Chronicity, Severity, and Timing of Maternal Depressive
Symptoms: Relationships With Child Outcomes at Age 5
Patricia A. Brennan
1, Constance Hammen
2, Margaret J. Andersen
3, William Bor
4, Jake M.
5, Gail M. Williams
1Department of Psychology, Emory University
3Department of Obstetrics and Gynecology, University of Queensland
Health Service, Royal Children's Hospital
Patricia A. Brennan, Department of Psychology, Emory University, 30322. Email:
2Department of Psychology, University of California
4Child and Youth Mental
5Department of Anthropology and Sociology, University of
6Department of Tropical Health, University of Queensland Address for Correspondence:
The relationships between severity, chronicity, and timing of maternal depressive symptoms and
child outcomes were examined in a cohort of 4,953 children. Mothers provided self-reports of
depressive symptoms during pregnancy, immediately postpartum, and when the child was 6 months
old and 5 years old. At the age 5 follow-up, mothers reported on children's behavior and children
completed a receptive vocabulary test. Results suggest that both the severity and the chronicity of
maternal depressive symptoms are related to more behavior problems and lower vocabulary scores
in children. The interaction of severity and chronicity of maternal depressive symptoms was
significantly related to higher levels of child behavior problems. Timing of maternal symptoms was
not significantly related to child vocabulary scores, but more recent reports of maternal depressive
symptoms were associated with higher rates of child behavior problems.
Many studies have documented the association between maternal depression and adverse outcomes
in children. Studies with clinical samples of depressed parents, focused primarily on mothers, have
shown elevated rates of depression in children as well as anxiety and disruptive behavior disorders
(reviewed in Downey & Coyne, 1990; Hammen, 1999). Studies with community samples of women
with self-reported symptoms of depression have also shown adverse outcomes in children. These
latter studies included infants, toddlers, and children of varying ages, and virtually all measures of the
children's affect and behaviors showed evidence of maladaptive reactions to their mothers' dysphoria
(reviewed in Downey & Coyne; Gelfand & Teti, 1990).
The consistency of detrimental effects across numerous samples and methodologies has doubtless
suggested that maternal depressive symptoms invariably have a negative impact on children's
behavior. However, most of the studies have been cross-sectional in design, and they provide little
information about the nature of maternal depression and how often it leads to problems for children.
Depression is extremely heterogeneous in its manifestations, ranging from mild and transitory mood
distress that is entirely normal to persisting and severe depressed mood accompanied by somatic,
cognitive, and behavioral disturbances that impair normal role functioning. In between these
extremes may be chronic but relatively mild symptoms or one or more periods of intense symptoms
of various durations with normal functioning in between. Lack of clinical information about the
features and history of depressive symptoms in the community studies plus the relative paucity of
longitudinal studies of children of depressed or dysphoric mothers make it difficult to characterize
the nature of depression and its consequences.
There are three separate but typically confounded characteristics of maternal depression that require
clarification: its severity, its chronicity, and the timing of the children's exposure to it. Greater
severity of depressive symptoms is likely to be accompanied by greater impairment of functioning.
Several studies have suggested that the severity and chronicity of maternal depression rather than its
diagnosis per se are related to children's outcomes (e.g., Keller et al., 1986; Sameroff, Barocas, &
Seifer, 1984). Hammen (1991) reported that children whose mothers had more depressive episodes
were themselves more likely to have more severe diagnoses. Warner, Mufson, and Weissman (1995),
in their most recent offspring study, demonstrated that only recurrent, early-onset major depression in
the parent was significantly associated with major depression in the offspring.
In a recent study focused on mothers with depressive symptoms and their infants, Campbell, Cohn,
and Meyers (1995) demonstrated that relatively fewer positive interactions between mothers and
their babies were observed only among women whose symptoms had persisted through 6 months
postpartum. Those who were depressed at 2 months postpartum but whose depressions remitted over
time did not differ from nondepressed comparison women. Frankel and Harmon (1996), in their
study of parentœchild observations, found that many depressed women did not perform differently
in laboratory tasks with their children than nondepressed women; however, those who had more
severe or chronic depressions were seen as significantly less emotionally available and had higher
rates of insecurely attached children than women with episodic depression only. Also, Teti and
colleagues (Teti, Gelfand, Messinger, & Isabella, 1995) found that the most chronically and severely
depressed women had infants and preschoolers with insecure attachments marked by less coherent
and organized strategies.
