VOLUME 111, NUMBER 3: 155–169 z MAY 2006 AMERICAN JOURNAL ON MENTAL RETARDATION

Meta-Analysis of Comparative Studies of

Depression in Mothers of Children With and
Without Developmental Disabilities
George H. S. Singer
University of California, Santa Barbara

Abstract
Meta-analysis was used to synthesize findings from comparative studies of depression in
mothers of children with and without developmental disabilities. Effect sizes were deter-
mined for 18 studies conducted between 1984 and 2003. A weighted effect size of .39
indicated an elevated level of depression in mothers of children with developmental dis-
abilities. Planned comparisons found that age of child and disability category moderated
effect sizes. Results show that mothers of children with developmental disabilities are at
elevated risk of depression compared to mothers of typically developing children. Depres-
sion in mothers of children with developmental disabilities is a condition that is presently
not being addressed on a wide scale, although promising interventions are available.

For over 4 decades, researchers have identified These mixed and seemingly contradictory
maternal depression as one problematic response findings are consistent with an emerging under-
to parenting children with disabilities (Cum- standing that there is a wide range of responses to
mings, Bayley, & Rie, 1966; Gath, 1977; Glidden parenting children with developmental disabilities
& Schoolcraft, 2003; Nixon & Singer, 1993; Wol- (Glidden, Kiphart, Willoughby, & Bush, 1993), in-
fensberger, 1967). Several researchers comparing cluding positive outcomes (Hastings & Taunt,
mothers of children with and without disabilities 2002). An earlier assumption of virtually universal
have reported significantly higher levels of de- and unvarying negative impacts on families (Wol-
pressive symptoms in mothers of children with fensberger 1967) has been displaced by a more
developmental disabilities (Amaral, 2003; Beck- complex understanding of family adaptation (Fer-
man, 1991; Breslau & Davis, 1988; Breslau, Sta- gusen, 2001; Glidden et al., 1993; Singer & Irvin,
ruch, & Mortimer, 1982; Kazak, 1987; Kazak & 1991; Turnbull, Turbiville, & Turnbull, 2000) and
Marvin, 1984; Miller, Gordon, Daniele, & Diller, of long-term resilience (Seltzer et al., 2001).
1992; Roach, Orsmond, & Barratt, 1999; Scott,
Despite the acknowledgement of variability in
Atkinson, Minton, & Bowman, 1997; Witt, Riley,
outcomes, there are reasons why it is important
& Coiro, 2003; Wolf, Noh, Fisman, & Speechley,
1989). This evidence, however, is more complex to determine whether there is, on average, an el-
because in several comparative studies, researchers evated level of depression in mothers of children
have also found no significant differences between with developmental disabilities and, if so, the
groups (Barakat & Linney, 1992; Bristol, Gallagh- magnitude of the difference in depression levels
er, & Schopler, 1988; Capelli, 1990; Gowen, John- between mothers of children with and those with-
son, Goldman, & Appelbaum, 1989; Kazak & out disabilities. Foremost among these reasons is
Marvin, 1987; Harris & McHale, 1989; Seltzer, the fact that both minor and major depression are
Greenberg, Floyd, Pettee, & Hong, 2001; Walker, linked to detrimental individual, familial, and so-
Ortiz-Valdes, & Newbrough, 1989). cietal outcomes (Glidden & Schoolcraft, 2003;

q American Association on Mental Retardation 155

VOLUME 111, NUMBER 3: 155–169 z MAY 2006 AMERICAN JOURNAL ON MENTAL RETARDATION

Depression in mothers G. H. S. Singer

Zimmer & Minkovitz, 2003). Even mild depres- in mothers of children with developmental dis-
sion is a cause for concern. People with minor abilities can be effectively reduced with certain
depression have shown considerable limitation kinds of interventions (Hastings & Beck, 2004).
and distress in social, work, and physical func- The weight of the evidence suggests that cognitive
tioning. Hays, Wells, Sherbourne, Rogers, and behavioral treatments with additional support ser-
Spritzer (1995) reported that elevated levels of de- vices are highly promising as ways to reduce ma-
pressive symptoms were associated with lowered ternal distress (Bristol, Gallagher, & Holt, 1993;
well-being, impaired role function, impaired social Hawkins, Singer, & Nixon, 1993; Nixon & Singer,
function, and poor general health. In their longi- 1993; Pelchat, Bisson, Ricard, Perreault, & Bou-
tudinal study of 1,790 persons, they found that chard, 1999; Singer et al., 1994; Singer, Irvin, &
people with minor depression had lower levels of Hawkins, 1988; Singer, Irvin, Irvine, Hawkins, &
functioning than people with serious chronic ill- Cooley, 1989; Singer et al., 1993). We note that
nesses. Major depression is associated with longer all of these interventions brought parents together
lasting and more severe limitations in functioning in groups as sources of mutual support and in-
(Kessler, Zhao, Blazer, & Swartz, 1997). If a group cluded structured content presented by a profes-
of people are at greater risk of depression, they sional facilitator, often psychoeducational mate-
should be an important social concern, deserving rial aimed at improving coping skills. These inter-
of preventive and ameliorative efforts. ventions, however, are not widely available in the
Another reason for addressing the question of United States. They also may not be sufficient for
whether or not mothers of children with disabil- some mothers and, consequently, the service sys-
ities experience more depression on average is that tem may need to make available more intensive
results may help researchers to better understand forms of psychological support to some mothers.
the etiology of depression in women in general. Singer et al. (1993) noted that in their interven-
It is well-established that women of child-bearing tion studies, a small number of mothers did not
age in the general population are at elevated risk respond sufficiently to psychoeducational inter-
for depression (Kessler, 2003). In studies using ventions and needed additional treatment, either
self-report questionnaires to assess prevalence of marital counseling or individual psychotherapy
depression among large samples of women, ap- for depression. The need for these services can be
proximately 20% have mild to major depression partly determined via meta-analysis.
(Lin, Dean, & Ensel, 1988). Investigators using Meta-analysis is a way to synthesize a body of
standardized clinical interviews to determine research literature statistically. It standardizes the
whether women experience depression have difference between groups in each individual
found prevalence rates ranging from 8% to 20% study, allowing them to be combined across stud-
(Kessler, 2003). One hypothesis about the causes ies. It has been particularly useful in medicine,
of depression is that it is, in part, the result of education, and psychology when a body of stud-
elevated stress in daily life. For example, Breslau ies has yielded inconsistent findings when re-
and Davis (1986) used a sample of mothers of viewed by the traditional method of comparing
children with developmental disabilities and the number of studies with significant and non-
chronic illnesses to assess the impact of presumed significant findings (Hedges & Olkin, 1985). This
chronic stress. Their findings were reported in the traditional approach is flawed because statistical
context of contributions to the larger theoretical significance is dependent, in part, on sample size.
question of the relationship between chronic Studies with small samples may fail to detect im-
stress and depression. portant differences, even though large differences
The question is also important for applied exist in the populations and, conversely, those
reasons, specifically, the design of developmental with large samples may find significance for very
disability service system and related social policy. small differences between populations. Further,
If mothers of children with developmental dis- simple counts of the number of studies with and
abilities are at higher risk for depression than without significant findings do not give a sense of
those of children without disabilities, the existing the magnitude of the difference between popula-
social service system for children with develop- tions.
mental disabilities and their families may need to Meta-analysis uses a standardized indicator of
specifically target this problem. Several random- the magnitude of difference between groups, ef-
ized studies have shown that depressive symptoms fect size. One commonly used statistic in research

