.....1... .L... :......
2.... . .1:

: ........
f
I, .
...::

q. .z: ...

 

 

.13.... .v...
2.5:.

 

:uz... . .,. ~
2. .. ,.

 

 

\. ti...

 

\ . ,

 
 

I: V B
. :L::,.

 

I. ,7.
If; r,

 

 

, LIBRARY
i”kahgan it-
11:1?th [‘3'

___. __ ...-v-O OW

This is to certify that the

dissertation entitled

THE INTERSECTION OF MENTAL HEALTH,'PREGNANCY AND RACE:
A CONTEXTUAL INVESTIGATION OF THE RELATIONSHIPS BETWEEN
SOCIAL FACTORS AND
MATERNAL PSYCHOLOGICAL DISTRESS

presented by

Renee Branch Canady

has been accepted towards fulfillment
of the requirements for

Doctor of Philosophy degree in Socioloqy

’ ' , M
Major professor

Date g‘XQ'Cj/

MSU is an Affirmative Action/Equal Opportunity Institution 0-12771

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

DATE DUE DATE DUE DATE DUE

.t a :2," «SI (3‘:

 

6/01 c1/CIRCIDmDIJop65-p. 15

 

 

 

THE INTERSECTION OF MENTAL HEALTH, PREGNANCY AND RACE:
A CONTEXTUAL INVESTIGATION OF THE RELATIONSHIPS BETWEEN
SOCIAL FACTORS AND
MATERNAL PSYCHOLOGICAL DISTRESS

BY

Renee Branch Canady

AN ABSTRACT OF A DlSSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Sociology

2001

Chairperson: Clifford L. Broman

ABSTRACT

THE INTERSECTION OF MENTAL HEALTH, PREGNANCY AND RACE:

A CONTEXTUAL INVESTIGATION OF THE RELATIONSHIPS BETWEEN
SOCIAL FACTORS AND

MATERNAL PSYCHOLOGICAL DISTRESS

BY

Renée Branch Canady

The birth of a baby. A time of rejoicing and gladness,

usually. The impetus for this study is another, still too

frequent pregnancy outcome, infant death. Infant mortality

rates among African American babies are double that of
Whites and have shown minimal improvement over the last
decade. Data collected from 1980—1990, show that black
infant mortality decreased by only 19% during that time
while white infants experienced an improvement of 30%.

Consequently, black babies remained twice as likely as

whites to die and the black—white gap in infant mortality

persisted.

This quantitative study utilizes a subsample of 5200
cases from the National Maternal and Infant Health Survey
(NMIHS), a nationally representative study conducted by the
National Center for Health Statistics. Analyses focus on
respondents to the pregnancy status measure who were between
the ages of 19 and 43 years.

Research in the field of pregnancy outcomes has
historically emphasized biomedical and clinical measures.

The more recent research agenda in pregnancy outcomes has

Renée B. Canady

emphasized the role of “social factors” in poor pregnancy

outcomes. While genetic and biological factors are

important in disease etiology, many race and health
researchers have argued that social factors may be more

important. This study investigates the relationships

between key social measures (race, socioeconomic status, and

marital status) and psychological distress. In particular,

this study investigates maternal psychological distress as a

potentially vital factor in the experience of pregnant

women. Considerable work has been done on depression after

pregnancy, however the work which addresses depressive

symptoms and maternal psychological distress prior to

delivery is comparatively less. In addition, analyses focus

on the disparate experiences of African American women.
This investigation provides empirical evidence that

black women experience significantly greater psychological

distress than women of other races. In addition, the

psychological state of women is predicted more by the social

contexts of their lives than by the immediate experience of

pregnancy. Furthermore, this study confirms the importance

of the role of marriage in the psychological well—being of

women who are pregnant. Lastly, it is the experience of

prior pregnancies, rather than the current pregnancy, which

effectively predicts psychological distress in women.

ACKNOWLEDGMENTS

For wisdom, encouragement, unyielding standards; for

true mentorship and training — I thank you, Dr. Cliff

Broman. For time, guidance, and support - I extend my

Stephen Gold, Dr. Angela Mertig, and Dr.

gratitude to Dr.
Linda Beth Tiedje.
I am indebted to the National Institutes of Mental

Health and the American Sociological Association for the

opportunities afforded me as a Mental Health Fellow in the

Dr. Felice Levine, Dr. Edward

Minority Fellowship Program.

Marguia and Dr. Alfonso Latoni have been instrumental in my

development as a scholar; my appreciation to each of you and

your staffs.

For walking with me through this journey, pushing me

when I was weary and loving me in spite of myself— I love

you, Mark Canady, my best friend. For understanding and

waiting patiently and loving me just the same ~ you are my

joy, Marcus, Alexander and Wesley.