The relative importance of the chronicity and severity of maternal depression is somewhat difficult to
gauge from previous studies. This is because chronicity and severity are commonly confounded–
higher levels of depressive symptoms generally last longer than mild depressions, and their distinct
relationships with child functioning have not been assessed. It is important, nevertheless, to try to
untangle the two dimensions, because they may have somewhat different theoretical and treatment
In order to examine these two properties of depression separately, a longitudinal study is required
that includes multiple assessments of maternal symptoms. The present study represents a unique
sample of women and their children studied over a 5-year period during which maternal depressive
symptoms were assessed at four points–during pregnancy, immediately postpartum, 6 months
postpartum, and at 5 years. The sample is well suited to the study of the issue of chronicity and
severity, because it is large enough (N = 4,953) to include adequate numbers of women who vary
in the characteristics of their depressive experiences. Clinical samples and most community
studies have been much too small to adequately explore variability in depression features.
Finally, the issue of the timing of maternal depression in relation to children's development is also
important to address. There may be relatively more crucial periods during which exposure to a
withdrawn or irritable mother has negative consequences for a child. It might be predicted, for
example, that the period of the first months of life is especially important because maternal
depression may impair the development of secure attachment. To date, there are several studies that
have linked concurrent maternal depression and infant attachment (e.g., Campbell, Cohn, Meyers,
Ross, & Flanagan, 1993; Radke-Yarrow, Cummings, Kuczynski, & Chapman, 1985; Teti et al.,
1995). However, only two studies of which we are aware have examined the timing of exposure to
maternal depression over a longer period. One prospective study of 131 mothers found that maternal
depression reported at 14 months postpartum was more predictive of child behavior disturbance than
was maternal depression reported at either 27 or 42 months postpartum (Ghodsian, Zajicek, &
Wolkind, 1984). In a retrospective study, Alpern and Lyons-Ruth (1993) found that type of child
maladjustment was differentially associated with timing of maternal symptoms, such that
postpartum-only depression was more associated with anxiety symptoms and recent depression-only
predicted hyperactivity symptoms. The sample was nonclinical, low income, and high adversity and
relied on retrospective accounts of depression. In the present study, by restricting the timing of
exposure to the first 5 years of the child's life and controlling for the confound of chronicity of
symptoms, we present a prospective but limited test of this question.
The goals of the present analyses, therefore, were to examine preliminary questions about severity,
chronicity, and timing of maternal depressive symptoms in relation to children's behavioral and
cognitive functioning at age 5. Specifically, we examined (a) the relationship between severity of
maternal depressive symptoms and child outcomes, (b) the relationship between chronicity of
maternal depressive symptoms and child outcomes, and (c) the interaction between maternal
depressive symptom severity and chronicity as it related to child outcomes. We also examined the
strength of the association between child outcomes and maternal depressive symptoms reported at
different times during the child's early development. Both behavior problems and cognitive
functioning measures were included as child outcomes, because both variables have been shown to
be sensitive to maternal depression (e.g., Lyons-Ruth, Connell, & Grunebaum, 1990; Murray, 1992;
Whiffen & Gotlib, 1989). In addition, demographic risk factors associated with maternal depressive
symptoms were included as statistical controls in our analyses.
The participants in this study were 4,953 mothers and their children born between 1981 and 1984 at
the Mater Misericordiae Mother's Hospital in Queensland, Australia (Keeping et al., 1989). The goal
of the original study was to examine social factors and children's health and development. Mothers in
this sample completed interviews and questionnaires about themselves and their children at four
different times–during pregnancy, 3 to 4 days after the birth of their children, 6 months after the birth
of their children, and when their children were 5 years of age. In addition, the children were directly
assessed for cognitive functioning at age 5. Children in this sample were representative of individuals
born in public hospitals in Queensland and therefore represented a relatively lower socioeconomic
sector (working and lower middle class) of the population of Australia.
The 4,953 participants included in this study are those children in the cohort who were followed up
through the age of 5 years. Of these children, 2,596 (52%) were male and 2,357 (48%) were female.
The large majority (92%) were of Caucasian ethnicity. Their mothers' mean age at the time of birth
was 25.4 years (SD = 5.0), and their mean birth order was 1.99 (SD = 1.06). Family yearly income
was reported at a mean of 3.6 (SD = 0.92) on the following scale: 1 = $0œ$5,199; 2 =
$5,200œ$10,399; 3 = $10,400œ$15,599; 4 = $15,600œ $20,799; 5 = $20,800œ$25,999; and 6 =
$26,000 or more (all in Australian dollars). Mothers' education was reported at a mean of 4.28 (SD =
1.29) on the following scale: 1 = preschool, 2 = primary school, 3 = started secondary school, 4 =
completed Grade 10, 5 = completed Grade 12, 6 = completed business, nursing, or secretarial college,
and 7 = completed university. Mothers in the sample reported a mean of 0.31 (SD = 0.61) changes in
marital status over the course of the 5 years of follow-up.
