Full text of DOI:10.1093/geronb/57.6.P559

Journal of Gerontology: PSYCHOLOGICAL SCIENCES
Copyright 2002 by The Gerontological Society of America
2002, Vol. 57B, No. 6, P559–P568
Stress and Depressive Symptoms
Among Mexican American Elders
David A. Chiriboga,1 Sandra A. Black,2 Maria Aranda,3 and Kyriakos Markides4
1Department of Aging and Mental Health, Florida Mental Health Institute, University of South Florida, Tampa.
2Department of Epidemiology and Preventive Medicine, University of Maryland, Baltimore.
3School of Social Work, University of Southern California, Los Angeles.
4Department of Preventive Medicine and Community Health, University of Texas Medical Branch, Galveston.
Although social stressors have successfully predicted depressive symptomatology in a number of populations, few
studies have examined the relevance of stressors for Mexican American elders. Results are reported here from a
multistage probability sample of 3,050 Mexican Americans aged 65 and older drawn from a 5-state region. Par-
ticipants reported low levels of education and income, and most reported difficulty in reading or writing in En-
glish. Deaths, illness of close other, and financial problems were the three most frequent life events, and many
reported financial strains. Depressive symptoms were then regressed on demographic indicators, cognitive status,
linguistic acculturation, social supports, and three types of stressors. Being a woman, lower income, decreased in-
come, chronic financial strain, and several health stressors were associated with greater symptomatology. Results
identified a cluster of economic stressors and conditions that may play a critical role in the etiology of depressive
symptoms in this minority population.
N this article we report the experiences of a population
about whose stress exposure and its consequences little is
Of course, stress exposure does not occur in a vacuum:
The impact of stressors may be modulated by specific con-
ditions. Almost from its inception, stress research placed an
emphasis on identifying factors associated with the stress
process (Chiriboga & Cutler, 1980). This effort has by and
large been successful, with the result that stress today has
well-established general models and measures (Glass, Kasl,
& Berkman, 1997; Hobfoll, 1998; Kaplan, 1996; Thoits,
1995). Common to all models are one or more social stres-
sors (i.e., life events, chronic stressors, hassles), mediating
and contextual elements (i.e., adaptive resources), and indi-
cators of outcome (i.e., physical health, depression). Given
our focus on stressors, in this article we sought to include as
many of these variables as possible, while holding to the
single criterion of depressive symptoms.
Past research suggests that the influence of stress upon
functioning is robust and transcends culture and race
(i.e., Kamarck et al., 1998; Krause, Liang, & Gu, 1998). On
the other hand, there has been a developing interest in how
stress applies in specific contexts (Turner, Lloyd, & Lloyd,
1995; Dohrenwend et al., 1993; Chou & Chi, 2001). Evi-
dence already exists that specific sociodemographic and cul-
tural factors help shape both stress exposure and its conse-
quences (Briones et al., 1990; Connell & Gibson, 1997;
Chou & Chi, 2001). As one example, studies of stress and
depressive symptomatology have consistently found that
women report higher levels of both stress events and symp-
toms (e.g., Chiriboga, 1984; Culbertson, 1997; Mirowsky,
1996; Turner, Lloyd, et al., 1995).
I
known: the Mexican American elder. The Hispanic popula-
tion in general is of interest not only because it constitutes
the nation’s second largest minority group but also because
it is exhibiting a growth rate seven times larger than the
country as a whole (Therrien & Ramirez, 2000). Within
the broader Hispanic American community, Mexican Amer-
icans are the largest subgroup: of the 32.8 million Hispanics
in the United States in 2000, approximately 66.1% are of
Mexican origin and reside primarily in the Southwest (Ther-
rien & Ramirez, 2000).
The need for research on stress and the Mexican Ameri-
can elderly population draws from the well-documented re-
lationship of stress exposure to physical and mental health.
As a way of beginning an exploration, this article focuses on
depressive symptomatology as an outcome. We chose de-
pressive symptomatology because there is clear evidence of
an elevated rate of depressive symptoms among older His-
panics, especially women (e.g., Black, Markides, & Miller,
1998; Kemp, Staples, & Lopez-Aqueres, 1987; Mendes de
Leon & Markides, 1988;), but reasons for the elevation are
unknown. At the same time there are a wealth of studies
demonstrating a detrimental effect of stress exposure upon
the depressive symptoms of Anglo-American and other pop-
ulations. The problem of heightened symptom levels, in
turn, is accentuated by the underutilization of mental health
services by middle-aged and older Mexican Americans
(López-Aqueres, Kemp, Plopper, Staples, & Brummel-Smith,
1984; Vega, Kolody, Aguilar-Gaxiola, & Catalano, 1999).
An enhanced understanding of the relationship of stress ex-
posure and depression in this population could facilitate the
delivery of services if Mexican American elders were found
to manifest specific and remediable stress vulnerabilities.
An area of growing interest is the relevance of stress to mi-
nority populations, especially to their older members. The
stress context of older minorities groups is relatively unex-
plored despite the manifest need for this type of research
(Marsella, Friedman, Gerrity, & Scurfield, 1996). The need
P559

P560
CHIRIBOGA ET AL.
is underscored by the fact that people from disadvantaged
backgrounds may, because of these very disadvantages,
have fewer resources with which to cope and adapt (Blan-
kenship, 1998). For example, Chou and Chi (2001) found
that among the relatively impoverished Chinese elderly pop-
ulation of Hong Kong, greater financial strain is strongly re-
lated to depressive symptom expression.
reporting of depressive symptomatology, and (e) because of
the marginal economic circumstances of many older Mexi-
can Americans, economic stressors will demonstrate the
strongest associations.
METHODS
There are marked variations in adaptive resources across
minority populations that render any uniform assessment of
risk suspect. In the instance of the Mexican American elder,
research suggests that the extended family is generally more
extensive and potentially supportive than in mainstream
Americans (Markides & Black, 1996; Mendes de Leon &
Markides, 1988). Their often robust social resources, how-
ever, may be offset by economic problems that are not only
more likely among Hispanic groups but that may reduce
coping resources and create greater vulnerabilities (Flack et
al., 1995; see also Smith, 1999). Financial strain may indeed
be a major risk factor for many disadvantaged groups. Fer-
raro and Su (1999) report that such strain was a predictor of
psychological distress in three of the four Western Pacific
localities they studied. Overall, there is a need for better un-
derstanding of what constitutes risk factors, and the distri-
bution of risk factors, for diverse cultural groups (Culbert-
son, 1997; Perez-Stable, Marin, & Marin, 1994).
In addition to those already mentioned, we included mea-
sures of two other potential risk factors in the present inves-
tigation because of their potential relevance to the study
population. Because of a likelihood of cognitive decline in
the age groups being studied, we included a measure of cog-
nitive status. We also included indicators of acculturation
since low levels of acculturation have been associated with
depression in Mexican American elders (Zamanian, Thack-
rey, & Starrett, 1992). Whereas the available measures of
acculturation dealt primarily with just its linguistic aspect,
the latter aspect has demonstrated content and predictive va-
lidity (e.g., Cuellar, Arnold, & Maldonado, 1995; Marin &
Gamba, 1996)
The Study
Analyses drew from the initial wave of a multiwave
panel study of Mexican American elders. We closely mod-
eled the research design after existing Established Popula-
tions for Epidemiological Studies of the Elderly (EPESE)
research and included many of the EPESE instruments (c.f.