156 q American Association on Mental Retardation

VOLUME 111, NUMBER 3: 155–169 z MAY 2006 AMERICAN JOURNAL ON MENTAL RETARDATION

Depression in mothers G. H. S. Singer

syntheses is d (Hedges & Olkin, 1985). Two recent moderating variables that partly account for dif-
meta-analyses illustrate the utility of this method ferences in effect sizes between studies. Based on
in synthesizing family research in the develop- theory and previous research, I identified three
mental disabilities field. Risdall and Singer (2004) variables as possible moderators for planned com-
reviewed 13 studies, comparing marital adjust- parisons: (a) year of publication, (b) age of child,
ment in families of children with and without de- and (c) disability label. Year of publication was
velopmental disabilities. The effect size, d 5 .21, chosen as a possible moderator because service
was small, indicating that the impact of raising systems in the United States and Canada have in-
children with developmental disabilities on mari- creasingly focused on family support, so that it is
tal relationships is negative but relatively minor. possible that maternal depression may have de-
Another meta-analysis by Rossiter and Sharpe clined in prevalence between 1982 and 2003. I
(2001) synthesized comparative studies on the im- selected child age as a possible moderator because
pact on siblings of having a brother or sister with recent research suggests that maternal depression
mental retardation. They found an overall effect declines over time (Glidden & Schoolcraft, 2003)
size of d 5 2.06, indicating that the overall im- and disability category because there is some evi-
pact across 79 comparisons was also negative but dence that parents with children who have autism
very small. These results help to clarify contradic- have unusually high levels of stress (Moes, Koegel,
tory outcomes between individual studies, and Schreibman, & Loos (1992).
they make it possible to generally characterize the The fourth and final question is a technical
strength of the impact of children with develop- one that is important in meta-analytic studies.
mental disabilities on their families. Both of these There is concern that studies showing significant
results contradict earlier historical assumptions of differences are more likely to be published than
more substantial negative impacts of children with are studies with nonsignificant findings, possibly
developmental disabilities on family relationships biasing the outcome. It is possible that there are
while also indicating that there is legitimate cause a number of unpublished studies in which re-
for concern about some couples and, to a lesser searchers did not find any significant differences,
extent, siblings in these families (Risdall & Singer, and these studies are languishing in file drawers.
2004). A statistical method, the fail-safe statistic, has
My purpose in conducting this study is to use been developed to determine how many such un-
meta-analysis to answer questions about the prev- published studies would be needed to reduce the
alence of depression in mothers of children with average effect size between groups in published
and without disabilities. My first research question studies to a negligible level (Hedges & Olkin,
is, Across studies, what is the difference in the 1985). The higher the fail-safe number of studies,
average level of depressive symptoms in mothers the less likely it is that the effect size misrepresents
of children with developmental disabilities com- the data because of publication bias. Therefore,
pared to mothers of nondisabled children as in- the fourth question is, how many unpublished
dicated by the effect size, d ? studies would need to exist to reduce the effect
Effect sizes are difficult to understand intui- size to a negligible number?
tively. In order to provide more information
about the differences between the two groups, one Method
way to give a sense of the size of these differences
is to look at simple percentage comparisons. For Search Procedures
the most commonly used measures of depressive I used the four approaches recommended by
symptoms, cut-off scores have been established Cooper (1989) to search for studies. Journals were
such that scores over a specified number indicate searched electronically using the databases
the likely existence of either minor or major de- PsychINFO, PubMed, and ERIC and unpublished
pression, permitting a comparison of the percent- dissertations were sought from the Dissertation Ab-
age of parents with depression in the two groups. stracts International database. The search was lim-
The second question is, What percentage of ited to studies published in the United States and
mothers in the two groups have depression when Canada because one research question concerned
it is measured as a dichotomous variable (de- whether there have been significant changes over
pressed or nondepressed)? time in effect sizes that might be associated with
The third question concerns whether there are the increase in early intervention and family sup-

q American Association on Mental Retardation 157

VOLUME 111, NUMBER 3: 155–169 z MAY 2006 AMERICAN JOURNAL ON MENTAL RETARDATION

Depression in mothers G. H. S. Singer

port services in these two countries during the ulation norms on standardized measures were ex-
past 20 years. My working definition of develop- cluded.
mental disability was a modification of the defi- Studies also were excluded if (a) sufficient sta-
nition in the Developmental Disabilities Act (PL tistical data were not available in order to calculate
98–527). Developmental disabilities were defined as an effect size, (b) if the samples combined did not
occurring before age 21, likely to continue indef- disaggregate data from mothers and fathers, and
initely, require professional services of lifelong du- (c) if they were later determined to be outliers be-
ration, and result in functional limitations in at cause of highly unusual effect size values (Hedges
least three major areas of life activities. I narrowed & Olkin, 1985).
this broad definition by including only those de- In the kind of meta-analysis used in this pa-
velopmental disabilities associated with mental re- per, namely, fixed-effect modeling, an assumption
tardation or a combination of intellectual and is that effect sizes are independent. Some inves-
physical impairments (i.e., physical disabilities tigators reported on the means and SDs for more
and chronic illnesses not associated with cognitive than one disability comparison group. In these in-
disabilities were excluded). Consequently, search stances, the depression scores were averaged across
terms for all databases included all possible com- groups with different developmental disability la-
binations of the terms disability, developmental dis- bels. Similarly, data from longitudinal studies
ability, handicap, autism, mental retardation, cerebral were averaged across time to yield one between-
palsy, traumatic brain injury, and spina bifida with groups comparison.
stress, depression; or distress with parents, mothers,
and families. Databases were searched for all years Statistical Analysis
available. When a journal was identified with ar- I calculated effect sizes for the differences be-
ticles that met the eligibility criteria, a title search tween the two groups for the individual studies
was conducted for all issues of that journal. Titles and an overall mean effect size, d, was produced,
and abstracts were read for all studies that includ- as recommended by Hedges and Olkin (1985).
ed the search terms. Studies reporting means and SDs were converted
All promising articles, dissertations, and book via the equation
chapters were obtained and reviewed. References
from these studies were searched, in turn, to iden- X ti 2 X ci
di 5 ,
tify other studies. References from reviews of the si
literature published as chapters in books were also
searched (Singer & Irvin, 1991). Finally, authors where d is the effect size, Xt is the mean of the
who had published two or more studies on the disability group, Xc is the mean of the nondisa-
topic were contacted by e-mail and asked whether bility group, and s is the pooled SD of the two
they had additional published or unpublished groups. Hedges and Olkin (1985) modified this
data sets or knew of other studies. In addition, effect size slightly by multiplying by a small cor-
authors were contacted by letter or e-mail when rection factor that reduced bias. In the kind of
further data were needed in order to determine an effect size analysis used in this study, a fixed ef-
effect size or clarify possible errors in published fects model, individual ds are averaged to deter-
data. mine an overall effect size. If researchers reported
To be included in the meta-analysis, research- frequencies or proportions, those were converted
ers had to have collected data on depressive symp- to an effect size via the equation d 5 (Pe 2 Pc)/
toms or general psychological distress, including spooled, where d is the study effect size, Pe is the
depression, by using published standardized self- proportion in the experimental group (mothers of
report measures with well-established psychomet- children with disabilities), Pc is the proportion in
ric properties. Studies with researcher-created mea- the comparison group (mothers of children with-
sures with unknown psychometric properties were out disabilities), and spooled is the pooled SD of the
not included. two groups. I calculated effect sizes for ANOVA F
Studies had to include mothers of children tests using the formula d 5 {F[(ne 1 nc)/(nenc)]}1/2
with developmental disabilities and a comparison where F is the ANOVA outcome, ne is the number
group of mothers of children without disabilities. of subjects in the disability group, and nc is the
Studies in which researchers compared one group number of subjects in the nondisability group
of mothers of children with disabilities with pop- (Hedges & Olkin, 1985).