The children who were followed to age 5 represented approximately 70% of the participants who
were born into the cohort. Those children lost to follow-up differed significantly from the retained
participants in terms of mothers' age (lost, M = 24.32 years, SD = 5.3; retained, M = 25.38 years, SD
= 5.0), mothers' parity (lost, M = 2.12, SD = 1.4; retained, M = 2.03, SD = 1.2), family income (lost,
M = 2.77, SD = 1.1; retained, M = 3.08, SD = 1.1), and mothers' education (lost, M = 4.13, SD = 1.1;
retained, M = 4.27, SD = 1.1). Compared to those who remained in the study, mothers lost to follow-
up were also significantly more likely to be single at the time of their entrance into the study (38%
vs. 20%) and to report more depressive symptoms at the time of their child's birth (lost, M = 1.17, SD
= 1.5; retained, M = 0.89, SD = 1.3). Because of this pattern of attrition, the results of this study may
provide a conservative estimate of the true association between maternal depressive symptoms and
Maternal depressive symptoms were indicated by self-report on the seven depression items of the
Delusions-Symptoms-States Inventory of Bedford and Foulds (1978):
I have been so miserable that I have had difficulty sleeping.
I have been depressed without knowing why.
I have gone to bed not caring if I never woke up.
I have been so low in spirit that I have sat up for ages doing absolutely nothing.
The future seems hopeless.
I have lost interest in just about everything.
I have been so depressed that I have thought of doing away with myself.
Mothers completed the Delusions-Symptoms-States Inventory four times, as noted earlier. A
response of —some of the time,“ —most of the time,“ or —all of the time“ to a depression item
from the Delusions-Symptoms-States Inventory indicated that a symptom of depression had been
endorsed by the mother. The internal consistency (alpha) of the depression items of the Delusions-
Symptoms-States Inventory ranged from .71 to .81 across the four administrations of this measure.
The Delusions-Symptoms-States Inventory had been chosen as the measure of maternal mental
health for the Mater Misericordiae hospital birth cohort study because it was a valid screening
instrument for mental health (e.g., Bedford & Foulds, 1977) and did not include symptoms that
might be confused with the effects of pregnancy or childbirth. The Delusions-Symptoms-States
Inventory is comparable to other validated self-report instruments that are known to be screens for
major depressive episodes. For example, five of its seven items overlap with the Beck Depression
Inventory (Beck, Ward, Mendelson, Mock, & Erbaugh, 1961), and four items overlap with the
Center for Epidemiological Studies Depression Scale (Radloff, 1977). In addition, the Delusions-
Symptoms-States Inventory depression items include one explicit suicidal item and two
hopelessness items that typically predict suicidal thoughts. These types of items identify significant
depression and have been found to be the self-report items that distinguish major depression from
dysthymic disorder (Steer, Beck, Brown, & Berchick, 1987).
Using data from the age 5 follow-up of this cohort, we compared the mothers' ratings of depressive
symptoms on the Delusions-Symptoms-States Inventory with the mothers' self-report of mental
health treatment. Mothers who reported high levels of depressive symptoms on the Delusions-
Symptoms-States Inventory were six times as likely to have received treatment for mental health
problems than were mothers who reported low levels of depressive symptoms, χ
192.8, p < .00001.
2(1, N = 4,856) =
We also administered the Delusions-Symptoms-States Inventory and the Beck Depression Inventory
to two separate samples of individuals to assess the convergent validity of these measures. In one
U.S. sample of 112 female undergraduates, the Beck Depression Inventory and the Delusions-
Symptoms-States Inventory were significantly correlated (r = .78, p < .01). And in a recent follow-up
of a subsample of 450 mothers from the Mater Misericordiae hospital cohort, a significant correlation
(r = .75, p < .01) was obtained between the Delusions-Symptoms-States Inventory and the Beck
Depression Inventory. The significance and the level of these correlations suggest that the Delusions-
Symptoms-States Inventory is a valid self-report measure of depressive symptoms.
Severity of maternal depressive symptoms was defined in two ways for the purposes of the present
study. We established a continuous measure of severity that reflected the maximum number of
depressive symptoms (from 0 to 7) ever reported on a single administration of the Delusions-
Symptoms-States Inventory (M = 1.7, SD = 1.7). For some analyses, we also operationalized
categories of maternal symptoms severity for each Delusions-Symptoms-States Inventory
administration as follows: Women who endorsed two or three symptoms of depression were
considered to have a moderate level of depressive symptoms, and women who endorsed four or more
symptoms of depression were considered to have a severe level of depressive symptoms. Using the
Beck Depression Inventory data and the Delusions-Symptoms-States Inventory data from our recent
follow-up of a subsample of these mothers (n = 450), we noted that women in the Delusions-
Symptoms-States Inventory category of moderate had a mean Beck Depression Inventory score of
11, and women in the Delusions-Symptoms-States Inventory category of severe had a mean Beck
Depression Inventory score of 19. According to scoring guidelines for the Beck Depression
Inventory, a score of 11 falls in the mild to moderate depression range, and 19 falls in the moderate to
severe range (Beck, Steer, & Garbin, 1988). These Beck Depression Inventory scores were
significantly different from one another, t(137) = 7.66, p < .001. In addition, both the moderate,
t(417) = 11.28, p < .001, and severe, t(400) = 21.92, p < .001, groups had significantly higher mean
Beck Depression Inventory scores than the other women in the subsample.
