.00, TLI = 1.00, RMSEA < .001, and was again preferred to the initial child anxiety model for its parsimony. Standardized parameter estimates for the final model are presented in Figure 3 and Table 4. Fewer weeks between the age 9 and Hurricane Sandy assessments (r = −.11, p < .05), lower family income (r = −.11, p = .05), and being female (r = .11, p < .05) were associated with higher pre-hurricane child anxiety symptoms. Lower family income (r = −.18, p < .01) and having a daughter (r = .11, p < .05) were associated with higher pre-Sandy mother depression. Pre-hurricane demographic characteristics were again unrelated to post-hurricane mother and child symptoms. Both actor effects were significant and positive from before to after the hurricane, indicating that child anxiety and mother depression symptoms were relatively stable pre- to post-Sandy. Both partner effects were nonsignificant, but Hurricane Sandy stress significantly moderated the effect of pre-Sandy child anxiety symptoms on post-Sandy mother depression symptoms. As shown in Figure 2 (bottom), pre-Sandy child anxiety significantly and positively predicted post-Sandy mother depression at low (β = .15, p < .05), but not at high (β = −.08, p =.14) levels of stress. Although uncorrelated before the hurricane, the residual variances of child anxiety and mother depression were positively correlated after the hurricane. There were no main effects of Hurricane Sandy stress on post-Sandy mother depression symptoms or child anxiety symptoms. The full model explained 18.1% and 54.9% of the variance in post-hurricane child anxiety symptoms and maternal depression symptoms, respectively.

Figure 3.

Figure 3.

Open in a new tab

Longitudinal actor-partner interdependence model of pre- and post-Sandy child anxiety and mother depression symptoms. Standardized parameter estimates are provided. Pre-Sandy covariates (child sex, time between the age 9 and post-Sandy assessments, and family income) are not shown in this figure, but were included in the analysis.

*p < .05, p** < .01, p*** < .001.

Table 4.

Actor-Partner Interdependence Model Predicting Post-Sandy Child Anxiety and Mother Depression Symptoms

Child Anxiety Symptoms Maternal Depression Symptoms
Predictor β SE P β SE P
Intercept 1.30 0.19 0.99 0.17
Child Gender (0 = Male; 1 = Female) 0.05 0.05 .55 −0.02 0.04 .52
Time Between Pre- and Post-Sandy Assessments 0.02 0.05 .70 0.02 0.04 .61
Family Income −0.04 0.06 0.50 −0.05 0.04 0.18
Hurricane Stress 0.03 0.6 .55 0.07 0.04 0.07
Pre-Sandy Child Anxiety Symptoms 0.35 a 0.05 <.0001 0.02b 0.04 0.69
Pre-Sandy Maternal Depression Symptoms 0.12b 0.06 0.06 0.70 a 0.05 <.0001
Interaction of Hurricane Stress and Pre-Sandy Child Anxiety Symptoms −0.05a 0.08 .51 −0.12 b 0.04 0.008
Interaction of Hurricane Stress and Pre-Sandy Maternal Depression Symptoms 0.15b 0.09 0.08 0.04a 0.08 0.57
Open in a new tab

Note. ...

a

actor effect

b

partner effect

*

p < .05

**

p < .01

***

p < .001.

Discussion

Despite the prominence of developmental and ecological models of mother-child interactions, little work has considered that children’s symptoms may influence their mothers’ functioning following natural disasters. The few studies (Juth et al., 2015; Shi et al., 2018) that have examined bidirectional influences have been limited to post-disaster assessments only, and hence could not examine pre-post disaster symptom change. We examined reciprocal effects of internalizing symptoms in mother-child dyads exposed to Hurricane Sandy, accounting for their pre-disaster symptoms. We also explored whether Hurricane Sandy stress exposure or child gender influenced the presence or magnitude of these effects.