For covering me in prayers and interceding for me

spiritually, emotionally and physically - I thank my family

and friends (you know who you are!). For longsuffering and

support above and beyond the call of duty, I thank my staff

and colleagues in the MSU College of Nursing.

iv

For strength sufficient for the task; for peace that

PASSED all understanding; for being my keeper and provider —

Jesus Christ, my Lord and Savior.

I give you praise,

In Mbmoriam

For Mark Howard Canady, II, whose brief life of seven

world of motherhood. My memory

months, escorted me into the

of him underscores the drive for this study. His life

frames the heart of who I am now and remains an integral
part of the heart of our family.

For Bernice White Branch, who compressed a life time of

mothering me into 23 short years and who epitomized the

privilege of motherhood in action and deed. Her blend of

parenting and friendship inspires me and I strive daily to

emulate her perfection of these roles.

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES
INTRODUCTION

CHAPTER 1:
RATIONALE/CONCEPTUAL FRAMEWORK

CHAPTER 2:
LITERATURE REVIEW . .
Mental Health and Pregnancy . . . . . . . . . .
Maternal Mental Health: The Stress and Social
Support Connection . . . . . . . . . . .
Maternal Mental Health: The Connection to the
Sociology of Mental Health . . . . . . . .
Maternal Mental Health: The Applicability of
Depression and Pregnancy Investigations . .
The Intersection of Mental Health, Pregnancy and Race
The Issue of Context . . .

CHAPTER 3:

METHODOLOGY . . . . . . . . . . . . . . .

The National Maternal and Infant Health Survey

The Present Study: Pregnancy and Psychological

Distress . . . . . . . . . . . . . . . .

Sample

Measures . . .

Data Analysis Plan . . . . . . . . . .
Considerations of Interaction Effects

CHAPTER 4:
RESULTS . . . .
Descriptive Findings . . . . . . . . . . . .
Factors Predicting Psychological Distress in Women
Progressive Adjustment through Hierarchical
Multiple Regression . . . . . . . . . . . . .
Interaction Effect and Subset Analysis by Race
Additional Findings from Interaction Effect
Analysis . . . . . . . . . . . . .
Psychological Distress in Pregnant Women
Contributions of Pregnancy Outcomes to Maternal
Distress

SUpplemental Considerations: Questions of Employment

and Parental Status

vi

Page

viii

14
14

16
19

24
35
44

50
50

54
58
62
77
79

83
83
92

92
99

101
105

109

116

TABLE OF CONTENTS (cont.)

Page
CHAPTER 5:
DISCUSSION . . . . . . . . . . . . . . . . . . . . . . 121
Conclusion . . . . . . . . . . . . . . . . . . . . . . 137
LIST OF REFERENCES . . . . . . . . . . . . . . . . . . 140

vii

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

TABLE

10:

ll:

12:

13:

LIST OF TABLES

Page

SAMPLE DEMOGRAPHICS COMPARISON . . . . . . . . . 6O
DESCRIPTIVES - TOTAL SAMPLE . . . . . . . . . . 64
DESCRIPTIVES BY RACE . . . . . . . . . . . . . . 86
DESCRIPTIVES BY PREGNANCY STATUS . . . . . . . . 86
BIVARIATE CORRELATIONS . . . . . . . . . . . . . 87
CES-D Scores - Descriptive analysis of High

School Attainment and Pregnancy Status by Race
[CES—D scores; (cases)] . . . . . . . . . . . . 91
CES—D Scores within Marital Status and

Pregnancy Status by Race . . . . . . . . . . . 93
Model 1.1-Multiple regression of Psychological
Distress on Race and Pregnancy Status . . . . . 93
Model 1.2—Multiple regression of Psychological
Distress on Race, Pregnancy Status, and
Socioeconomic Status (SES) measures (maternal
education and household income) . . . . . . . . 93
Model 1.3—Mu1tiple regression of

Psychological Distress on Race, Pregnancy

Status, Socioeconomic Status (SES) and

Marital Status . . . . . . . . . . . . . . . . 94

PROGRESSIVE ADJUSTMENT: Demonstration of
measure of Psychological Distress among women
categorized by race and pregnancy status,
adjusting for socioeconomic status, marital

status and age 98
Non-additive Effects of Race and Pregnancy on
Psychological Distress 100
Non-additive Effects of Pregnancy and Marital 02
1

Status on Psychological Distress

viii

LIST OF TABLES (Continued)

Page

TABLE 14: Regression Coefficients for Current Pregnancy

Status regressed on CES—D by Race and Marital
Status . . . . . . . . . 104

TABLE 15: Model 2.1- Multiple regression of
Psychological Distress on Race among
Pregnant Women . . . . . . . . . . . . . . . 106

TABLE 16: Model 2.2-Multiple regression of
Psychological Distress on Race and
Socioeconomic Status (SES) measures
(maternal education and household income)
among Pregnant Women . . . . . . . . . . . . 106

TABLE 17: Model 2.3—Multiple regression of
Psychological Distress on Race,
Socioeconomic Status (SES), Marital Status
and Age among Pregnant Women . . . . . . . . 106