158 q American Association on Mental Retardation

VOLUME 111, NUMBER 3: 155–169 z MAY 2006 AMERICAN JOURNAL ON MENTAL RETARDATION

Depression in mothers G. H. S. Singer

Studies with larger samples are likely to pro- indicates that the participants in all of the studies
duce more accurate results than those with smaller can be treated as one large pooled sample.
samples. Thus, individual effect size statistics need Planned comparisons. Rosenthal, Rosnow, and
to be refined by giving more weight to larger stud- Rubin (2000) presented a procedure for conduct-
ies (Hedges & Olkin, 1985). Further, as discussed ing planned comparisons in meta-analysis in order
below, studies by researchers using better sam- to identify significant moderating variables. Three
pling procedures and designs and those reporting comparisons were planned based on theory or
key variables are also likely to better approximate previous research findings. I hypothesized that the
the population effect size. Consequently, a year of publication would be associated with a
weighted effect size was calculated to account for downward trend, such that effect sizes would be
both sample size and quality (Shadish & Had- smaller the more recent the publication date. A
dock, 1994). The following formula was used to second planned comparison was based on the hy-
calculate an overall effect size weighted by both pothesis that studies with samples of parents of
sample size and quality: children with autism would have larger effect sizes
than studies of parents of children with mental

O wqT k

i i i
retardation or spina bifida. Finally, I conducted a
planned comparison to test the hypothesis that
T 5
i51

i5Owq
e , k
parents of younger children would have higher
i i
levels of depressive symptoms than mothers of
i51 adults with developmental disabilities. The for-
mula I used to conduct planned comparisons as
where Te is the estimate of the population effect recommended by Rosenthal, Rosnow, and Rubin
size, wi is the inverse of the variance for each ef- (2000) is

O Mili 4
fect size, Wi is a score on a quality index, and Ti
is the individual effect size for each study (Shadish  1 
Fcontrast 5  
O
& Haddock, 1994). The inverse of the variance w MSwithin li

was calculated using the sample sizes ne and nc for  
each study.  ni 
To determine whether all samples (effect siz-
es) were drawn from a common population of ef- in which Mi is the mean for the ith group of effect
fect sizes with a shared mean, I performed a test sizes, li is a weight assigned to each group of ef-
of homogeneity. This Q statistic was calculated fect sizes based on the hypothesis or previous re-
using Hedges and Olkin’s (1985) formula: search findings, MSwithin is the mean sum of
squares within groups, and ni is the number of
effect sizes that make up a group. For example, a

1O w T 2
k 2
planned comparison can be developed based on

O wT
i i
k i51 the hypothesis from previous research findings
Q5 2
O
i i
2
. k that studies of parents of children with autism will
i51
w i
have significantly higher effect sizes than will
i51 studies of parents of children with mental retar-
dation or studies of parents of children with spina
The Q statistic is tested via traditional inferential bifida. The effect sizes are then aggregated into
methods. If this homogeneity test produces a re- three groups, one for each disability category. In
sult that is significant at the .05 level (of chi- order to model the hypothesis, lambda weights of
square at k 2 1 dfs), the individual studies are 12, 21, and 21 were assigned to the groups, re-
considered to be more variable than expected due spectively, thereby weighting the autism group
to sampling error alone; thus, if the Q is signifi- higher than the other two groups, which are as-
cant, a 5 .05, the effect sizes are heterogeneous sumed to be equivalent. Fcontrast is based on tradi-
and cannot be assumed to have been drawn from tional ANOVA and is analogous to it.
the same population of studies. It is important to Fail-safe statistic. The formula used to calculate
note that the homogeneity of variance refers to the fail-safe number is k0 5 k(d 2 dc)/dc in which
the differences between studies, not subjects with- k0 is the number of studies with null results need-
in studies. In effect, homogeneity across studies ed to render the average effect as negligible and k