Cornoni-Huntley et al., 1990). Between July 1993 and April
1994, we conducted home interviews with an area probabil-
ity multistage sample of 3,050. All participants were aged
65 or older, noninstitutionalized, and resided in a five-state
region that contains the majority of Mexican Americans:
Texas, California, Mexico, Colorado, and Arizona.
Sampling Frame
Louis Harris and Associates Poll, Inc. developed and im-
plemented a multistage area probability sampling frame for
the study. As described in detail by Miller, Markides, and
Black (1997), we included four stages in sampling proce-
dures: (a) we selected counties from a complete listing of
counties in the five states; (b) we randomly selected 300
census tracts (primary sampling units); (c) we randomly se-
lected blocks from each census tract; and (d) we conducted
house-to-house surveying, requesting interviews of all Mex-
ican Americans, aged 65 and older. The response rate, at
86%, approximates the response rate found in the other
EPESE (see Cornoni-Huntley et al., 1990). Of the 3,050
participants included in the sample, 2,873 were interviewed
in person and 177 were interviewed by proxy, either be-
cause of illness or low cognitive capacity as judged by their
score on the Mini-Mental State Examination (MMSE).
A final point is that although the general relationship of
stress to health is well established, there is some suggestion
that it is now time for the entire paradigm, and its constitu-
ent building blocks, to be reexamined (e.g., Brown, Harris,
& Hepworth, 1994). Studies of diverse populations may fa-
cilitate this reexamination process at the same time that they
generate useful information concerning the relationships
among variables within a specific group. In the present in-
stance, the availability of a large, five-state probability study
provides a unique opportunity for rigorous examination of
the applicability of stress measures to an at-risk minority
population. Although our general interest was in beginning
an assessment of the role of stress for older Mexican Amer-
icans, we developed five hypotheses on the basis of existing
literature: (a) consistent with past research, the level of de-
pressive symptom expression will be relatively high; (b)
consistent with past research, women and those with lower
incomes will be more depressed; (c) because of the influ-
ence of cultural factors, those who are more highly accultur-
ated will demonstrate lower levels of depressive symptoms;
(d) because of the general robustness of the stress paradigm,
stressful experiences of all types will be associated with the
Conceptual Model
The model applied to the cross-sectional data represents a
relatively “no frills” direct effects approach incorporating
variables common to most stress models (e.g., Lazarus &
Folkman, 1984; Hobfoll, 1998) but does include additional
variables appropriate to the study population. Five hierar-
chical levels of explanatory variables guided the regression
analyses, entered by sets according to whether they tapped
preexisting or stable social characteristics (e.g., age, gender,
income), were more or less current, or might act as condi-
tioning or predisposing factors (e.g., cognitive status, accul-
turation, supports) for depression. In order of entry, the sets
included: (a) sociodemographic characteristics; (b) cogni-
tive status; (c) acculturation; (d) social resources and sup-
ports; and (e) three domains of stressors: recent life events,
chronic strains, and recent health events. The criterion se-
lected for study was depressive symptomatology, one of the
most widely used indicators of stress response (e.g., Anesh-
ensel, 1996; Brown, Harris, & Hepworth, 1994). The evi-
dence for heightened levels of depressive symptoms among

STRESS AND DEPRESSIVE SYMPTOMS
P561
Mexican American elders was critical to our selection of
this criterion (e.g., Black et al., 1998).
reliability for these seven items was high (Cronbach’s ꢀ ꢁ
.91), but we did not include a summary scale in the follow-
ing analyses because we were most concerned with identi-
fying individual items of significance to stress and depres-
sion. To round out the breadth of assessment, we included also
two additional indicators (see Cuellar, Arnold, & Maldonado,
1995) of acculturation: (a) whether the R was born in the
United States or not, and (b) whether the interview was con-
ducted (per participant request) in English or Spanish.
Measures
Where possible, we drew measures from other EPESE
studies, especially the one conducted at Duke University. If
suitable measures were not available from the EPESE
studies, we drew measures were drawn from the Hispanic
Health and Nutrition Examination Survey (HANES) or cre-
ated ones specifically for this study. The measures analyzed
for this article represent a subset of those available in the
larger study. A Spanish version of the interview schedule
was available, and all interviewers were bilingual.
Social resources.—Social resources are not only the
most commonly studied mediators in stress research, but
also represent a construct of particular significance to
studies of Mexican Americans because of the generally
larger size of the extended family. Given the importance of
the overall construct, we included 11 items. Two dealt with
how often the R felt he or she could (a) count on some fam-
ily or friends in “times of trouble”, and (b) talk to some fam-
ily or friends about “your deepest problems” (coded from 1
[most of the time] to 3 [hardly ever]). The remaining nine
items dealt with characteristics of the family network. Items
included: (a) how many different family generations lived in
the household; (b) how many individuals under 18, 18–24,
25–49, 50–64, and 65 and older lived in the household; (c)
number of living sons and number of living daughters; and
(4) how many children the R saw at least once a month.
Cronbach’s ꢀ for the 11 items, at .65, was relatively low. In-
spection of the zero-order intracorrelations indicated great
variation in magnitude, which is not surprising given that we
did not intend for the included items to constitute a scale.
Sociodemographic.—The six items include: (a) age, (b)
gender, (c) educational attainment (grade level), (d) personal
income, (e) work status (fully employed, partially employed,
not employed), and (f) marital status (in analyses coded as
presently married and spouse present vs. all alternatives).
Cognitive status.—We included a Spanish version (Bird,
Canino, Rubio-Stipec, & Shrout, 1987; Escobar et al., 1987)
of the MMSE (Folstein, Folstein, & McHugh, 1975), as a
means of assessing the individual’s general adaptive capac-
ity. Note that we substituted spelling world (or the Spanish
equivalent, mundo) backwards for the option of serial sev-
ens subtraction. Scores can range from 0 to 30, with higher
scores indicating fewer errors. For mainstream American
populations, scores from 0 to 17 suggest severe cognitive
impairment, 18–23 indicate mild impairment, and scores
from 24 to 30 indicate probable lack of deficit (Weissman et
al., 1985; Yu et al., 1989; Zarit & Zarit, 1998). Although
there is evidence that education can affect the significance
of scores, no change in the above cutoffs was made in the
presentation of frequencies. Cronbach’s alpha level of inter-
nal reliability for the present data set was .65.
Stressors.—Measures representing three different do-
mains of stressor were included.