Consistent with prior literature (Norris et al., 2002a), pre-Sandy internalizing symptoms predicted higher post-Sandy symptoms in mothers and children. More importantly, maternal and child symptoms were clearly linked, mirroring findings in non-disaster exposed mother-adolescent dyads over time (Hughes & Gullone, 2010). Modest correlations among mothers’ and children’s internalizing symptoms after the disaster suggest their symptoms were linked to some degree in response to the disaster, particularly given that symptoms were uncorrelated at pre-Sandy. We also found little evidence of reciprocal mother and child symptoms effects from before to after Hurricane Sandy. Pre-Sandy maternal depression symptoms predicted higher post-Sandy child depression symptoms, but this was not the case for post-Sandy child anxiety symptoms. Although children’s internalizing symptoms did not have main effects on post-disaster maternal depression, each of these paths was significantly moderated by degree of hurricane stress exposure. Surprisingly, however, pre-hurricane child anxiety symptoms predicted greater post-hurricane mother depression symptoms only at lower levels of Sandy-related stress; pre-hurricane child anxiety and post-Sandy maternal depression were not associated at higher levels of stress. Hurricane-related stress also moderated the effect of pre-Sandy child depression symptoms on post-Sandy mother depression symptoms. At lower levels of stress, pre-hurricane child depression predicted somewhat greater post-hurricane mother depression, whereas at higher levels of stress, pre-Sandy child depression predicted somewhat lower post-Sandy maternal depression. However, when plotted at 1 SD above and below the mean, neither of these simple slopes was significant on its own. The Johnson-Neyman test also indicated that there were no significant regions. The significant interaction therefore reflected the differing directions of the effects. Finally, findings did not differ for boys and girls, as there were no significant interactions with child gender.

Mother’s depression symptoms predicted children’s depressive symptoms, regardless of their degree of disaster exposure. This finding is consistent with a large body of research that has documented that mothers’ depression symptoms influence their children’s later functioning in community samples and samples exposed to natural and human disasters (Bonanno et al., 2010; Goodman et al., 2011). Interestingly, similar mechanisms (e.g., parenting, marital distress, parent-child relationship quality; Bonanno et al., 2010; Goodman et al., 2011; Masten & Narayan, 2010; Norris et al., 2002a) have been observed in both disaster-exposed and nonexposed community samples. Therefore, it is possible that the negative influence of mothers’ symptoms on children’s symptoms represents a process of intergenerational transmission that was not affected by natural disaster exposure. Alternatively, natural disaster exposure may strengthen this process for families exposed to disasters. Children typically look to their parents for models of coping and support following stressful events, like natural disasters (Prinstein et al., 1996), but such family social support has been shown to be eroded by natural disasters (Banks & Weems, 2014). As our entire sample were exposed to the hurricane and lived in FEMA-declared disaster areas, we cannot resolve this. Innovative methods and collaborative research efforts will be needed to compare the effects of child and parental internalizing symptoms on each other longitudinally in families exposed to natural disasters and age-matched unexposed controls.

Our finding that children’s pre-Sandy anxiety symptoms predicted mothers’ post-Sandy depression symptoms only at low, and not at high, levels of hurricane stress is intriguing. Research has demonstrated that parenting a child with emotional problems is challenging and stressful, and this has been shown to mediate the relationship of youth depression on their mothers’ later depression in a non-disaster exposed community sample (Raposa et al., 2011). Exposure to higher levels of hurricane stress may have disrupted the influence of children’s symptoms on their mothers’ emotional adjustment (e.g., Allmann et al., 2016; Hughes & Gullone, 2010; Raposa et al., 2011; Sellers et al., 2016). Under low hurricane stress, mothers’ may focus on their child’s problems, causing them distress, but in the context of higher stress, their focus may shift to coping with more pressing threats. If this is the case, we would expect that as the effects of the hurricane recede, the interaction will fade and a main effect of children’s symptoms on mothers’ depression will emerge. Because this analysis was exploratory and the finding unexpected, caution is needed in evaluating this result and our interpretation until it can at least be replicated.

Strengths of the present study included the availability of data on symptoms prior to the hurricane and using separate informants for child and mother internalizing symptoms to reduce informant and method biases. However, the study also had limitations. First, our sample was demographically homogenous. Second, fathers may be an important source of support or distress within a family system following a natural disaster, but data were only available in the present study to examine reciprocal processes within mother-child dyads. Third, the present study may have overestimated partner effects because, contrary to prior work (e.g., Goodman et al., 2011), child internalizing symptoms and mother depression symptoms were uncorrelated before the hurricane. This may reflect developmental influences on the quality of children’s self-reports of their internalizing symptoms (e.g., Kiss et al., 2007; Renouf & Kovacs, 1994). The higher coefficient alphas for the child report measures in the post-Sandy assessments than the age 9 assessments are consistent with this interpretation. Fourth, overall, our sample was exposed to modest levels of hurricane stressors so we were unable to examine exposure thresholds for spillover or co-escalation effects. Our findings may also not be applicable to families experiencing more significant disaster exposure. Power to detect significant interactions may also have been limited in the present study. Relatedly, data were not available to examine maternal anxiety nor mother or child post-traumatic stress (PTS) symptoms, which may be more sensitive to the effects of disaster stress. Fifth, we asked only mothers to report on their families’ hurricane stress exposure, rather than obtaining separate reports from mothers and children, although parent and adolescent reports were highly correlated in Shi et al. (2018; r = .60 a year after the earthquake). This may explain why our models accounted for more variance in mother’s symptoms than children’s symptoms. Sixth, post-Sandy child internalizing and maternal depression symptoms were assessed concurrently with Hurricane Sandy stress exposure; although mothers and children were asked to rate symptoms either based on the period since the hurricane or one to two weeks prior to the post-Sandy assessment, it is possible that changes in some of these symptoms preceded the hurricane. Seventh, mothers’ post-Sandy depression symptoms may have influenced their reporting of Hurricane Sandy stress exposure. Informant or interview measures of Hurricane Sandy stress exposure and parent symptoms would be ideal, although the DID correlates well with clinician ratings of depression symptoms (Zimmerman et al., 2004). Eighth, the clinical significance of these findings is unclear, as we did not include post-Sandy diagnostic interviews and symptoms were generally not in the clinical range. Further, although differences between included and dropped cases were minor, more severely anxious children and their mothers were less likely to participate in the study. Finally, we are unable to determine how long the effects we observed persist after Hurricane Sandy. This is an important direction for future research, given that most studies find that their effects are fleeting following natural disasters.