q American Association on Mental Retardation 159

VOLUME 111, NUMBER 3: 155–169 z MAY 2006 AMERICAN JOURNAL ON MENTAL RETARDATION

Depression in mothers G. H. S. Singer

the number of studies in the synthesis, d 5 the method of obtaining the child samples; namely,
overall average effect size, and dc is the effect size whether they were convenience groups or based
at a negligible level; in this analysis it was set at on larger population-sampling methods. If studies
.01 (Hedges & Olkin, 1985). drew both samples of children from large popu-
To conduct the effect size analyses for this lation samples using methods of random selec-
study, I utilized a commercially available com- tion, they were coded with a 1 and those with
puter program, METAWIN (Rosenberg, Adams, convenience sample were assigned a 0.
& Gurevitch, 1997) specifically designed for per- Another important quality dimension had to
forming meta-analyses. METAWIN does not in- do with whether researchers tested for demograph-
clude the procedure for calculating the quality ic differences between samples. If significant dif-
weighted effect size, so a program developed with ferences were identified, studies were rated for
the assistance of a statistician was used. whether or not these differences were statistically
Weighting effect sizes for quality. To deal with controlled. For studies in which researchers either
variability in the quality of studies, I assigned established equivalence of the two groups or sta-
quality weightings. The process of developing and tistically controlled for differences in demograph-
applying these weights is, in fact, a detailed effort ic variables a 1 was assigned; if not, a 0.
to analyze and summarize methodological rigor. Additional quality variables had to do with
Following a validity framework recommended by the level of peer review applied to each study and
Lipsey (1994), quality weights were determined for the dependent measures. Doctoral dissertations
each effect size based on (a) suggestions from nar- and presentations were assigned a 0 and publica-
rative reviews and other relevant literature, (b) tions in peer-reviewed journals, a 1. The last qual-
conceptual and methodological considerations, ity variable was the construct validity and psycho-
and (c) key variables identified in the literature metric maturity of the measures used in these
that might confound the relationship between studies. Widely used and psychometrically ‘‘ma-
groups and the dependent measures. Using these ture’’ instruments were assigned a weight of 1.
criteria, I rated several variables for quality. These measures were the Beck Depression Inven-
The first concern was the extent of demo- tory (Beck, Steer, & Garbin, 1988); the Center for
graphic information provided about the mothers Epidemiology Depression Scale (Radloff, 1977);
in both groups in each study. Several demograph- the Langer Symptom Checklist (Langer, 1962),
ic variables correlated significantly with the prev- the Brief Symptom Inventory (Derogatis & Meli-
alence of elevated levels of depressive symptoms saratos, 1983), and the Symptom Checklist 90 Re-
in large population studies (Lin et al., 1986). The vised (Derogatis, 1977). Researchers using instru-
variables were (a) family income, (b) employment, ments that have been challenged for their con-
(c) race or ethnicity, (d) marital status, and (e) ed- struct validity as a measure of depressive symp-
ucational attainment. Each identified demograph- toms or those with psychometric properties that
ic variable was assigned a weight of either 1 if are unknown were assigned a 0. The Parent Stress
reported or 0 if not reported. Index Depression subscale (Abidin, 1995) was as-
A similar methodological concern had to do signed a 0 because the author of the instrument
with the amount of descriptive information re- asserted in the manual that this subscale was best
ported about the children in both groups and the understood as a measure of a parent’s ability to
extent to which the samples had homogeneous marshal energy for parenting tasks rather than as
diagnostic categories. The following variables were traditional depression. In a study by Witt et al.
judged as important based on a reading of the (2003), an index consisting of three questions
literature: (a) child’s age, (b) child’s primary di- from the National Health Information Survey was
agnosis, and (c) an indicator of the severity of the also rated as 0 because no psychometric properties
condition. Studies were assigned a weight of 1 for of this index were available.
each reported variable or 0 if not reported. Child Percentage comparisons. To aid interpretation of
characteristics were coded by assigning a 0 for the effect sizes, I also used an additional ap-
samples that combined children who had devel- proach, namely, comparisons of the percentages
opmental disabilities with children who had of mothers scoring above or below clinical cutoff
chronic illnesses and a 1 for studies with samples scores for depression. In order to understand this
solely made up of children with developmental approach, it is important to note that depression
disabilities. Quality also varied according to the is sometimes reported as a continuous variable

160 q American Association on Mental Retardation

VOLUME 111, NUMBER 3: 155–169 z MAY 2006 AMERICAN JOURNAL ON MENTAL RETARDATION

Depression in mothers G. H. S. Singer

and sometimes as a dichotomous one (depressed (Amaral, 2003; Cappelli, 1990). I rejected Amar-
or not depressed). Two of the measures most com- al’s study as an outlier because Amaral reported
monly used in survey studies of depression, the an effect size eight times larger than the highest
Center for Epidemiology Depression Scale and effect size in the other 18 studies (Hedges & Olk-
the Beck Depression Scale, have established clin- in, 1985). Subsequently, 18 studies were included
ical cutoff scores for determining whether or not in the analyses. Table 1 lists each of the studies
a person falls into the category of having depres- alphabetically. It includes data on the authors and
sion. These cut-off scores are based on studies publication dates and the measure of depression,
comparing scores on self-report measures with re- the means and SDs of the depression scores or
sults of psychiatric interviews. Percentage of par- the percentage over a clinical cut-off score, quality
ticipants in each group falling at or above the clin- weightings, effect sizes, and 95% confidence in-
ical cutoff score of 16 and 10, respectively, were tervals. As noted in the footnotes, when mean
calculated and compared. Because most of the re- scores and SDs were not available, I calculated ef-
searchers in this synthesis did not report on de- fect sizes from percentages, F tests, t tests, or chi-
pression as a dichotomous variable, it was neces- square tests as reported in the studies.
sary to extrapolate from the data on means and Table 2 presents demographic information,
SDs by first standardizing depression scores with including the mothers’ education levels, mothers’
the standardized z statistic and then finding the age, ethnicity or race, family income or SES, and
percentage of subjects above and below these cut- marital status. For the children age and disability
off scores using a z distribution table. The resul- categorical diagnosis were included.
tant percentages are approximations based on the
assumption that scores in the studies were nor- Homogeneity of Variance
mally distributed. Although percentages provide When the 18 studies were analyzed for ho-
only a rough indicator, they may help to make mogeneity of variance, the Qw statistic was not sig-
intuitive sense of the effect sizes. In two studies nificant, indicating that all of the effect sizes
(Blacher, Lopez, Shapiro, & Fusco, 1997; Scott et could be combined and the participants treated as
al., 1997), the researchers did report on the per- one pooled sample consisting of 6,641 parents of
centages of subjects above and below clinical cut- children with disabilities and 26,438 parents of
off scores, and the published data rather than z children without disabilities.
scores were used.
Reliability. To determine interobserver agree- Effect Size Estimates
ment on quality weightings, a second reader cod- The average unweighted d for the 18 studies
ed 100% of the studies. Percentage of interobserv- was .35, with the 95% confidence interval of
er agreement was calculated by dividing the total 1.29/1.39. When weighted by quality indicators
number of agreements by agreements plus dis- and the reciprocals of the variances, the weighted
agreements and multiplying by 100. In instances effect size, d, was .39,. with a 95% confidence in-
of disagreement, the author and the observer dis- terval of 1.31/1.47. Effect sizes of this magni-
cussed the texts and arrived at an agreed upon tude are, as a rough rule of thumb, considered
correct data point for final calculation of the over- between small and moderate (Cohen, 1988).
all weighted effect size.
Fail-Safe Number
Results The fail-safe calculations indicated that 34
Sample of Studies studies with null results would need to exist in
In an initial search of the categories that yield- order to reduce the average d to a negligible level.
ed the largest number of studies (family and dis-
ability as title words), I found 276 articles in the Percentage With Elevated Levels of
PsychInfo database, 39 in the Dissertation Ab- Depressive Symptoms
stracts database, and 43 in ERIC. Of these 358 A common sense, albeit crude, indicator of
articles, 19 met the inclusion criterion. Sixteen the difference between categorical variables is a
were published in peer-reviewed journals, 1 was a simple contrast of percentages. Table 3 presents
paper presented at a conference (Seltzer et al., the percentage of mothers in each group whose
1989), and 2 were unpublished dissertations scores fell on or above the clinical cut-off of 16

q American Association on Mental Retardation 161

VOLUME 111, NUMBER 3: 155–169 z MAY 2006 AMERICAN JOURNAL ON MENTAL RETARDATION

Depression in mothers G. H. S. Singer

Table 1. Outcomes of Studies Comparing Mothers of Children With and Without Developmental
Disabilities Including Effect Size and Quality Ratings