1. Life events. These were 13 acute life events reported by
the R as occurring within the past year. Seven items were
from the Duke EPESE (Cornoni-Huntley et al., 1990; see
also Holmes & Rahe, 1967): death of a close other, wid-
owed, illness of close other, finances got worse, change
in residence, retired, and divorced. We draw six addi-
tional items from several inventories developed for older
populations (e.g., Chiriboga & Cutler, 1980; Murrell, Nor-
ris, & Hutchins, 1984): victim of crime, became caregiver,
spouse hospitalized, married, separated, moved to USA (a
subset of the general question concerning change in resi-
dence). Note that for one general item (“Did anyone close
to you die?”), we included probes to ascertain relationship
of the deceased for up to two reported deaths. Because the
two probes elicited reports of more deaths than did the ini-
tial yes or no response, results are based on recodings that
drew upon responses to all three items. Internal reliability
for all items, at .45, is quite low but typical of inventories
designed to assess the incidence of adventitious events over
a limited period of time.
Acculturation.—Following the lead of the Hispanic
HANES, the overall study included 13 acculturation items
drawn from the indices of Cuellar, Harris, and Jasso (1980)
and Hazuda, Stern, and Haffner (1988). The majority of
items assessed linguistic acculturation, only one dimension
of what is generally recognized to be a multidimensional con-
struct (e.g., Cuellar, Arnold, & Maldonado, 1995; Keefe &
Padilla, 1987; Padilla, 1980). Additional questions probed for
the cultural background of social others. Because several of
the items had relatively high levels of missing data as a re-
sult of vacated or never occupied role status (e.g., language
currently spoken with parents: in most cases the parent was
deceased), the present analyses employed only five of the
linguistic items and two of the cultural background items.
Included items each had 2% or less missing data, and in-
cluded self-reports of: (a) how well the respondent (R)
spoke English, (b) how well the R was able to read English,
(c) language used with most of R’s friends, (d) language
used at family gatherings, (e) language used with most of
neighbors, (f) whether neighbors had been primarily Mexi-
can American or Anglo American over the life course, and
(g) whether close friends had been primarily Mexican Ameri-
can or Anglo American over the life course. The internal
2. Chronic economic strains. Because research has empha-
sized the important role of chronic stress (e.g., Cohen et
al., 1998), we included five items. The focus of these
items was disproportionately on economic stress, be-
cause past research has indicated that Mexican Ameri-
cans in general experience higher rates of poverty, and

P562
CHIRIBOGA ET AL.
we hypothesized that economic strain might be particu-
Overall, these adjustments allowed us to make generaliza-
tions regarding distributions and relationships manifest in
the Mexican American elderly population living in the five
southwestern states from which the sample was drawn.
larly important in the lives of older members of this
target population. Four items came from the Pearlin,
Menaghan, Lieberman, and Mullen (1981) Index of Chronic
Economic Strain (reported Cronbach’s ꢀ ꢁ .73; 3-year
test–retest reliability ꢁ .9; Chronbach’s ꢀ for present
study ꢁ .76).
RESULTS
3. Personal health events. These were 10 items that assess
onset or initial diagnosis of either acute health events or
chronic health problems within the past year. Such health
stressors are problematic in that they often introduce a
spurious effect into predictions of health-related out-
comes. In the analyses presented in this article, we
treated health stressors as a distinct category. We were
concerned that because the outcome measure of depres-
sion includes several somatic items, it was possible that
any relationship could be due to explanatory and out-
come measures both tapping health status. Of the in-
cluded items, two dealt with experiences within the past
three months: (a) having a health problem that limited
activities, and (b) having a health problem that required
bed rest. The remaining eight items dealt with health ex-
periences over the past year: (a) an illness or injury re-
quiring hospitalization; (b) total days spent in a hospital;
(c) heart attack; (d) stroke; (e) bone fracture; and (f) di-
agnosis with cancer, hypertension, or diabetes. Internal
reliability, at .21, was low but again we did not presume
health events to constitute a scaleable set of items.
Basic Characteristics
Demographic.—Because the demographic characteris-
tics of the sample have implications for how Mexican
American elders fare when exposed to stressors, we present
these characteristics in some detail. As shown in Table 1,
and reflecting the fact that the Mexican American popula-
tion has proportionately fewer older people than the Anglo-
American population, the age distribution was highly
skewed. Two thirds of the sample were aged 65 to 74 years,
an age period often referred to as the young old stage of life;
only slightly more than 7% were aged 85 and older. Even this
oldest group, however, was sizable when extrapolated to the
population of the five-state region: over 35,000 Mexican
Americans were estimated to fall into the oldest-old category.
Women accounted for 56% of the sample, fewer than
would be anticipated in an Anglo sample. More striking,
perhaps, are the distributions for income and education. Well
over a third reported personal incomes of less than $5,000,
and half had not completed elementary school. These results
demonstrate that, compared to other groups of older adults,
Mexican Americans have fewer economic resources with
which to cope with adverse events and conditions.
Depressive symptoms.—We included the 20-item Cen-
ter for Epidemiologic Studies–Depression (CES-D) scale
(Radloff, 1977). The CES-D contains 16 negatively oriented
items and 4 positively oriented items (reverse scored) and
has proved to be a valid and reliable instrument when used
with Spanish-speaking populations (Angel & Guarnaccia,
1989). Although previous factorial studies of the CES-D in
mainstream samples have reported from two to four factors,
there is evidence that there are two major dimensions, and
this has been found to be the case with our own sample
(Miller, Markides, & Black, 1997). We used the translation
developed for the Hispanic HANES; at .88 the internal reli-
ability for this study was moderately high. A score of 16 or
higher is often used as a cutoff for heightened probability
for clinical depression, although the CES-D is not consid-
ered a diagnostic or clinical tool (Radloff, 1977; Zarit &
Zarit, 1998).
Cognitive status.—Although the levels of educational
disadvantage evident in our sample have been associated
with increased MMSE errors in both American and interna-
tional studies (Crum, Anthony, Bassett, & Folstein, 1993),
the results suggested that this was a relatively intact group.
Nearly two thirds fell within what is generally considered
the normal range for Anglos, with another 30% falling
within the mild impairment category (Zarit & Zarit, 1998).
Fewer than 7% had scores indicative of severe cognitive
deficit. This low proportion translated into a projection that
well over 30,000 older, community-dwelling Mexican
Americans in the five-state region may suffer from severe
cognitive impairment.
Acculturation.—By and large the sample displayed rela-
tively low levels of linguistic acculturation. For example,
approximately 37% was unable to read English and 30% re-
ported that they could not speak English. In contrast, about
58% reported they spoke only Spanish with their spouse,
and a full 74% reported that they spoke only Spanish with
their (middle-aged) children. At family gatherings, the pro-
portion relying solely on Spanish dropped to 47%, perhaps
reflecting out-marriages and greater use of English among
younger generations. Two questions asked about character-
istics of neighbors and friends over the participant’s life-
time, with responses suggesting that large numbers of older
Mexican Americans live rather segregated lives. Over 76%
reported that their friends had been mostly Mexican Americans
Analyses
All computations included adjustments based on the mul-
tistage sampling design. For example, although the actual
frequency distributions shown in Table 1 are unweighted,
the percentages represent normalized population estimates,
and the extrapolated frequencies represent projections based
on census data (N ꢁ 498,176) for Mexican Americans aged
65 and older in the region covered (U.S. Census, 1991). We
adjusted also all derived scores, such as means, standard de-
viations, or regression coefficients to represent normalized
estimates. In calculating levels of significance, we used the
SUDAAN procedure (Version 7.5.2; Shah, Barnwell, &
Bieler, 1997) to adjust analyses for weight and design effects.