In sum, mothers’ and children’s internalizing symptoms changed in tandem and influenced each other in nuanced ways over the period spanning Hurricane Sandy. Mothers’ depression symptoms before the hurricane contributed to increases in their children’s depression symptoms after the hurricane. In addition, at low levels of hurricane stress, children’s prior anxiety symptoms impacted their mothers’ later depression symptoms. In contrast, at high levels of hurricane stress, child symptoms did not influence mother’s post-Sandy symptoms. Thus, reciprocal effects of internalizing symptoms between mothers and children were evident only at low levels of hurricane stress. These findings highlight the importance of considering both commonly observed developmental processes in parent-child relationships and the potential disruption of such processes after exposure to natural disasters.

Acknowledgments

This work was supported by the National Institute of Mental Health (NIMH) grant R01 MH069942 (DNK).

Footnotes

Maternal depression data in this manuscript have been reported in two other publications, but all child data and the ideas in this manuscript have not been previously disseminated.

1

At post-Sandy, 13 mothers exceeded the threshold for MDD on the DID. Eleven children endorsed clinically significant post-Sandy depression symptoms (raw CDI scores >85th percentile), and 54 children exhibited clinically significant anxiety symptoms after the hurricane (SCARED scores >25).

References

  1. Aiken LS, & West SG (1991). Multiple regression: Testing and interpreting interactions. Newbury Park: Sage.
 [Google Scholar]
  2. Allmann AES, Kopala-Sibley DC, & Klein DN (2016). Preschoolers’ psychopathology and temperament predict mothers’ later mood disorders. Journal of Abnormal Psychology, 44(3), 421–432. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Anderson ER, & Hope DA (2008). A review of the tripartite model for understanding the link between anxiety and depression in youth. Clinical Psychology Review, 28(2), 275–287. [DOI] [PubMed] [Google Scholar]
  4. Baer R (2003). Mindfulness training as a clinical intervention: A conceptual and empirical review. Clinical Psychology: Science and Practice, 10, 125–143. [Google Scholar]
  5. Banks DM, & Weems CF (2014). Family and peer social support and their links to psychological distress among hurricane-exposed minority youth. American Journal of Orthopsychiatry, 84, 341–352. [DOI] [PubMed] [Google Scholar]
  6. Bell RQ (1968). A reinterpretation of the direction of effects in studies of socialization. Psychological Review, 75, 81–95. [DOI] [PubMed] [Google Scholar]
  7. Birmaher B, 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 American Academy of Child and Adolescent Psychiatry, 38, 1230–1236. [DOI] [PubMed] [Google Scholar]
  8. Bonanno GA, Brewin CR, Kaniasty K, & La Greca AM (2010). Weighing the costs of disaster: Consequences, risks, and resilience in individuals, families, and communities. Psychological Science in the Public Interest, 11, 1–49. [DOI] [PubMed] [Google Scholar]
  9. Bromet EJ, Goldgaber D, Carlson G, Panina N, Golovakha E, Gluzman SF, … Schwartz JE (2000). Children’s well-being 11 years after the Chornobyl catastrophe. Archives of General Psychiatry, 57(6), 563–571. [DOI] [PubMed] [Google Scholar]
  10. Bronfenbrenner U, & Morris PA (2006). The bioecological model of human development. In Lerner RM & Damon W (Eds.), Handbook of Child Psychology: Theoretical Models of Human Development (pp. 793–828). Hoboken, NJ: John Wiley & Sons, Inc. [Google Scholar]
  11. Cook WL, & Kenny DA (2005). The actor-partner interdependence model: A model of bidirectional effects in developmental studies. International Journal of Behavior and Development, 29, 101–109. [Google Scholar]
  12. Costa NM, Weems CF, & Pina AA (2009). Hurricane Katrina and youth anxiety: The role of perceived attachment beliefs and parenting behaviors. Journal of Anxiety Disorders, 23, 935–941. [DOI] [PubMed] [Google Scholar]
  13. Cuijpers PC, 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, 237–245. [DOI] [PubMed] [Google Scholar]