Quality 95%
n Mean (SD) weight- Effect confidence
Study Measurea With Without With Without ing size interval
1. Barakat & BSI 29 28 1.88 1.84 10 .28 2.24/1.80
Linney (.13) (.15)
(1992)
2. Beckman PSI-D 27 27 25*** 20.5 10 .60 1.06/11.15
(1991) (7.8) (6.6)
3. Blacher et CES-D 148 101 50%* 33% 11 .35c .09/.60
al. (1997)
4. Breslau et CES-D 310 357 11.76*** 8.31 10 .38 1.22/1.53
al. (1988) (9.87) (8.25)
5. Breslau et Langer 369 450 1.8**** 1.4 9 .29 1.15/1.42
al. (1982) (1.5) (1.3)
6. Bristol et al. CES-D 31 25 14.39 9.12 11 .54 1.008/11.08
(1988) (10.66) (7.93)
7. Capelli CES-D 46 46 10.6 9.0 9 .19 2.21/1.60
(1990) (9.1) (7.2)
8. Gowen et CES-D 21 20 10.2 8.0 9 .27 2.35/1.88
al. (1989) (6.5) (9.3)
9. Harris & BDI 30 30 5.77 3.87 10 .40 2.11/1.91
McHale (5.0) (4.24)
(1989)
10. Kazak Langer 125 127 4.21* 3/16 3 .39 1.14/1.64
(1987) (3.19) (2/1)
11. Kazak & Langer 53 53 — — 8 .50b 1.11/1.89
Marvin
(1987)
12. Miller et al. BSI 69 63 54.6*** 51 10 .37 1.02/1.71
(1992) (10.74) (8.68)
13. Roach et al. PSI-D 41 58 18.88* 16.84 8 .48 1.069/1.88
(1999) (4.5) (4.1)
14. Scott et al. BDI 46 46 5.66* 4.63 9 .36b 1.05/1.77
(1997)
15. Selzer et al. CES-D 92 126 10.1 9.1 11 .13 2.14/1.40
(1998) (8.8) (7.1)
16. Walker et CES-D 24 24 9.67 10.12 8 2.04 2.50/1.42
al. (1989) (9.75) (10.08)
17. Witt et al. NHIS 5,089 24,820 17.6%* 6.5% 11 .40d 1.37/.43
(2003)
18. Wolf et al. BDI 61 31 9.25* 6.06 6 .45 1.36/1.1
(1989) (7.36) (6.36)
a
BSI 5 Brief Symptom Inventory, PSI-D 5 Parent Stress Inventory-Depression subscale, CES-D 5 Center for Epide-
miological Studies-Depression Scale, Langer 5 Langer Symptom Inventory, BDI 5 Beck Depression Inventory, NHIS
5 National Health Interview Study. bEffect size calculated on F score. cCalculated from chi-square, (2, N 5 249) 5 7.63,
p , .02. dCalculated based on %.
*p # .05. **p # .01. ***p # .001.

162 q American Association on Mental Retardation

VOLUME 111, NUMBER 3: 155–169 z MAY 2006 AMERICAN JOURNAL ON MENTAL RETARDATION

Depression in mothers G. H. S. Singer

from the studies utilizing the Center for Epide- ers of all children in the United States has been
miology Depression Scale and 10 on the Beck De- recognized as an important public health problem
pression Inventory. Table 3 shows that, on aver- (National Institute on Child Health and Human
age, 29% of mothers of children with develop- Development, 1999). The additional prevalence in
mental disabilities had elevated levels of depres- mothers parenting a child with a developmental
sive symptoms compared to 19% of parents in the disability is cause for further concern. Almost one
comparison group, a difference in prevalence of third, 29%, of the mothers of children with de-
10%. velopmental disabilities in these studies experi-
enced depression. These results are consistent
Planned Comparisons and with research suggesting that increased environ-
Interobserver Reliability mental stress is linked to depression (Witt et al.,
Three planned comparisons were conducted 2003).
in order to identify moderating variables. Suffi- The finding that effect sizes did not decrease
cient data and a rationale for planned compari- over the period from 1982 to 2003 suggests that
sons warranted comparisons for date of publica- increases in services during this period of time did
tion, child’s age, and disability category. The com- not impact depression, although they are likely to
parison for year of publication yielded a nonsig- have had other benefits. Future researchers should
nificant result, indicating that the year in which include measurements of service utilization and
the study was conducted was not a moderator of satisfaction in order to better understand the re-
effect size differences between studies. The com- lationship between services, their stated missions,
parison of effect sizes by child age was significant, and maternal satisfaction with them. These vari-
F(2,15) 5 11.99, p 5 .001, indicating a significant ables were not assessed in any of the studies under
difference between effect sizes in studies of par- review.
ents of adult children and effect sizes from studies These meta-analytic findings help to support
of parents of children in middle and early child- the argument that there is need for more widely
hood. The planned comparison for disability cat- available interventions designed to prevent and
egory indicated that in studies focused on parents treat depression in mothers of children with de-
of children with autism, researchers found higher velopmental disabilities. The fact that a large ma-
effect sizes compared to studies of parents of chil- jority of mothers in these studies were not de-
dren with mental retardation and studies of par- pressed suggests that depression is not an inevi-
ents of children with spina bifida, F(2, 15), p 5 table consequence of parenting children with de-
.005. velopmental disabilities. Hastings and Beck (2003)
Interobserver agreement on the quality reviewed the literature on interventions for stress
weighting was 93.6% (range 5 80% to 100%). in this population of mothers and concluded that
group interventions using cognitive–behavioral
stress management methods were promising. Sing-
Discussion er et al. (1989) reported that half of the parents
It is clear from these data that mothers of chil- in their intervention study of stress management
dren with disabilities are at elevated risk of de- combined with additional support services
pression compared to the population of mothers showed clinically significant improvements in
of nondisabled children sampled in these studies. mental health. Promising interventions have all
Small to moderate effect sizes were relatively con- involved professionally led psychoeducational
sistent over the period from 1982 through 2003. groups in which parent-to-parent support was en-
When examined in terms of percentages of wom- couraged. Attention should be given to replicating
en with scores over clinical cutoffs, there was a intervention studies for the purpose of establish-
relatively consistent increase in prevalence of ing treatments of choice and confirming efficacy
roughly 10%. Although women of child-bearing and generalizability. The problem appears to be
age in general are at high risk of depression, the widespread, suggesting that intervention studies
addition of extra challenges associated with care- should be tested on a large scale as part of com-
giving further raises the risk for psychological dis- monly available service systems. Group stress
tress. The prevalence of elevated levels of depres- management interventions may not be sufficient
sive symptoms in the general population of moth- for some parents who are likely to require more

q American Association on Mental Retardation 163

 164
Table 2. Descriptive Information on Mothers and Children With Disabilities
VOLUME

Mothers Children
111,

Age Marital Age

Study mean (SD) Educationa Income mean (SD) Race/Ethnicity status (%) Disabilityb mean (SD)
NUMBER