STRESS AND DEPRESSIVE SYMPTOMS
P563
Table 1. Characteristics of the Sample (n ꢁ 3,050), by Actual Number, Normalized Percent, and Inflated Number
Inflated
Inflated
Characteristic
n
%
n
Characteristic
n
%
n
Demographic
Age
Life Events Within Year
Death of close otherb
Illness of close otherb
Finances got worseb
Changed residencesb
Spouse hospitalized
Became caregiver
734 22.4 111,821
Young-old (65–74)
Old-old (75–84)
Oldest-old (85ꢂ)
2003 66.1 329,218
833 26.9 133,786
598 19.3
420 14.1
256 10.2
96,118
70,118
50,876
46,242
28,833
12,590
214
7.2
35,172
Women
Married
Income ꢃ $5000
Up to sixth-grade education
1775 56.7 282,394
1693 55.2 275,216
1094 34.3 158,483
2208 71.1 347,991
291
187
78
46
60
13
6
9.3
5.8
2.5
Widowedb
Retiredb
Victim of crime
Moved to United Statesb
Divorcedb
Separated
Married
2.3 11,211
Acculturation
1.8
0.8
0.3
0.2
0.1
9,120
3,943
1,354
755
Doesn’t speak English
Doesn’t read English
Only Spanish with friends
1016 30.0 149,648
1200 36.7 178,027
1630 50.3 250,453
1501 47.7 237,677
1643 49.7 247,828
2516 73.1 363,869
2514 76.5 377,727
1347 46.9 233,569
2374 72.6 361,676a
4
4
Only Spanish at family gatherings
Only Spanish with neighbors
Neighbors mostly Mexican-American
Close friends mostly Mexican-American
Born outside the United States
Spanish used for interview
Mini-Mental State Examination Score
0 to 17 correct (“severe” range)
Social
557
Chronic Financial Strain
Not enough for food: fairly often or more
Difficulty w/monthly bills: some of time or more
Not enough for medical care: fairly often or more 1609
Not enough at end of month
Acute Health Events
199
7.2
31,387
244 55.4 241,370
8.9
481 17.5
38,831
75,825
152
6.7
31,454
Health limited activities within past 3 months
Bedfast within past 3 months
Illness and/or injury limited activitiesb
Hospitalized 1ꢂ times in past yearb
Hypertension diagnosed
Diabetes diagnosed
Heart attack
Stroke
Cancer diagnosed
Bone fracture
820 27.0 133,734
597 19.0 94,896
625 20.5 101,885
Lives alone
Up to two people at home
640 20.9 104,234
1975 61.9 308,573
1752 57.6 286,867
1747 56.7 282,428
2502 81.0 391,713
579 19.2
95,875
29,149
10,155
15,545
12,522
6,200
Has two or fewer living sons
Two or fewer living daughters
No children to age 18 at home
Sees one or no children per month
Can count on family most times
Can talk about problems most times
152
60
83
47
35
22
5.9
2.0
3.1
2.5
1.2
0.7
498 22.1
90,722
2024 68.5 321,184
1970 66.2 310,725
3,534
CES-D Scale
Score of 16 and above
687 25.7 120,770
Notes: Percentages were computed by using normalized inflation weights to allow estimates to the population of Mexican Americans who are aged 65 and older
and living in the five states in the United States covered by the study. Frequencies are unweighted; inflated frequencies project to estimated population of Mexican
Americans aged 65 and older (N ꢁ 498,176) in the five-state region under study [U.S. Bureau of the Census. (1991). Race and Hispanic Origin (Census Profile No. 2).
Washington, DC: Department of Commerce]. CES-D ꢁ Center for Epidemiologic Studies–Depression.
aEstimated number of older Mexican Americans who would have requested an interview in Spanish.
bLife event item included in Duke Established Populations for Epidemiologic Studies of the Elderly.
(or Mexican), and for 73% the same was true of neighbors.
Reflecting the overall level of linguistic acculturation in this
sample, nearly 73% requested that the interviews be con-
ducted in Spanish.
Finally, it should be noted that about 47% of the partici-
pants—nearly half—had been born outside of the United
States of America. Although this high proportion of immi-
grants might suggest greater than usual vulnerability, many
of these individuals had spent a substantial length of time
here: The average number of years in the United States was
42.7 (SD ꢁ 22.34). Only 13 had moved to this country
within the past year; at the other extreme, a total of 7 had
immigrated 90 or more years ago.
average number of persons in the household was 2.75). Ap-
proximately 57% reported having only one or two living
sons, and roughly the same proportion reported one or two
living daughters. A series of questions asking about the
numbers of people in the household of differing ages again
suggested that the majority of the participants did not live
with people of differing generations. On the other hand, al-
though generally not living as part of an extended family,
the majority of these elders were not socially isolated:
68.5% reported that they could count on family members to
help out most of the time, and 66.2% reported having a fam-
ily member or friend with whom they could talk “about
your deepest problems” most of the time. Only 22% re-
ported seeing one or no children each month.
Social supports.—Turning to social support characteris-
tics (Table 1), the results suggested that the Mexican Ameri-
can elders were generally well integrated into the family
network. Contrary to stereotypes, the majority did not live
as part of a large and multigenerational household: Over
20% lived alone and this percentage increased to approxi-
mately 62% when two person households were counted (the
Life events.—Several of the included events were re-
ported with some frequency. As shown in Table 1, 22.4% of
the subjects reported the death of a family member or close
friend within the past year, and 19.3% reported the serious
illness of a family member. Slightly over 14% reported a
worsening of their financial situation and about 10% reported

P564
CHIRIBOGA ET AL.
Table 2. Results of Hierarchical Set Regression of Depressive
a residential change of some sort. Several other events oc-
curred with less frequency, but are still of interest because
the events were not trivial: 9.3% reported that their spouse
was hospitalized within the past year; 5.8% reported they
had become caregivers, and 2.5% reported becoming a
widow (a specific instance of “death of a close other”).
Symptoms on Assumed Predictors (n ꢁ 2,852)
(Block R2 Change)
Block
Beta Coefficient
Block 1: Demographic
Gender
(.05)
.16***
ꢄ.05
ꢄ.07*
ꢄ.06*
Marital status
Personal income
Education
Chronic strains.—Perhaps the most striking information
contained in the distributions for more durable stressors was
that over 55% of participants reported difficulty in paying
monthly bills some or most of the time. This percentage
translated into an estimate that over 240,000 Mexican Ameri-
cans elders in the five-state region frequently encountered
problems paying their bills. Approximately 18% reported
that they did not have enough money left at the end of the
month to make ends meet. Lower proportions reported not
having enough money for food (7.2%) and medical care
(8.9%) some or most of the time.