  14. Cummings CM, Caporino NE, & Kendall PC (2014). Comorbidity of anxiety and depression in children and adolescents: 20 years after. Psychological Bulletin, 140(3), 816–845. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Galea S, Brewin CR, Gruber M, Jones RT, King DW, King LA, ...Kessler RC (2007). Exposure to hurricane-related stressors and mental illness after Hurricane Katrina. Archives of General Psychiatry, 64, 1427–1434. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Goodman SH, & Garber J (2017). Evidence-based interventions for depressed mothers and their young children. Child Development, 88(2), 368–377. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Goodman SH, & Gotlib IH (1999). Risk for psychopathology in the children of depressed mothers: A developmental model for understanding mechanisms of transmission. Psychological Review, 106, 458–490. [DOI] [PubMed] [Google Scholar]
  18. Goodman SH, Rouse MH, Connell AM, Broth MR, Hall CM, & Heyward D (2011). Maternal depression and child psychopathology: A meta-analytic review. Clinical Child and Family Psychology Review, 14, 1–27. [DOI] [PubMed] [Google Scholar]
  19. Hu L, & Bentler PM (1999). Cutoff criteria for fix indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. [Google Scholar]
  20. Hughes EL, & Gullone E (1010). Reciprocal relationships between parent and adolescent internalizing symptoms. Journal of Family Psychology, 24(2), 115–124. [DOI] [PubMed] [Google Scholar]
  21. Johnson PO, & Fay LC (1959). The Johnson-Neyman technique, its theory and application. Psychometrika, 15, 349–367. [DOI] [PubMed] [Google Scholar]
  22. Juth V, Silver RC, Seyle DC, Widyatmoko CS, & Tan ET (2015). Post-disaster mental health among parent-child dyads after a major earthquake in Indonesia. Journal of Abnormal Child Psychology, 43(7), 1309–1318. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kiss E, Gentzler AM, George C, Kapornai K, Tamas Z, Kovacs M, & Vetro A (2007). Factors influencing mother-child reports of depressive symptoms and agreement among clinically referred depressed youngsters in Hungary. Journal of Affective Disorders, 100(1–3), 143–151. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Klein DN, & Finsaas MC (2017). The Stony Brook Temperament Study: Early antecedents and pathways to emotional disorders. Child Development Perspectives, 11, 257–263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kopala-Sibley DC, Danzig AP, Kotov R, Bromet EJ, Carlson GA, Olino TM, . . . Klein DN (2016a). 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), 471–481. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. 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. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kopala-Sibley DC, Kotov R, Bromet EJ, Carlson GA, Danzig AP, Black SR, & Klein DN (2016b). Personality diatheses and Hurricane Sandy: Effects on post-disaster depression. Psychological Medicine, 46(4), 1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Koplewicz HS, Vogel JM, Solanto MV, Morrissey RF, Alonso CM, Abikoff H, … Novick RM (2002). Child and parent response to the 1993 World Trade Center bombing. Journal of Traumatic Stress, 15, 77–85. [DOI] [PubMed] [Google Scholar]
  29. Kovacs M (1985). The children’s depression inventory. Psychopharmacology Bulletin, 21, 995–998. [PubMed] [Google Scholar]
  30. Masten AS, & Narayan AJ, (2010). Child development in the context of disaster, war, and terrorism: Pathways of risk and resilience. Annual Review of Psychology, 63, 227–257. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Minuchin P (1985). Families and individual development: Provocations from the field of family therapy. Child Development, 56, 289–302. [PubMed] [Google Scholar]
  32. Mirzamani M, & Bolton D (2002). PTSD symptoms of mother’s following occurrence of a disaster affecting their children. Psychological Reports, 90, 431–483. [DOI] [PubMed] [Google Scholar]
  33. Morris MC, Ciesla JA, & Garber J (2008). A prospective study of the cognitive-stress model of depressive symptoms in adolescents. Journal of Abnormal Psychology, 117, 719–734. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Neill EL, Weems CF, & Scheeringa MS (2018). CBT for child PTSD is associated with reductions in maternal depression: Evidence for bidirectional effects. Journal of Clinical Child & Adolescent Psychology, 47(3), 410–420. [DOI] [PubMed] [Google Scholar]