1 34.90 12.38 34.48 3% minority, 97% SB 8.45

Depression in mothers

(1.61) (12.73) White (1.74)

2 32.00 15.30 50.84 23% White All married 10, CP; 2, autism 3.80
(4.90) (3.0) 3% African American 5, MD; 10, DD (1.27)
3: 155–169 z

3 Low SESc Latina immigrants Married 88.5 MR 11.3

MAY

Single 21.5 (4.2)

4 41.00 12.00 Income: White 79% Married 68 CP, SB, CF, MD Range 8–23
2006

(7.90) (2.30) ,15,000, 33% Black 21% Single 32

,20,000, 12.5%
,25,000, 12.5%
.25,000, 42%
5 37.00 11.90 $17 76% White Married 75 CF, CP, MD 10.10
(2.40) Single 25 (4.50)
6 30.47 5 41.63 White 76% Autism 3.82
(4.86) (13.59) African American (1.04)
24%
7 12.80 HS, 56%; $40,000–$50,000 Canadian All married Spina bifida 7.25
(5.40) Some jr. college, 13% (4.35)
Jr. college, 6%
Some university, 4%
University, 17%
Graduate, 2%
6 post HS
6, college
7, graduate
8 31.15 18% White Married 20 7 DS, 2 DD, 4 CP; 1.50
(6.64) Single 1 8 DD
(Table 2 continued)
AMERICAN JOURNAL ON MENTAL RETARDATION

G. H. S. Singer

q American Association on Mental Retardation

 VOLUME

Table 2. Continued
Mothers Children
111,

Age Marital Age

Study mean (SD) Educationa Income mean (SD) Race/Ethnicity status (%) Disabilityb mean (SD)
NUMBER

Depression in mothers

9 34.52 $15,000–$30,000 All married MR 7.70

(7.93)
10 All married 36 MR; 43 PKU Range 9–30
3: 155–169 z

46 SB
11 35 $17.90 85 married; 15
MAY

single
12 32 Employed, 42%; ,$25e, 4 orthopedic, 4
2006

21%; $25,000–39,900, quadriplegia, 3

13%; $40,000–59,900, SB, 2 TBI

q American Association on Mental Retardation

24%; $60,000–above,
42%
13 35.80 50.51 All married DS 3.50
(5.70) (11.00) (0.80)
14 Middle class DS 1.18
15 12.90 $40,850 82 married DD adults .30.00
18 single
16 Middle class 71 married 24 MR 8–19d
29 single
17 Some HS, 20% 20% below poverty White 69.70% 65 married Limited function- 6–10, 41%
HS grad, 38.9% threshold Black 17.30% 35 single ing (DD and CI) 11–14, 35.2%
Some college, 24.90% Hispanic 10.60% 15–17, 23.4%
Other 2.40%
18 35.50 13.2 $22,400 30 autism, 31 MR 9.20
(10.40) (4.70) ($5,800) (4.20)
a
HS 5 high school. bSB 5 spina bifida, CP 5 cerebral palsy, DD 5 developmental disability or delay, MR 5 mental retardation, MD 5 multiple disabilities, CF 5
cystic fibrosis, TBI 5 traumatic brain injury. cMedian Hollingshead category. dFathers and mothers’ scores were not disaggregated. eReported in thousands.
AMERICAN JOURNAL ON MENTAL RETARDATION

G. H. S. Singer

165
VOLUME 111, NUMBER 3: 155–169 z MAY 2006 AMERICAN JOURNAL ON MENTAL RETARDATION

Depression in mothers G. H. S. Singer

Table 3. Mothers With and Without Children With Developmental Disabilities Scoring Above Clinical
Cut-Off Scores on the Depression Measures
Mean (SD) Z score .Cut-Off (%) Difference
Measure /Study
a
With Without With Without With Without (%)
BDI
Harris & McHale, (1989) 5.77 (5.0) 3.87 (4.24) .85 1.45 20 7 13
Wolf et al. (1989) 9.24 (7.27) 6.05 (5.88) .92 1.69 17 4 13
28 16 6
Scott et al. (1997) 5.66 4.63 39 20 19
CES-D
Blacher et al. (1997) 50 30 20
Breslau et al. (1986) 11.76 (9.87) 8.31 (8.25) .42 .93 33 17 16
Bristol et al. (1988) 14.39 (10.66) 9.12 (8.25) .15 .83 44 20 24
Capelli (1990) 10.6 (9.1) 9.00 (7.20) .59 .97 27 16 11
Gowen et al. (1989) 10.20 (8.00) 9.26 (6.5) .63 1.23 26 11 15
Selzer et al. (1989) 10.1 (8.8) 9.1 (7.1) .68 .97 24 17 7
Walker et al. (1989) 10.46 (8.83) 10.12 (10.08) .66 .58 25 28 23
a
BDI 5 Beck Depression Inventory, cut-off score .10. CES-D 5 Center for Epidemiology Depression Scale, cut-off
.16.

individualized and intensive forms of treatment dependent variables. Only two studies drew from
(Singer et al., 1989). large population samples (Seltzer et al., 2001; Witt
Using a planned comparison approach, I et al., 2003), whereas the others relied on samples
found significantly higher effect sizes in the two of convenience. Although difficult, future inves-
studies that disaggregated data on parents of chil- tigators should draw from population samples
dren with autism. Higher levels of depression as- that avoid the problems of external validity raised
sociated with this condition again raise questions by use of convenience samples.
about the need for specific services targeted at this Further, future researchers should compare
population as well as further research to identify depression in mothers from minority ethnic and
more specifically why there is more distress in this linguistic groups, including recent immigrants.
group. The finding that parents of adult children With the exception of a few studies, the subjects
had a significantly lower effect size than did par- were predominantly middle-class White women.
ents of children in early and middle childhood is Authors who did use more diverse samples did
consistent with research suggesting that maternal not disaggregate the data by ethnicity. Blacher et
distress gradually decreases over time (Glidden & al. (1997) found high levels of depression in La-
Schoolcraft, 2003). It is important to note that tina mothers of children with developmental dis-
there was only one available comparative study abilities, suggesting that further study of minority
with a group of parents of adults, and there were populations, particularly recent immigrants, is
no studies of adolescents that could be included warranted.
in the planned comparison. The possibility of an Finally, fully 71% of mothers of children with
age-related trend should be investigated. developmental disabilities in the studies analyzed
The quality of these studies was highly vari- in this paper did not experience problems with
able. In the rating system I used to develop a qual- depression. It is remarkable how little is known
ity index for each study, the highest possible score about the large majority of mothers of children
was a 13. The average quality rating across the 18 with developmental disabilities who are faring as
studies was 8.9 (range 5 3 to 12). Authors of sev- well, on average, as the rest of the women of
eral studies failed to report important demograph- child-bearing age in the United States. The find-
ic information that should be routinely included ing that roughly 29% of mothers experience ele-
in future research along with means and SDs of vated levels of depressive symptoms should serve