Block 2: Cognition
(.01)
Number correct on MMSE
ꢄ.09*
Block 3: Acculturation
(.00)
.08*
ꢄ.05
Language with family gatherings
Ethnicity of neighbors
Block 4: Social Supports
Someone to talk to
(.03)
.13***
.06*
.05
ꢄ.03
Number of living sons
Has someone to count on
Number of people in household
Block 5: Life Events
Illness in close other
Finances get worse
Retired
(.05)
.04
.21***
ꢄ.04***
.04
Health stressors.—For this sample, the most frequently
reported health events were those that either limited activi-
ties or required hospitalization (Table 1). The most com-
monly reported health event, at 27%, was having to cut
down on activities within the past 3 months because of ill-
ness or injury. Approximately 19% reported having been
bedfast all or most of a day within the past 3 months be-
cause of illness or injury. At least one hospitalization during
the past year was reported by 19.2%, and nearly 17% re-
ported having spent 2 or more days in the hospital. The fre-
quency of being diagnosed during the past year with hyper-
tension, diabetes, and cancer was relatively low, as was the
incidence of stroke and heart attack. Still lower was the inci-
dence of bone fractures, a finding that builds on reports indi-
cating that the prevalence of osteoporosis among older
Mexican Americans was relatively low (Markides, Rudkin,
Angel, & Espino, 1997).
Newly widowed
Block 6: Chronic Money
(.04)
.11*
.08
Not enough money for food
Not enough money for medical care
Not enough money for bills
ꢄ.06
Block 7: Health Events
Days in hospital
Activities limited
Stroke
(.07)
.06**
.23***
ꢄ.04*
.04
Cancer
Bedfast within year
.03
Note: Beta coefficients were computed by using normalized inflation
weights; probability levels were computed by using SUDAAN, Version 7.5.2x.
MMSE ꢁ Mini-Mental State Examination.
*p ꢅ .05; **p ꢅ .01; ***p ꢅ .001.
Depressive symptoms.—The outcome measure used in
this study was the CES-D. With the standard cutoff of 16 or
higher, the results suggested that over one quarter of the Mex-
ican American elderly population in the five-state region cov-
ered by the study manifest symptom levels indicative of de-
pression. This translated into a five-state estimate of 120,770
older Mexican Americans who may be in need of treatment.
Demographic variables accounted for 5% of the variance
in depression (Table 2). As reported in other studies of Mex-
ican Americans (e.g., Chiriboga, 1996), as well as studies of
the majority population, women were substantially more
likely to report depressive symptoms. Also in accord with
studies of Anglo-American and other populations, partici-
pants with lower levels of education and lower incomes re-
ported more symptoms.
Prediction of Depressive Symptomatology
To evaluate the contribution of a general stress model to-
ward an understanding of depressive symptomatology in
this sample, we employed a hierarchical regression ap-
proach that followed the conceptual model outlined earlier
in this article. Each included variable has been found in pre-
vious research to associate with depressive symptoms. As
shown in Table 2, at nearly every level the hierarchical anal-
ysis generated significant results. Overall, the predictive set
accounted for 25% of the variance in depressive symptoma-
tology. It should be noted that all predictors retained for the
final analysis had previously demonstrated probability lev-
els of .10 or lower, when we ran initial regressions without
the use of inflation weights. We adjusted these probabilities
during the final SUDAAN procedures.
The MMSE contribution was significant and suggested
that older Mexican Americans with lower scores on cogni-
tion were likely to report more depressive symptoms. At the
same time, the very low proportion of explained variance
that came with adding the cognitive score suggested that
cognitive factors play only a minimal role in the origins of
depressive symptoms. One possible explanation for the
quite minimal MMSE contribution was that cultural or lin-
guistic factors may have affected responses to the cognitive
test and thereby vitiated its meaningfulness. However, al-
though those who took the MMSE in English scored signifi-
cantly higher than those who took the test in Spanish, the
mean difference was minor (means of 24.77 vs 24.55,
t(2851) ꢁ 12.12, p ꢁ .00).

STRESS AND DEPRESSIVE SYMPTOMS
P565
Turning to linguistic acculturation, this two-measure
DISCUSSION
(only two variables contributed in initial regressions) set ac-
counted for less than 1% of the variance in depressive
symptoms. The one significant association indicated that
those elders who were more likely to speak Spanish at fam-
ily gatherings were significantly more likely to report symp-
toms of depression. On the whole, however, the results sug-
gest that lower levels of acculturation do not dramatically
increase the risk of experiencing depressive symptoms for
older Mexican Americans.
This article focuses on factors associated with depressive
symptoms among older Mexican Americans. Although anal-
yses are inherently limited by virtue of the cross-sectional
design, a unique characteristic of the investigation is its abil-
ity to extrapolate numbers and percentages to the entire five-
state region included within the sampling frame. The result-
ing figures, as depicted in the first table, show the extent of
need and risk in the population. Confirming our initial hy-
pothesis (regarding levels of depression), the proportion re-
porting depressive symptoms is surprisingly high. Indeed, at
25.7% the proportion who report 16 or more symptoms on
the CES-D (a cut-off score associated with a heightened risk
of clinical depression) is substantially above the 15.1% re-
porting the same symptom range in the primarily Anglo-
American New Haven EPESE (White, Kohout, Evans, Cor-
noni-Huntley, & Ostfeld, 1986). Such findings emphasize the
need for clinical intervention in the population under study.
The level of symptomatology is not particularly surpris-
ing, however, given the generally sparse economic and edu-
cational resources available to these older Mexican Ameri-
cans. We had expected relatively low incomes on the basis
of existing census data, but the fact that well over one third
(37.5%) of our subjects report personal incomes of under
$5000 is particularly striking. This proportion is lower than
levels reported by any of the other EPESE populations.
Comparable figures include 25.6% for the East Boston
EPESE, 17.2% for the Iowa EPESE, 29.0% for New Haven,
and 25.3% for North Carolina (Brock, Manton, & Woodbury,
1990; Brock, Wineland, Freeman, Lemke, & Scherr, 1986).
In light of the extreme poverty manifest in our sample, it
is also not surprising that chronic and acute economic prob-
lems are among the most common stressors reported. Over
50% report difficulty meeting monthly bills at least some of
the time, and over 17% report not having enough money to
make ends meet at the end of the month. Among more acute
stressors, the two most frequent events are related to the
death or illness of a close other. The third most frequent
event is a worsening of financial status within the past year.
Medical events are also common, with over a quarter of the
sample reporting that their health had limited activities dur-
ing the 3 months prior to the interview.
The findings for social supports are very much in line with
results from studies of other populations of older adults. As
one example, it has been known for over 30 years that having
someone to talk to about problems is associated with mental
wellness (e.g., Lowenthal & Haven, 1967). Thus, it is not re-
ally a surprise to find that having someone to talk to about
problems was associated with fewer symptoms in this Mexi-
can American group. A finding that is at the very least pro-
vocative was that those with more living sons were likely to
report more symptoms (number of living daughters did not
contribute in either direction). An examination of zero-order
correlations indicated that having more living sons was sig-
nificantly (p ꢁ. 001) associated with being male, being less
linguistically acculturated, and living with more generations
in the same residence. More investigation is obviously
needed to parse out the relationships with mental health.
The last three sets in the hierarchical regression all dealt
with various domains of life stress and together accounted
for 16% of the variance in symptoms. In the first set of stres-
sors, we entered the more traditional life-event measures.