For the assessment of chronicity, a continuous measure was operationally defined according to the
number of Delusions-Symptoms-States Inventory administrations (from 0 to 4) on which the mother
reported either moderate or severe levels of depressive symptoms (M = 0.7, SD = 1.0). Because the
measure of chronicity is in part dependent on the measure of severity, these two variables are highly
correlated (r = .81).
In part to counteract the potential issue of collinearity in our analyses, and in part to tease apart the
distinct relationships between severity versus chronicity of maternal depressive symptoms and child
outcome, we also created four nonoverlapping comparison groups on the basis of maternal
Delusions-Symptoms-States Inventory scores. These groups represented orthogonal measures of the
severity (moderate vs. severe) and chronicity (one report of at least moderate level vs. two or more
reports) of maternal depressive symptoms and allowed for a comparison of the following types of
maternal depressive symptoms in terms of their relationships with child outcome: (a) neither severe
nor chronic, (b) chronic but not severe, (c) severe but not chronic, and (d) both chronic and severe.
For analyses examining the timing of depressive symptoms, it was essential to control for the
relative factors of severity and chronicity. Timing, therefore, was defined within a restricted range of
severity and chronicity: Children of mothers who reported a severe level of depressive symptoms on
only a single administration of the Delusions-Symptoms-States Inventory (i.e., during pregnancy
only, at birth only, at 6 months only, or at 5 years only) were compared with one another, as were
children of mothers who reported a moderate level of depressive symptoms on only a single
administration of the Delusions-Symptoms-States Inventory.
Child behavior problems
A total behavior problem scale score (α = .90, M = 14.9, SD = 8.5) was obtained from mothers'
reports on a subset of 33 items from the Child Behavior Checklist (Achenbach, 1991) when the child
was 5 years of age. This shortened form of the Child Behavior Checklist was used for the purposes of
reducing the time required for questionnaire administration. Items for the shortened form were
chosen on the basis of face validity as those most likely to be relevant to the behavioral functioning
of 5-year-old children. The full range of child behavior problems–aggression, oppositional behavior,
hyperactivity, anxiety, withdrawal, and depression–was reflected in these items. In a sample of 76
mothers of 5-year-olds in Australia, the total behavior problem score from this shortened form
correlated .98 with the total behavior problem score obtained from the standardized Child Behavior
Checklist form developed by Achenbach (Bor et al., 1997).
Child cognitive functioning
At the age of 5, the children who were assessed directly completed the Peabody Picture Vocabulary
Test–Revised (Dunn & Dunn, 1981). Research assistants who administered the Peabody test were
blind to the hypotheses of the study. The Peabody test is a standardized measure of vocabulary
development that does not rely on expressive language skills. Raw scores were converted to standard
scores based on chronological age norms (M = 99.6, SD = 13.6). Peabody vocabulary data were
available on 3,767 children; because of staffing patterns or scheduling difficulties, not all children
could be tested directly. Children who did not take the Peabody test were not significantly different
from those who did in terms of behavior problem scores, t(4951) = 0.26, p = .80; maternal education,
t(4937) = 0.71, p = .48; family income, t(4948) = 1.41, p = .16; gender, t(4964) = 0.76, p = .45; or
maternal depressive symptom chronicity scores, t(4964) = 1.75, p = .08. Children without Peabody
test data had more changes in mother's marital status, t(4964) = 5.29, p < .001, mothers with a
younger maternal age at birth, t(4964) = 4.14, p < .001, and lower birth order, t(4929) = 4.49, p <
.001, than did children for whom the Peabody test data were available. In addition, those without
Peabody test data had higher maternal depressive symptom severity scores, t(4964) = 3.80, p < .001,
than those children for whom these data were available. Therefore, our analyses may represent a
conservative test of the association between maternal depressive symptoms and child cognitive
Multiple regression analyses were performed to examine the relationship between the severity and
chronicity of maternal depressive symptoms and child outcomes. Child outcomes included behavior
problems (Child Behavior Checklist scores) and cognitive functioning (Peabody test scores).