166 q American Association on Mental Retardation

VOLUME 111, NUMBER 3: 155–169 z MAY 2006 AMERICAN JOURNAL ON MENTAL RETARDATION

Depression in mothers G. H. S. Singer

as a caution against blanket assumptions about Adaptation and spousal support. Developmen-
negative impacts of parenting a child with dis- tal Psychology, 24, 441–451.
abilities while making it clear that a serious prob- *Cappelli, M. (1990). Marital interaction of cou-
lem needs to be addressed. ples with children with spina bifida: A case-
control study (UMI No. NN68849). Disserta-
tion Abstracts International, 53(03), 4366.
References1 Cohen, J. (1988). Statistical power analysis for the
behavioral sciences (2nd ed.). New York: Aca-
Abidin, R. R. (1995). Parenting Stress Index profes- demic Press.
sional manual (3rd ed.). Odessa, FL: Psycho- Cooper, H. (1989). Integrating research: A guide for
logical Assessment Resources. literature reviews. Newbury Park, CA: Sage.
Amaral, R. (2003). How do children with devel- Cummings, S. T., Bayley, H. C., & Rie, H. E.
opmental disabilities impact their parents’ (1966). Effects of the child’s deficiency on the
parent satisfaction, self-esteem, symptoms of mother: A study of mentally retarded, chron-
stress, ways of coping, marital satisfaction and ically ill, and neurotic children. American Jour-
family support. Dissertation Abstracts Interna- nal of Orthopsychiatry, 36, 395–408.
tional, 64(04), 1183. (UMI No. 3089126) Derogatis, L. (1977). SCL-R-90 Version: Manual-I.
*Barakat, L. P., & Linney, J. A. (1992). Children Baltimore: Johns Hopkins University.
with physical handicaps and their mothers: Derogatis, L. R., & Melisaratos, N. (1983). The
The interaction of social support, maternal Brief Symptom Inventory: An introductory
adjustment, and child adjustment. Journal of report. Psychological Medicine, 13, 595–605.
Pediatric Psychology, 17, 725–739. Fergusen, P. (2001). Mapping the family: Disabil-
Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). ity studies and the exploration of parental re-
Psychometric properties of the Beck Depres- sponse to disabilities. In G. Albrecht, K. Seel-
sion Inventory: Twenty-five years of evalua- man, & M. Bury (Eds.), Handbook of disability
tion. Clinical Psychology Review, 8, 77–100. studies (pp. 373–395). Thousand Oaks, CA:
*Beckman, P. J. (1991). Comparison of mothers’ Sage.
and fathers’ perceptions of the effect of young Gath, A. (1977). The impact of an abnormal child
children with disabilities. American Journal on upon the parents. British Journal of Psychiatry,
Mental Retardation, 95, 585–595. 130, 405–410.
*Blacher, J., Lopez, S., Shapiro, J., & Fusco, J. Glidden, L. M., Kiphart, M. J., Willoughby, J. C.,
(1997). Contributions to depression in Latina & Bush, B. A. (1993). Family functioning
mothers of with and without children with when rearing children with developmental
retardation: Implications for caregiving. Fam- disabilities. In A. P. Turnbull, J. M. Patterson,
ily Relations, 46, 325–334. S. K. Behr, D. L. Murphy, J. G. Marquis, &
*Breslau, N., & Davis, G. C. (1986). Chronic M. Blue-Banning (Eds.), Cognitive coping, fam-
stress and major depression. Archives of Gen- ilies, and disability: Participatory research in ac-
eral Psychiatry, 43, 309–314. tion (pp. 183–194). Baltimore: Brookes.
*Breslau, N., Staruch, K. S., & Mortimer, E. A. Glidden, L. M., & Schoolcraft, S. A. (2003). De-
(1982). Psychological distress in mothers of pression: Its trajectory and correlates in moth-
disabled children. American Journal of Diseases ers rearing children with intellectual disabili-
of Children, 136, 682–686. ties. Journal of Intellectual Disability Research.
Bristol, M. M., Gallagher, J. J., & Holt, K. D. Special Issue on Family Research, 47(4–5), 250–
(1993). Maternal depressive symptoms in au- 263.
tism: Response to psychoeducational inter- *Gowen, J. W., Johnson-Martin, N., Goldman, B.
vention. Rehabilitation Psychology, 38, 3–10. D., & Appelbaum, M. (1989). Feelings of de-
*Bristol, M. M., Gallagher, J. J., & Schopler, E. pression and parenting competence of moth-
(1988). Mothers and fathers of young devel- ers of handicapped and nonhandicapped in-
opmentally disabled and nondisabled boys: fants: A longitudinal study. American Journal
on Mental Retardation, 94, 259–271.
*Harris, V. S., & McHale, S. M. (1989). Family
1
References with asterisks at the beginning are the life problems, daily caregiving activities, and
primary studies synthesized in this meta-analysis. the psychological well-being of mothers of