Three associations remained significant after the SUDAAN-
based regression, with the strongest indicating that people
who had recently experienced a worsening of their financial
situation were more depressed. In addition, those who had
retired within the year were less depressed than those who
retired over a year ago. This latter finding suggests that dis-
tance from retirement may represent a proxy for chronic
economic problems or for age itself. Finally, people whose
spouse had died within the year reported more symptoms.
Overall, the life-event measures explained 5% of the vari-
ance in depressive symptoms, an amount equal to that ex-
plained by demographic characteristics.
The indices of chronic financial problems also made an
independent, and significant, contribution, accounting for ap-
proximately 4% of the variance in symptom level. The only
significant contributor was a measure indicating how often
the subject reported not having enough money to buy food.
After we adjusted the significance level by the SUDAAN pro-
cedure, the significance levels of measures indicating diffi-
culty with medical bills and bills in general fell to what is
sometimes indicated to be a “trend” level (p ꢃ .10).
The final set we entered into the equation (i.e., that deal-
ing with recent changes in one’s personal health) made the
greatest overall contribution to levels of symptom expres-
sion: 7%. The two major contributors were variables that
seem to reflect the consequences of many different health
problems. The first was a self-report of the amount to which
one’s activities had become limited by ill health within the
past 3 months. The second was a self-report of how many
days had been spent in a hospital over the entire year.
The investigation not only evaluated prevalence of poten-
tial risk factors but sought in addition to evaluate which of
these factors are associated with problems. We, therefore,
examine how depressive symptoms are related to stressors
and other elements of what is generally considered to be the
basic stress mode. The results demonstrate that several
stressors—as well as several other social and personal char-
acteristics—do in fact manifest an association. Generally,
the factors that are associated with symptom expression in
this population are very similar to those factors associated
with symptoms in Anglo Americans, African Americans,
and other groups and nationalities. The role of demograph-
ics, for example, is essentially the same as has been reported
in numerous studies of mainstream Americans. Specifically,
and supporting our second hypothesis, women and those
with lower incomes report more symptoms.
Cognitive status, the second hierarchical set, is not gener-
ally included in stress paradigms. We include cognition

P566
CHIRIBOGA ET AL.
because the perception of stress is widely acknowledged as
important in the stress literature, and in older populations
cognition might play a critical role as a conditioning factor
for stress responsivity in addition to having a direct predic-
tor role with regard to depressive symptoms. The results in-
dicated that cognitive status played only a minor role with
regard to symptom expression, and the triad of stressor
types (acute, chronic, and health related) continued to be
significantly associated with symptoms after the removal of
variance attributed to cognition.
Partial support is provided for our third hypothesis (re-
garding the role of acculturation). Only one acculturation
measure is associated with symptoms: Those who are less
likely to speak Spanish at family gatherings report fewer
symptoms. This particular indicator may indirectly tap the
level of overall family acculturation: It can be hypothesized
that the more likely an elder is to speak Spanish in family
gatherings, the less acculturated is his or her extended fam-
ily. One result may be that the extended family is less able
to provide help, including linkages with the formal provid-
ers. Two caveats deserve mention here. First, given that
most of our measures tap just one dimension of accultura-
tion, the scarcity of associations may reflect limitations of
our instrumentation rather than limitations of the concept it-
self. Second, being more or less acculturated in itself is not
clearly linked to adaptation: The individual and context play
important roles (Garza & Gallegos, 1995; Padilla, 1980;
Ward, 2001).
In contrast to cognition and acculturation, but providing
support for the fourth hypothesis, all three sets of stress in-
dicators do account for moderate proportions of the variance
in depressive symptoms. In total, the 12 measures included
in the final analysis generate 16% of the 25% of the variance
accounted for in the full model. Although these levels are
certainly not striking, they hold interest for at least three
reasons. The first reason is that we have relatively few items
within each of the three stressor inventories. Shorter inven-
tories such as these often are lacking in content validity:
They may not detect the stress experiences of subjects (e.g.,
Chiriboga & Cutler, 1980). On the other hand, the present
results suggest that a relatively brief array of potential stres-
sors will generate results if the items are carefully selected
for population relevance.
Given their income levels, it is not surprising that low in-
come is significantly associated with higher levels of de-
pressive symptomatology. However, the influence of eco-
nomic factors clearly does not end with income. Even after
the variance associated with the latter variable is partialed
out of the equation, financial life events and chronic eco-
nomic strains, in their respective sets, are significantly asso-
ciated with symptom expression. In other words, it is not
just level of income, but the experience of acute and chronic
economic events and strains, that independently and addi-
tively affect their mood.
These results are in agreement with past research sug-
gesting that economic constraints are profoundly important
in the lives of older adults (e.g., Brenner, 1997; Krause,
1997). Low income is perhaps the most common of con-
straint, with either acute or chronic financial stressors occa-
sionally added or employed in substitution. For example,
Ferraro and Su (1999), who refer to money as a power re-
source, used financial strain as their primary economic indi-
cator in a study of psychological distress in four Western
Pacific countries. In the present study, we find that for older
Mexican Americans there may be a triad of economic con-
straints with implications for emotional well-being. It is not
simply low income. It is not simply a specific financial loss.
It is not simply the chronic burden of meeting monthly bills.
It is in fact all three (income, economic events, and chronic
strains) that exert an independent and additive influence on
symptoms of depression. For health professionals these re-
sults suggest a new strategy for risk screening: Although
low income remains an important risk factor for depression
among Mexican American elders, attention should be di-
rected towards identifying persons with chronic and acute
financial stressors. These stressors play a role in depres-
sive symptom expression that is independent of the effect
of income.
Acknowledgments
This research was supported by National Institute on Aging Grant R01
AG10939-01.
Address correspondence to David Chiriboga, Department of Aging and
Mental Health, Louis de La Parte Florida Mental Health Institute, Univer-
sity of South Florida, 13301 Bruce B. Downs Blvd., Tampa, FL 33612.
E-mail: dchiriboga@fmhi.usf.edu
The second reason is that most studies of stress focus on
only one type of stressor: either life events or chronic strains.
In this study, we include both events and chronic strains. We
also separate out those events that can be classified as health
related, so that the contribution of each can be examined in-
dependently. It is noteworthy that all three types of stress
measure contribute to our findings.
The third reason the stress measures are of interest is that
they told a coherent story. Supporting the fifth hypothesis,
results from the life events and chronic strain sets of the re-
gression analysis suggest vulnerabilities related to socioeco-
nomic factors. This vulnerability to economic stress occurs
in the context of socioeconomic disadvantage: Over 35%
have an income of less than $5,000, and over 71% report a
sixth-grade education or less. The substantial proportions
reporting limited or no knowledge of English may also rep-
resent at least a potential barrier to health services.
References
Aneshensel, C. S. (1996). Consequences of psychosocial stress: The uni-
verse of stress outcomes. In H. B. Kaplan (Ed.), Psychosocial stress
(pp. 111–136). San Diego: Academic Press.
Angel, R., & Guarnaccia, P. J. (1989). Mind, body, and culture: Somatiza-
tion among Hispanics. Social Science and Medicine, 28, 1229–1238.