Analyses of covariance (ANCOVAs) were also performed to assess the relationships between these
child outcomes and maternal depressive symptom severity, chronicity, and timing. Initial analyses
separated by gender revealed similar patterns of results for boys and girls. Therefore, rather than
presenting all the results for each gender separately, we combined the data for both genders and
included gender as a statistical control in our analyses. All analyses presented included the following
variables as statistical controls or covariates: gender and birth order of child, mother's age and
education, family income, and number of changes in mother's marital status.
Severity of Depressive Symptoms
First we examined the relationship between the severity of the worst maternal depressive symptoms
and child behavior problems and cognitive functioning. In these regression analyses, the covariates
were entered as a block in the first step, and the highest maternal depressive symptom severity scores
were then entered as the predictor in the second step. Betas for covariates and depressive symptom
severity, as well as the significance of these variables, are shown in Table 1. Severity of maternal
depressive symptoms made a significant contribution to the prediction of both total behavior problem
scores and Peabody vocabulary scores in this sample. The direction of the relationship was as
predicted–the higher the maternal depressive severity score, the more the behavior problems and the
lower the vocabulary scores for the children. It should be noted that although the relationship was
statistically significant, the percentage of the variance in Peabody test scores explained by the
severity of maternal depressive symptoms was close to zero.
Severity of Maternal Depressive Symptoms and Child Outcomes at Age 5
Chronicity of Depressive Symptoms
Next we examined the relationship between chronicity of mothers' depressive symptoms and child
behavior problems and cognitive functioning. As in the analyses for severity, in these regression
analyses covariates were entered as a block in the first step, and maternal depressive symptom
chronicity scores were then entered in the second step (see Table 2). Chronicity of maternal
depressive symptoms was significantly related to Child Behavior Checklist total behavior problem
scores. Chronicity was also related to Peabody vocabulary scores. Again, the direction of the
relationship was as predicted–more chronic maternal depressive symptoms were associated with
higher levels of child behavior problems and lower scores on the receptive vocabulary test. Again,
the percentage of variance in Peabody test scores explained by maternal depressive symptoms was
close to zero.
Chronicity of Maternal Depressive Symptoms and Child Outcomes at Age 5
Severity þ Chronicity Interactions
Additional regression analyses were undertaken to examine whether chronicity and severity of
maternal depressive symptoms might interact in their prediction of child outcomes. In these analyses,
covariates were entered as a block in the first step, severity and chronicity were then entered as a
block in the second step, and the Severity þ Chronicity interaction term was entered in the last step.
The interaction term was not significant in predicting Peabody test scores, ∆F(1, 3702) = 0.37, p =
.55, for the children in this sample. However, the interaction of severity and chronicity of maternal
depressive symptoms was significant in predicting total behavior problem scores, ∆F(1, 4894) = 3.90,
p < .05, with a beta of −.08 and a ∆R
2 of .001.
To better interpret this interaction finding and to counteract the problem of collinearity in the above
analysis (chronicity and severity correlated .81), we examined maternal depressive symptom severity,
chronicity, and their interaction in four nonoverlapping —mother depressed“ comparison groups. As
outlined in the Method section, these groups represented orthogonal measures of the severity
(moderate vs. severe) and chronicity (one report of at least a moderate level vs. two or more reports)
of maternal depressive symptoms, allowing for a comparison of the following types of maternal
depressive symptoms in relation to child outcome: (a) neither severe nor chronic, (b) chronic but not
severe, (c) severe but not chronic, and (d) both chronic and severe. Mean behavior problem scores for
these groups are presented in Figure 1. A 2 þ 2 ANCOVA revealed a significant Severity þ
Chronicity interaction in relation to child behavior problems, F(1, 2011) = 6.53, p < .05, d = .11. Post
hoc Duncan multiple comparisons revealed that the children whose mothers had experienced both
chronic and severe depressive symptoms had significantly more behavior problems than the children
in the other three maternal depressive symptom comparison groups.
Severity and chronicity of maternal depression as related to child behavioral and cognitive
Timing of Depressive Symptoms
As noted, timing was assessed for those women who had only a single elevated score in order to
avoid confounding timing with severity and chronicity. ANCOVAs were performed that controlled
for gender and birth order of child, mother's age and education, family income, and number of
changes in mother's marital status. In these analyses, the main effect of timing of moderate levels of
maternal depressive symptoms on child behavior problems was significant, F(3, 881) = 2.87, p < .05,
d = .19, as was the main effect of timing of severe levels of mother depressive symptoms, F(3, 205) =
5.54, p < .01, d = .33. Duncan multiple range tests revealed that behavior problem scores were higher
for children whose mothers reported moderate levels of depressive symptoms only at 6 months (n =
167, M = 17.21, SD = 8.3) or at 5 years (n = 236, M = 17.17, SD = 8.7) than for children whose
mothers reported moderate levels only during pregnancy (n = 273, M = 15.38, SD = 8.7) or at birth (n
= 206, M = 15.42, SD = 8.6). In addition, there was a significantly higher level of behavior problems
in children whose mothers reported severe levels of depressive symptoms only at age 5 (n = 97, M =
19.51, SD = 10.0) than in children whose mothers reported severe levels only during pregnancy (n =
47, M = 14.91, SD = 8.2), only at birth (n = 29, M = 14.93, SD = 7.6), or only at 6 months of age (n =
33, M = 13.36, SD = 8.2). There were no significant relationships between the timing of moderate
maternal depressive symptoms, F(3, 655) = 2.48, p = .06, or severe maternal depressive symptoms,
F(3, 153) = 0.91, p = .44, and child Peabody test scores.