q American Association on Mental Retardation 167

VOLUME 111, NUMBER 3: 155–169 z MAY 2006 AMERICAN JOURNAL ON MENTAL RETARDATION

Depression in mothers G. H. S. Singer

mentally retarded children. American Journal book of research synthesis (pp. 111–124). New
on Mental Retardation, 94, 231–239. York: Russell Sage Foundation.
Hastings, R. P., & Beck, A. (2004). Practitioner *Miller, A. C., Gordon, R. M., Daniele, R. J., &
review: Stress intervention for parents of chil- Diller, L. (1992). Stress, appraisal, and coping
dren with intellectual disabilities. Journal of in mothers of disabled and nondisabled chil-
Child Psychology and Psychiatry, 45, 1338–1349. dren. Journal of Pediatric Psychology, 17, 587–
Hastings, R. P., & Taunt, H. M. (2002). Positive 605.
perceptions in families of children with de- Moes, D., Koegel, R., Schreibman, L., & Loos, L.
velopmental disabilities. American Journal on M. (1992). Stress profiles for mothers and fa-
Mental Retardation, 107, 116–127. thers of children with autism. Psychological Re-
Hawkins, N. E., Singer, G. H. S., & Nixon, C. ports, 71, 1272–1274.
(1993). Short term behavioral counseling for National Institute of Child and Human Devel-
families of persons with disabilities. In G. H. opment Early Child Care Research Network
S. Singer & L. E. Powers (Eds.), Families, dis- (NICHDECCRN). (1999). Chronicity of ma-
ability and empowerment (pp. 317–342). Balti- ternal depressive symptoms, maternal sensitiv-
more: Brookes. ity, and child functioning at 36 months. De-
Hays, R. D., Wells, K. B., Sherbourne, C., Rogers, velopmental Psychology, 35, 1297–1310.
W., & Spritzer, K. (1995). Functioning and Nixon, C., & Singer, G. H. S. (1993). A group
well-being outcomes of patients with depres- cognitive behavioral treatment for excessive
sion compared with chronic general medical parental self-blame and guilt. American Journal
illnesses. Archives of General Psychiatry, 52, 11– on Mental Retardation, 97, 665–672.
19. Pelchat, D., Bisson, J., Ricard, N., Perreault, M.,
Hedges, L. V., & Olkin, I. (1985). Statistical meth- & Bouchard, J. M. (1999). Longitudinal ef-
ods for meta-analysis. New York: Academic fects of an early family intervention pro-
Press. gramme on the adaptation of parents of chil-
*Kazak, A. E. (1987). Families with disabled chil- dren with a disability. International Journal of
dren: Stress and social networks in three sam- Nursing, 36, 465–477.
Radloff, L. (1977). The CES-D scale: A self-report
ples. Journal of Abnormal Child Psychology, 15,
depression scale for research in the general
137–146.
population. Applied Psychological Measurement,
*Kazak, A. E., & Marvin, R. S. (1984). Differences,
1, 385–401.
difficulties and adaptation: Stress and social
Risdall, D., & Singer, G. H. S. (2004). Marital ad-
networks in families with a handicapped
justment in parents of children with disabili-
child. Family Relations: Journal of Applied Fam-
ties: A historical review and meta-analysis. Re-
ily and Child Studies, 33, 67–77.
search and Practice for Persons with Severe Dis-
Kessler, R. C. (2003). Epidemiology of women abilities, 29, 95–103.
and depression. Journal of Affective Disorders, *Roach, M. A., Orsmond, G. I., & Barratt, M. S.
74, 5–13. (1999). Mothers and fathers of children with
Kessler, R. C., Zhao, S., Blazer, D. G., & Swartz, Down syndrome: Parental stress and involve-
M. (1997). Prevalence, correlates, and course ment in child care. American Journal on Mental
of minor depression and major depression in Retardation, 104, 422–436.
the national comorbidity survey. Journal of Af- Rosenberg, M., Adams, D. C., & Gurevitch, J.
fective Disorders, 45(1–2), 19–30. (1997). MetaWin. Sunderland, MA: Sinauer.
Langer, T. S. (1962). A twenty-two item screening Rosenthal, R., Rosnow, R. L., & Rubin, D. B.
score of psychiatric symptoms indicating im- (2000). Contrasts and effect sizes in behavioral
pairment. Journal of Health and Social Behavior, research: A correlational approach. Cambridge:
3, 269–276. Cambridge University Press.
Lin, N., Dean, A., & Ensel, W. (1986). Social sup- Rossiter, L., & Sharpe, D. (2001). The siblings of
port, life events, and depression. San Diego: Ac- individuals with mental retardation: A quan-
ademic Press. titative integration of the literature. Journal of
Lipsey, M. W. (1994). Identifying potentially in- Child and Family Studies, 10, 65–84.
teresting variables and analysis opportunities. *Scott, B. S., Atkinson, L., Minton, H. L., & Bow-
In H. Cooper & L. Hedges (Eds.), The hand- man, T. (1997). Psychological distress of par-

168 q American Association on Mental Retardation

VOLUME 111, NUMBER 3: 155–169 z MAY 2006 AMERICAN JOURNAL ON MENTAL RETARDATION

Depression in mothers G. H. S. Singer

ents of infants with Down syndrome. Ameri- Turnbull, A. P., Turbiville, V., & Turnbull, H. R.
can Journal on Mental Retardation, 102, 161– (2000). Evolution of family professional part-
171. nerships: Collective empowerment as the
*Seltzer, M. M., Greenberg, J. S., Floyd, F., Pettee, model for the early twenty-first century. In J.
Y., & Hong, J. (1998, November). Life course P. Shonkoff & S. J. Meisels (Eds.), Handbook
impacts of parenting a child with a disability. Pa- of early childhood intervention (pp. 630–650).
per presented at the Annual Scientific Meet- New York: Cambridge University Press.
ing of the Gerontological Society of America, *Walker, L. S., Ortiz-Valdes, J. A., & Newbrough,
Philadelphia. J. R. (1989). The role of maternal employment
Shadish, W. R., & Haddock, C. K. (1994). Com- and depression in the psychological adjust-
bining estimates of effect size. In H. Cooper ment of chronically ill, mentally retarded, and
& L. Hedges (Eds.), The handbook of research well children. Journal of Pediatric Psychology, 4,
synthesis. New York: Russell Sage Foundation. 357–370.
Singer, G. H. S., Gang, A., Powers, L., Cooley, E., *Witt, W. P., Riley, A. W., & Coiro, M. J. (2003).
Nixon, C., Kerns, K., & Williams, D. (1994). Childhood functional status, family stressors,
A comparison of two psychosocial interven- and psychological adjustment among school-
tions for parents of children with acquired aged children with disabilities in the United
brain injury: An exploratory study. Journal of States. Archives of Pediatric Adolescent Medicine,
Head Trauma Rehabilitation, 9, 38–49. 157, 687–695.
Singer, G. H. S., & Irvin, L. K. (1991). Supporting *Wolf, L. C., Noh, S., Fisman, S. N., & Speechley,
families of persons with disabilities: Emerging M. (1989). Brief report: Psychological effects
findings, practices, and questions. In L. H. of parenting stress on parents of autistic chil-
Meyer, C. A. Peck, & L. Brown (Eds.), Critical dren. Journal of Autism and Developmental Dis-
issues in the lives of people with severe disabilities orders, 19, 157–166.
(pp. 271–312). Baltimore: Brookes. Wolfensberger, W. (1967). Counseling the parents
Singer, G. H. S., Irvin, L. K., & Hawkins, N. of the retarded. In A. A. Baumeister (Ed.),
(1988). Stress management training of parents Mental retardation: Appraisal, education, and re-
of children with severe handicaps. Mental Re- habilitation (pp. 329–400). Chicago: Aldine.
tardation, 26, 269–277. Zimmer, K. P., & Minkovitz, C. S. (2003). Mater-
Singer, G. H. S., Irvin, L. K., Irvine, B., Hawkins, nal depression: An old problem that merits
N. J., & Cooley, E. (1989). Evaluation of com- increased recognition by child healthcare
munity based support services for families of practitioners. Current Opinion in Pediatrics, 15,
persons with developmental disabilities. Jour- 636–640.
nal of the Association for the Severely Handi-
capped, 14, 312–323.
Received 9/9/03, accepted 7/6/05.
Singer, G. H. S., Irvin, L. K., Irvine, B., Hawkins,
Editor-in-charge: Frank Floyd
N. E., Hergreness, J., & Jackson, R. (1993).
Helping families adapt positively to disability:
Overcoming demoralization through com- Requests for reprints should be sent to George
munity supports. In G. H. S. Singer & L. Pow- Singer; University of California, Santa Barbara,
ers (Eds.), Families, disability, and empowerment Department of Education, Santa Barbara, CA
(pp. 67–84). Baltimore: Brookes. 93106. E-mail: singer@education.ucsb.edu

q American Association on Mental Retardation 169