Bird, H. R., Canino, G., Rubio-Stipec, M., & Shrout, P. (1987). Use of the
Mini-Mental State Examination in a probability sample of a Hispanic
population. Journal of Nervous and Mental Disease, 175, 731–737.
Black, S. A., Markides, K. S., & Miller, T. Q. (1998). Correlates of depres-
sive symptomatology among community-dwelling Mexican American
elderly. Journal of Gerontology: Social Sciences, 53B, S198–S208.
Blankenship, K. (1998). A race, class, and gender analysis of thriving.
Journal of Social Issues, 54, 393– 404.
Brenner, M. H. (1997). Heart disease mortality and economic changes, in-
cluding unemployment, in Western Germany 1951–1989. Acta Physio-
logica Scandinavica, 640 (Suppl. 161), 149–152.
Briones, D. F., Heller, P. L., Chalfant, H. P., Roberts, A. E., Aquirre-Hauch-
baum, S. F., & Farr, W. F. (1990). Socioeconomic status, ethnicity,

STRESS AND DEPRESSIVE SYMPTOMS
P567
psychological distress, and readiness to utilize a mental health facility.
American Journal of Psychiatry, 47, 1333–1340.
sonal choice: A humanistic perspective on acculturation. In A. M.
Padilla, (Ed.), Hispanic psychology: Critical issues in theory and re-
search (pp. 3–14). Thousand Oaks, CA: Sage.
Brock, D. B., Manton, K. G., & Woodbury, M. A. (1990). Demographic
characteristics. In J. Cornoni–Huntley, D. G. Blazer, M. E. Lafferty, D. F.
Everett, D. B. Brock, & M. E. Farmer (Eds.), Established Populations
for Epidemiologic Studies of the Elderly, volume II: Resources data
book (NIH Publication No. 90–495, pp. 8–19). Washington DC: U.S.
Department of Health and Human Services.
Brock, D. B., Wineland, T., Freeman, D. H., Lemke, J. H., & Scherr, P. A.
(1986). Demographic characteristics. In J. Cornoni-Huntley, D. B.
Brock, A. M. Ostfeld, J. O. Taylor & R. B. Wallace (Eds.), Established
Populations for Epidemiologic Studies of the Elderly: Resources data
book (NIH Publication No. 86–244, pp. 11–32). Washingon, DC: U.S.
Department of Health and Human Services.
Brown, G. W., Harris, T. O., & Hepworth, C. (1994). Life events and en-
dogenous depression: A puzzle reexamined. Archives of General Psy-
chiatry, 51, 525–534.
Chiriboga, D. A. (1984). Social stressors as antecedents of changes. Jour-
nal of Gerontology, 39, 468–477.
Chiriboga, D. A. (1996). In search of continuities and discontinuities across
time and culture. In V. Bengtson (Ed.), Adulthood and aging: Research
on continuities and discontinuities (pp. 173–199). New York: Springer.
Chiriboga, D. A., & Cutler, L. (1980). Stress and adaptation: Life span per-
spectives. In L. Poon (Ed.), Aging in the 1980s: Psychological issues
(pp. 347–362). Washington, DC: American Psychological Association.
Chou, K.-L., & Chi, I. (2001). Financial strain and depressive symptoms in
Hong Kong elderly Chinese: The moderating or mediating effect of
sense of control. Aging & Mental Health, 5, 23–30.
Glass, T. A., Kasl, S. V., & Berkman, L. F. (1997). Stressful life events and
depressive symptoms among the elderly. Journal of Aging and Health,
9, 70–89.
Hazuda, H. P, Stern, M. P., & Haffner, S. M. (1988). Acculturation and as-
similation among Mexican Americans: Scales and population-based
data. Social Science Quarterly, 3, 687–706.
Hobfoll, S. E. (1998). Stress, culture, and community: The psychology and
philosophy of stress. New York: Plenum Press.
Holmes, T., & Rahe, R. (1967). The Social Readjustment Rating Scale.
Journal of Psychosomatic Research, 11, 213–218.
Kamarck, T. W., Shiffman, S. M., Smithline, L., Goodie, J. L., Paty, J. A.,
Gnys, M., et al. (1998). Effects of task strain, social conflict, and emo-
tional activation on ambulatory cardiovascular activity: Daily life con-
sequences of recurring stress in a multiethnic adult sample. Health
Psychology, 17, 17–29.
Kaplan, H. B. (1996). Perspectives on psychosocial stress. In H. B. Kaplan,
(Ed.), Psychosocial stress (pp. 3–24) San Diego, CA: Academic Press.
Keefe, S. E., & Padilla, A. M. (1987). Chicano ethnicity. Albuquerque:
University of New Mexico Press.
Kemp, B., Staples, F., & Lopez-Aqueres, W. (1987). Epidemiology of de-
pression and dysphoria in an elderly Hispanic population: Prevalence and
correlates. Journal of the American Geriatrics Society, 35, 920–926.
Krause, N. (1997). Anticipated support, received support, and economic
stress among older adults. Journal of Gerontology: Psychological Sci-
ences, 52B, P284–P293.
Cohen, S., Frank, E., Doyle, W. J., Skoner, D. P., Rabin, B. S., & Gwaltney,
J. M., Jr. (1998). Types of stressors that increase susceptibility to the
common cold in healthy adults. Health Psychology, 17, 214–223.
Connell, C. M., & Gibson, G. D. (1997). Racial, ethnic, and cultural differ-
ences in dementia caregiving: Review and analysis. The Gerontologist,
37, 355–364.
Cornoni-Huntley, J., Blazer, D. G., Lafferty, M. E., Everett, D. F., Brock,
D. B., & Farmer, M. E. (Eds.). (1990). Established Populations for Ep-
idemiologic Studies of the Elderly, volume II: Resource data book (NIH
Publication No. 90–495). Washington, DC: U.S. Department of Health
and Human Services.
Cornoni-Huntley, J. C., Brock, D. B, Ostfeld, A. M., Taylor, J. O., & Wal-
lace, R. B. (Eds.). (1986). Established populations for epidemiologic
studies of the elderly: Resources data book (NIH Pub No. 86–2443).
Washington, DC: U.S. Department of Health and Human Services.
Crum, R. M., Anthony, J. C., Bassett, S. S., & Folstein, M. F. (1993). Popu-
lation-based norms for the Mini-Mental State Examination by age and
educational level. Journal of the American Medical Association, 289,
2386–2391.
Cuellar, J., Arnold, B., & Maldonado, R. (1995). Acculturation Rating
Scale for Mexican Americans–II: A revision of the original ARSMA
scale. Hispanic Journal of Behavioral Sciences, 17, 275–304.
Cuellar, J., Harris, L. C., & Jasso, R. (1980). An acculturation scale for
Mexican American normal and clinical populations. Hispanic Journal
of Behavioral Sciences, 2, 199–246.
Krause, N., Liang, J., & Gu, S. (1998). Financial strain, received support,
anticipated support, and depressive symptoms in the People’s Republic
of China. Psychology and Aging, 13, 58–68.
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New
York: Springer.