Although several studies have suggested that chronicity and severity must be taken into account in
assessing the relationship between maternal depressive symptoms and children's outcomes, this is
the first study of which we are aware that has separately evaluated these typically overlapping or
confounded features on a longitudinal sample. Moreover, previous studies have not examined the
relative importance of these maternal depression characteristics, nor have they indicated whether
severity and chronicity would interact in their prediction of children's outcomes. Previous studies
have also rarely addressed the issue of the importance of the timing of the maternal depressive
symptoms and whether the child's age at exposure in the early years is differentially related to
Several of these questions were explored in a large community sample of women reporting on their
depressive symptoms on four occasions in their children's early development. With respect to the
severity of maternal depressive symptoms, there was a significant relationship with both children's
behavioral problem scores and their receptive vocabulary scores, such that increasingly severe
reported depressive symptoms were associated with increasingly negative outcomes. Regarding
chronicity, there was a similar pattern for behavioral problem scores and for Peabody test scores.
When chronicity and severity were examined together, a significant interaction occurred for child
behavior problems but not for child cognitive functioning. Children whose mothers' depressive
symptom history was characterized as both chronic and severe had higher levels of behavior
problems than did the other children in the sample. The combination of severity and chronicity
appears to be potent in relation to child behavior outcome. Timing of depressive symptoms–based on
data from women with a single episode of elevated symptoms–appeared to be only modestly related
to behavior problems scores and unrelated to Peabody vocabulary scores. The most pronounced
relationships were noted between child behavior outcome and maternal depressive symptoms
reported at the child's age of 5 years, with little apparent relationship between child behavior
problems and one-time maternal depressive symptoms reported during pregnancy or immediately
The overall results confirm results obtained with clinical as well as other nonclinical samples
indicating significant associations between maternal depressive symptoms and negative outcomes in
children's behavior and cognitive functioning (reviewed in Downey & Coyne, 1990; Gelfand & Teti,
1990; Hammen, 1999). The effect sizes for Peabody vocabulary scores were small. The results
suggest that demographic factors such as maternal education are more strongly related to child
cognitive functioning than are maternal depressive symptoms.
The effect sizes for behavior problems were small to medium in size. This may be due to the fact that
this study was based on a community sample rather than a treatment-seeking sample and that attrition
may have resulted in more conservative estimates of the true association between maternal depressive
symptoms and child behavior outcomes. Statistical significance does not necessarily imply clinical
significance of results, especially when sample sizes are large. To better assess the clinical
significance of our results we completed follow-up analyses in which we examined the relationship
between maternal depressive symptoms and behavior problem scores higher than the 90th percentile.
These analyses revealed that over 30% of children of mothers who reported severe, chronic
depressive symptoms had scores higher than the 90th percentile on the Child Behavior Checklist,
compared to fewer than 5% of the children whose mothers did not report depressive symptoms. (Note
that these are percentiles for this sample, not t-score-based percentiles, because t scores were not
available for the abbreviated form of the Child Behavior Checklist used in this study.)
The relationships between severity, chronicity, and timing of maternal depressive symptoms and
child behavior problems were significant when maternal age, child birth order and gender, maternal
education, family income, and mothers' changes in marital status were controlled. This significance
suggests that these relationships with child outcome are not due solely to social risks associated
with maternal depressive symptoms. However, it is important to emphasize that the present study
cannot claim that the depressive symptoms were the direct cause of the negative impact on children.
This study is not an experiment, and therefore we cannot assess causal relations between maternal
depressive symptoms and child outcomes. It has been argued that the negative effects of parental
depression on children reflect a multifactorial process that may include not only biological or
genetic mechanisms but also correlated risk factors such as stress, marital difficulties, poor-quality
parentœchild relationships, as well as mutual influences between children's behavior and maternal
mood (e.g., Cummings & Davies, 1994; Goodman, Brogan, Lynch, & Fielding, 1993; Hammen,
1991, 1999). The current study does not address these more complex processes and relationships
between maternal depression and child outcome.
The results of the present study should also be interpreted in light of the limitations of the measures.
The measure of maternal depressive symptoms is not well known, which makes it difficult to
compare the results with those of other samples for which more standard measures were used.