  35. Neria Y, & Shultz JM (2012). Mental health effects of Hurricane Sandy: Characteristics, potential aftermath, and response. Journal of the American Medical Association, 308, 2571–2572. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Norris FH, Friedman MJ, Watson PJ, Byrne CM, Diaz E, & Kaniasty K (2002a). 60,000 disaster victims speak: Part I. An empirical review of the empirical literature, 1981–2001. Psychiatry, 65(3), 207–239. [DOI] [PubMed] [Google Scholar]
  37. Norris FH, Friedman MJ, & Watson PJ (2002b). 60,000 disaster victims speak: Part II. Summary and implications of the disaster mental health research. Psychiatry, 65(3), 240–260. [DOI] [PubMed] [Google Scholar]
  38. Norris FH, Sherrieb K, & Galea S (2010). Prevalence and consequences of disaster-related illness and injury from Hurricane Ike. Rehabilitation Psychology, 55, 21–230. [DOI] [PubMed] [Google Scholar]
  39. Patterson GR (1982). Coercive family process. Eugene, OR: Castalia. [Google Scholar]
  40. Perloe A, Esposito-Smythers C, Curby TW, & Renshaw KD (2014). Concurrent trajectories of change in adolescent and maternal depressive symptoms in the TORDIA study. Journal of Youth and Adolescence, 43(4), 612–628. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Pfefferbaum B, Houston JB, North CS, & Regens JL (2008). Youth’s reactions to disasters and the factors that influence their response. Prevention Research, 15(3), 3–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Pfefferbaum B, & North CS (2008). Children and families in the context of disasters: Implications for preparedness and response. Family Psychology, 24(2), 6–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Pfefferbaum B, Seems CF, Scott BG, Nitiema P, Noffsinger MA, Pfefferbaum RL, Varma V, & Chakraburtty A (2013). Research methods in child disaster studies: A review of studies generated by the September 11, 2001, terrorist attacks; the 2004 Indian Ocean Tsunami; and Hurricane Katrina. Child Youth Care Forum, 42, 285–337. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Prinstein MJ, La Greca AM, Vernberg EM, & Silverman WK (1996). Children’s coping assistance: How parents, teachers, and friends help children cope after a natural disaster. Journal of Clinical Child Psychology, 25(4), 463–475. [Google Scholar]
  45. Raposa EB, Hammen CL, & Brennan PA (2011). Effects of child psychopathology on maternal depression: The mediating role of child-related acute and chronic stressors. Journal of Abnormal Child Psychology, 39, 1177–1186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Renouf AG, & Kovacs M (1994). Concordance between mothers’ reports and children’s self-reports of depressive symptoms: A longitudinal study. Journal of the American Academy of Child and Adolescent Psychiatry, 33(2), 208–216. [DOI] [PubMed] [Google Scholar]
  47. Sameroff A (1975). Transactional models in early social relations. Human Development, 18(1–2), 65–79. [Google Scholar]
  48. Sellers R, Hammerton G, Harold GT, Mahedy L, Potter R, Langley K, . . . Collishaw S (2016). Examining whether offspring psychopathology influences illness course in mothers with recurrent depression using a high-risk longitudinal sample. Journal of Abnormal Psychology, 125(2), 256–266. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Shi X, Zhou Y, Geng F, Li Y, Zhou J, Lei B, Chen S, Chen X, & Fan F (2018). Posttraumatic stress disorder symptoms in parents and adolescents after the Wenchuan earthquake: A longitudinal actor-partner interdependence model. Journal of Affective Disorders, 226, 301–306. [DOI] [PubMed] [Google Scholar]
  50. Weems CF, & Overstreet S (2008). Child and adolescent mental health research in the context of Hurricane Katrina: An ecological needs-based perspective and introduction to the special section. Journal of Clinical Child and Adolescent Psychology, 37, 487–494. [DOI] [PubMed] [Google Scholar]
  51. Weems CF, Pina AA, Costa NM, Watts SE, Taylor LK, & Cannon MF (2007). Predisaster trait anxiety and negative affect predict posttraumatic stress in youths after Hurricane Katrina. Journal of Consulting and Clinical Psychology, 75(1), 154–159. [DOI] [PubMed] [Google Scholar]
  52. Zimmerman M, Sheeran T, & Young D (2004). The Diagnostic Inventory for Depression: A self-report scale to diagnose DSM-IV major depressive disorder. Journal of Clinical Psychology, 60, 87–110. [DOI] [PubMed] [Google Scholar]

RESOURCES