López-Aqueres, W., Kemp, B. J., Plopper, M., Staples, F., & Brummel-
Smith, K. (1984). Use of health care services by older Hispanics. Jour-
nal of the American Geriatrics Society, 32, 434–440.
Lowenthal, M. F., & H aven, C. (1967). Interaction and adaptation: Inti-
macy as a critical variable. American Sociological Review, 33, 20–30.
Marin, G., & Gamba, R. J. (1996). A new measurement of acculturation for
Hispanics: The Bidimensional Acculturation Scale for Hispanics
(BAS). Hispanic Journal of Behavioral Sciences, 18, 297–316.
Markides, K. S., & Black, S. (1996). Aging and health behaviors in Mexi-
can Americans. Family & Community Health, 19, 11–18.
Markides, K., Rudkin, L., Angel, R. J., & Espino, D. V. (1997). Health sta-
tus of Hispanic elderly in the United States. In L. Martin & B. Soldo
(Eds.), Racial and ethnic differences in the health of older Americans
(pp. 285–300). Washington, DC: National Academy Press.
Marsella, A. J., Friedman, M. J., Gerrity, E. T., & Scurfield, R. M. (Eds.).
(1996). Ethnocultural aspects of posttraumatic stress disorder: Issues,
research and clinical applications. Washington, DC: American Psycho-
logical Association.
Mendes de Leon, C., & Markides, K. S. (1988). Alcohol consumption and
physical symptoms among Mexican Americans. Drug and Alcohol De-
pendence, 16, 369–379.
Miller, T. Q., Markides, K. S., & Black, S. A. (1997). The factor structure
of the CES-D in two surveys of elderly Mexican Americans. Journal of
Gerontology: Social Sciences, 52B, S259–S269.
Mirowsky, J. (1996). Age and the gender gap in depression. Journal of
Health and Social Behavior, 37, 362–380.
Murrell, S. A., Norris, F. H., & Hutchins, G. L. (1984). Distribution and de-
sirability of life events in older adults: Population and policy implica-
tions. Journal of Community Psychology, 12, 310–311.
Padilla, A. M. (1980). The role of cultural awareness and ethnic loyalty in
acculturation. In A. M Padilla. (Ed.), Acculturation: Theory, models
and some new findings (pp. 47–84). Boulder, CO: Westview.
Pearlin, L. I., Menaghan, E. G., Lieberman, M. A., & Mullan, J. T. (1981).
The stress process. Journal of Health and Social Behavior, 22, 337–356.
Perez-Stable, E. J., Marin, G., & Marin, B. (1994). Behavioral risk factors:
A comparison of Latinos and non-Latino whites in San Francisco.
American Journal of Public Health, 84, 971–975.
Culbertson, F. M. (1997). Depression and gender: An international review.
American Psychologist, 52, 25–31.
Dohrenwend, B. P., Raphael, K. G., Schwartz, S., Stueve, A., & Skodol, A.
(1993). The structured event probe and narrative rating method for
measuring stressful life events. In L. Goldberger & S. Bernitz (Eds.),
Handbook of Stress (pp. 174–199). New York: The Free Press.
Escobar, J., Burnam, A., Karno, M., Forsythe, A., Landsverk, J., & Gold-
ing, J. (1986). Use of the Mini-Mental Status Examination (MMSE) in
a community population of mixed ethnicity: Cultural and linguistic ar-
tifacts. Journal of Nervous and Mental Disease, 174, 607–614.
Ferraro, K. F., & Su, Y.-P. (1999). Financial strain, social relations, and
psychological distress among older people: A cross-cultural analysis.
Journal of Gerontology: Social Sciences, 54B, S3–S15.
Flack, J. M., Amaro, H., Jenkins, W., Kunitz, S., Levy, J., Mixon, M., et al.
(1995). Panel I: Epidemiology of minority health. Health Psychology,
14, 592–600.
Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). Mini-Mental
State: A practical method for grading the cognitive state of patients for
the clinician. Journal of Psychiatric Research, 12, 189–198.
Garza, R. T., & Gallegos, P. I. (1995). Environmental influences and per-
Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for
research in the general population. Applied Psychological Measure-
ment, 1, 385–401.

P568
CHIRIBOGA ET AL.
Shah, B. V., Barnwell, B. G., & Bieler, G. S. (1997). SUDAAN user’s manual,
Release 7.5. Research Triangle Park, NC: Research Triangle Institute.
Smith, J. P. (1999). Healthy bodies and thick wallets: The dual relation be-
tween health and economic status. Journal of Economic Perspectives,
13, 145–166.
Therrien, M., & Ramirez, R. R. (2000). The Hispanic population in the
United States: March 2000. Current population reports (Publication
No. P20-535). Washington, DC: U.S. Bureau of the Census.
Thoits, P. A. (1995). Stress, coping and social support processes: Where are
we? What next? Journal of Health and Human Behavior, Special Issue,
53–79.
Turner, R., Lloyd, J., & Lloyd, D. A. (1995). Lifetime traumas and mental
health: The significance of cumulative adversity. Journal of Health and
Social Behavior, 36, 360–376.
Turner, R. J., Wheaton, B., & Lloyd, D. A. (1995). The epidemiology of so-
cial stress. American Sociological Review, 60, 104–125.
U.S. Bureau of the Census. (1991). Race and hispanic origin (Census Pro-
file No. 2). Washington, DC: Department of Commerce.
Vega, W. A., Kolody, B., Aguilar-Gaxiola, S., & Catalano, R. (1999). Gaps
in service utilization by Mexican Americans with mental health prob-
lems. American Journal of Psychiatry, 156, 928–934.
Weissman, M. M., Myers, J. K., Tischler, G. L., Holzer, C. E., III, Leaf, H. O.,
& Brody, J. (1985). Psychiatric disorders (DSM-III) and cognitive im-
pairment among the elderly in a U.S. urban community. Acta Psychiat-
rica Scandinavica, 71, 366–379.
White, L. R., Kohout, F., Evans, D. A., Cornoni-Huntley, J., & Ostfeld, A. M.
(1986). Related health problems. In J. C. Cornoni-Huntley, D. B.
Brock, A. M. Ostfeld, J. O. Taylor, & R. B. Wallace (Eds.), Established
Populations for Epidemiologic Studies of the Elderly: Resources data
book (NIH Publication No. 86–2443, pp. 129–165). Washington, DC:
U.S. Department of Health and Human Services.
Yu, E. S. H., Liu, W. T., Levy, P., Zhang, M.-Y., Katzman, R., Lung, C.-T.,
et al. (1989). Cognitive impairment among elderly adults in Shanghai,
China. Journal of Gerontology: Social Sciences, 44, S97–S106.
Zamanian, K., Thackrey, M., & Starrett, R. (1992). Acculturation and de-
pression in Mexican American elderly. Clinical Gerontology, 11, 109–
121.
Zarit, S. H., & Zarit, J. M. (1998). Mental disorders in older adults: Funda-
mentals of assessment and treatment. New York: Guilford Press.
Ward, C. (2001). The A, B, Cs of Acculturation. In D. Matsumoto (Ed.),
The handbook of culture & psychology (pp. 411–445). New York: Ox-
ford University Press.
Received May 29, 2001
Accepted May 15, 2002