Moreover, the relationship of elevated scores to clinical diagnoses and actual duration of depression
remains to be studied. We also acknowledge that because of relatively infrequent testing, especially
in the child's early years, some women might have been called —nondepressed“ who actually
experienced periods of depressive symptoms that were undetected. A further issue is that depressive
symptoms as measured in the present study may not necessarily be specific to depression diagnoses
and, as with any self-report depression scale, may be elevated in the presence of other psychiatric
disorders and even medical problems.
A question can also be raised concerning our focus on the term chronicity and how well our measures
reflect this concept. Chronicity implies that the symptoms the mothers reported are long lasting and
were in place continuously from one follow-up period to the next. In fact, repeated reports of
maternal depressive symptoms in our study might reflect the recurrence of symptoms rather than
their chronicity. This is a particular concern when a long period of time exists between maternal
reports of symptoms, as in the case of the extended time period between the 6-month and the 5-year
follow-ups. In response to this concern, we completed additional analyses of the chronicity of
maternal depressive symptoms that excluded the age 5 maternal symptom data, and our results were
highly similar. Although it is not possible to disentangle chronicity from recurrence with our
available data, it appears that our findings do reflect, in large part, the relationship between chronicity
of maternal depressive symptoms and child outcomes.
An additional methodological weakness of this study is that behavior problem scores based on the
Child Behavior Checklist were provided by the mothers rather than independent judges. Also, at the
age 5 assessment, the behavior problem data were collected at the same time as maternal depressive
self-report scores. Some studies have suggested that depressed mothers exaggerate symptoms in their
children and that the apparent association between maternal and child symptoms might be somewhat
spurious. However, a review of the research by Richters (1992) suggests that there is no clear
evidence for negative distortion or bias as such, and indeed, several studies have shown that
depressed women are actually more accurate in detecting true disorders in their children than are
nondepressed women (e.g., Conrad & Hammen, 1989; Weissman et al., 1987). Thus, mothers' reports
of child functioning should not be discounted–and the finding of somewhat similar patterns of
association between maternal depressive symptoms and children's receptive vocabulary scores
obtained through objective testing further bolsters the validity of the results. Nonetheless, we
acknowledge that multiple informants would have provided a more complete picture of child
functioning, and a replication of these results based on multiple informant data would be an
important goal for future research.
The evaluation of timing of maternal depressive symptoms was limited to those women who
experienced a single elevated depression score so as not to confound timing with chronicity of
depressive symptoms. It may be that the larger issue of timing concerns broad periods such as
preschool, school age, or adolescence and also that more persisting depressive symptoms rather than
a single episode of elevated scores during a particular period might be differentially related to child
Despite these measurement and design shortcomings, the multiple assessments and large sample
allowed some disentangling of the crucial dimensions of maternal depressive symptom severity,
chronicity, and timing. Our results underscore the utility in examining the dimensions of both
chronicity and severity of maternal depressive symptoms as they relate to child behavior problems.
Burlington: University of Vermont, Department of Psychiatry.
Alpern, L. & Lyons-Ruth, K. (1993). Preschool children at social risk: Chronicity and timing
of maternal depressive symptoms and child behavior problems at school and at home. Development
and Psychopathology, 5, 371-387.
Beck, A. T., Steer, R. A. & Garbin, M. G. (1988). Psychometric properties of the Beck
Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8, 77-100.
Beck, A. T., Ward, C. H., Mendelson, M., Mock, J. & Erbaugh, J. (1961). An inventory for
measuring depression. Archives of General Psychiatry, 4, 561-571.
Bedford, A. & Foulds, G. (1977). Validation of the Delusions-Symptoms –States Inventory.
British Journal of Medical Psychology, 50, 163-171.
Bedford, A. & Foulds, G. (1978). Delusions-Symptoms -States Inventory of Anxiety and
Depression.. Windsor, England: National Foundation for Educational Research.
Bor, W., Najman, J. M., Andersen, M. J., O'Callaghan, M. O., Williams, G. M. & Behrens, B.
C. (1997). The relationship between low family income and psychological disturbance in young
children: An Australian longitudinal study. Australian and NewZealand Journal of Psychiatry, 31,
Campbell, S. B., Cohn, J. F. & Meyers, T. A. (1995). Depression in first-time mothers:
Mother-infant interaction and depression chronicity. Developmental Psychology, 31, 349-357.
Campbell, S. B., Cohn, J. F., Meyers, T. A., Ross, S. & Flanagan, C. (1993). Chronicity of
maternal depression and mother-infant interaction.. In D. Teti (Chair), Depressed mothers and their
children: Individual differences in mother-child outcome. Symposium conducted at the meeting of
the Society for Research in Child Development, New Orleans, LA.
Conrad, M. & Hammen, C. (1989). Role of maternal depression in perceptions of child
maladjustment. Journal of Consulting and Clinical Psychology, 57, 663-667.
Achenbach, T. M. (1991). Integrative guide to the 1991 CBCL, YSR, and TRF profiles..
Journal of Child Psychology and Psychiatry, 35, 73-112.
Downey, G. & Coyne, J. C. (1990). Children of depressed parents: An integrative review.
Psychological Bulletin, 108, 50-76.
Dunn, L. & Dunn, L. M. (1981). The Peabody Picture Vocabulary Test–Revised.. Circle
Pines, MI: American Guidance Services.
Frankel, K. A. & Harmon, R. J. (1996). Depressed mothers: They don't always look as bad as
they feel. Journal of the American Academy of Child and Adolescent Psychiatry, 35, 289-298.
Gelfand, D. M. & Teti, D. M. (1990). The effects of maternal depression on children. Clinical
Psychology Review, 10, 320-354.
Ghodsian, M., Zajicek, E. & Wolkind, S. (1984). A longitudinal study of maternal depression
and child behavior problems. Journal of Child Psychiatry, 25, 91-109.
Goodman, S. H., Brogan, D., Lynch, M. E. & Fielding, B. (1993). Social and emotional
competence in children of depressed mothers. Child Development, 64, 516-531.
Hammen, C. (1991). Depression runs in families: The social context of risk and resilience in
children of depressed mothers.. New York: Springer-Verlag.
Hammen, C. (1999). Children of affectively ill parents.. In H. C. Steinhausen & F.Verhulst
(Eds.), Risks and outcomes in developmental psychopathology (pp. 38œ53).Oxford, England: Oxford
Keeping, J. D., Najman, J. M., Morrison, J., Western, J. S., Andersen, M. J. &Williams, G. M.
(1989). A prospective longitudinal study of social, psychological, and obstetrical factors in
pregnancy: Response rates and demographic characteristics of the 8,556 respondents. British Journal
of Obstetrics and Gynecology, 96, 289-297.
Keller, M. B., Beardslee, W. R., Dorer, D. J., Lavori, P. W., Samuelson, H. & Klerman, G. R.
(1986). Impact of severity and chronicity of parental affective illness on adaptive functioning and
psychopathology in children. Archives of General Psychiatry, 43, 930-937.
Lyons-Ruth, K., Connell, D. & Grunebaum, H. (1990). Infants at social risk: Maternal
depression and family support services as mediators of infant development and security of
attachment. Child Development, 61, 85-98.
Murray, L. (1992). The impact of postnatal depression on infant development. Journal of
Child Psychology and Psychiatry, 33, 543-561.
Radke-Yarrow, M., Cummings, E. M., Kuczynski, L. & Chapman, M. (1985). Patterns of
attachment in two- and three-year-olds in normal families and families with parental depression.
Child Development, 56, 884-893.
Radloff, L. S. (1977). The CES-D scale: A new self-report depression scale forresearch in the
general population. Applied Psychological Measurement, 1, 385-401.
Cummings, E. M. & Davies, P. T. (1994). Maternal depression and child development.
of the evidence for distortion. Psychological Bulletin, 112, 485-499.
Sameroff, A. J., Barocas, R. & Seifer, R. (1984). The early development of children born to
mentally ill women.. In N. Watt, E. J. Anthony, L. Wynne, & J. Rolf (Eds.), Children at risk for
schizophrenia (pp. 482œ514). New York: Cambridge University Press.
Steer, R., Beck, A. T., Brown, G. & Berchick, R. (1987). Self-reported depressive symptoms
that differentiate recurrent-episode major depression from dysthymic disorders. Journal of Clinical
Psychology, 43, 246-250.
Teti, D. M., Gelfand, D. M., Messinger, D. S. & Isabella, R. (1995). Maternaldepression and
the quality of early attachment: An examination of infants, preschoolers,and their mothers.
Developmental Psychology, 31, 364-376.
Warner, V., Mufson, L. & Weissman, M. M. (1995). Offspring at high and low risk for
depression and anxiety: Mechanisms of psychiatric disorder. Journal of the American Academy of
Child & Adolescent Psychiatry, 34, 786-797.
Weissman, M., Wickramarante, P., Warner, V., John, K., Prusoff, B. A., Merikangas, K. R. &
Gammon, D. (1987). Assessing psychiatric disorders in children: Discrepancies between mothers'
and children's reports. Archives of General Psychiatry, 44, 747-753.
Whiffen, V. & Gotlib, I. (1989). Infants of postpartum depressed mothers: Temperament and
cognitive status. Journal of Abnormal Psychology, 98, 274-279.
Richters, J. E. (1992). Depressed mothers as informants about their children: A critical review