TABLE 18: PROGRESSIVE ADJUSTMENT: Psychological
Distress among Pregnant women categorized
by race and adjusting for age, socioeconomic
status and marital status . . . . . . . . . 108

TABLE 19: Dummy regression with Progressive Adjustment
of Psychological Distress on Race, Pregnancy
Status and Pregnancy History . . . . . . . . 111

TABLE 20: Regression of Psychological Distress on Race,
Pregnancy Status and Pregnancy History . . . 111

TABLE 21: Reiterative Analysis for Role of Prior
Pregnancy Outcome . . . . . . . . . . . . . 114

TABLE 22: Exploratory Sub—sample Analysis . . . . . . . 119

ix

LIST OF FIGURES

Page
Figure 1: Proposed model of Mental Health, Pregnancy,
and Race . . . . . . . . . . . . . . . . . . . 9
Figure 2: Revised model of Mental Health, Pregnancy,
and Race . . . . . . . . . . . . . . . . . . 125

INTRODUCTION

The birth of a baby. A time of rejoicing and gladness,
usually. The impetus for this study is another, still too
frequent pregnancy outcome, infant death. The United States
(U.S.) is a country of comparative wealth and privilege yet,
Black women are twice as likely to experience a poor
pregnancy outcome than their White counter—parts (David,
1991). Infant mortality rates among African American babies
are double that of Whites and have shown minimal improvement
over the last decade (Carmichael, 1998; Hogue, 1993). Data
collected from 1980-1990, show that Black infant mortality
decreased by only 19% during that time. However, during
that same time period, Whites infants experienced an
improvement of 30%. Consequently, Black babies remained
twice as likely as Whites to die and the Black—White gap in
infant mortality has persisted (Flack, 1995). Overall U.S.
infant mortality rates reached an all time low of 9.8 deaths
per 1000 births in 1989, yet the infant death rate among
African Americans looms precariously above that all time low
at 12.8 deaths per 1000 (Hummer, 1993).

Much of the motivation for recent work in this field
springs not only from a recognition of poor outcomes
overall, but also from persistent differences in birth
outcomes among racially diverse populations. In the late

1800’s, Black babies were twice as likely to die as White

1

infants and incredulously, over a century later, Black
babies remain twice as likely to die as White infants
(Hogue, 1993). The U.S. Surgeon General has mandated the
reversal of racial and ethnic disparities in disease
outcomes, including infant mortality, by 2010 (Arno, 1998;
House, 1998). Not only are health professionals and
researchers now charged with continuing to improve prenatal
outcomes among all Americans, but also with abolishing the
gap between Black and White health status measures.

Research in the field of pregnancy outcomes has
historically emphasized biomedical and clinical measures.
The lack of consistency and resultant confusion in measuring
pregnancy outcomes is openly acknowledged in the field
(Bramwell, 1997). Outcomes such as pre—term delivery, intra-
uterine growth retardation, low birth weight, and other
clinical indices have been studied in relation to factors
such as access to prenatal care and health behaviors such as
smoking, alcohol consumption, and diet. The more recent
research agenda in pregnancy outcomes has emphasized the
role of “social factors” in poor pregnancy outcomes. While
genetic and biological factors are important in disease
etiology, many race and health researchers, such as David
Williams (1992a), have argued that social factors may be
more important. The production of disease and illness
clearly results from the convergence of multiple social

dynamics upon biological factors. It is increasingly

2

apparent that the disparity of pregnancy outcomes among
Blacks and Whites will not be effectively mediated without
evaluating the underlying problems of social structure and
status and their contributions to the health and illness
experience.

As a mental health researcher, sociologist and the
survivor of a seven month old, pre-term son, I have chosen
to narrow the focus in this very large problem to the
question of maternal psychological distress and race and
their contribution to the pregnancy experience. This study
investigates the relationships between key social measures
(race, socioeconomic status, and marital status) and
psychological distress in a sample of pregnant women. In
order to expand the understanding of this multi-factorial
problem, this study will investigate maternal psychological
distress as a potentially vital factor in the experience of
pregnant women. Considerable work has been done on
depression after pregnancy (post-partum depression), however
the work which addresses depressive symptoms and maternal
psychological distress prior to delivery is comparatively
less (DaCosta, 2000; Seguin, 1999; Kitamura, 1996). The
preponderance of work on depression and pregnancy is also
clinically focused, concentrating on various medical
diagnoses of mental illness and pregnancy related mood
disorders (Steinberg, 1999). Little work in this field has

gone the additional step of investigating maternal

3

psychological distress as a factor in the pregnancy outcome
(Benedict, 1999; Schmeelk, 1999). Furthermore, the
attention to race in the present study attends to an
additional factor which has frequently been overlooked,
despite disparate trends in the African American community
as previously noted. Much of the recent work on race and
mental health during pregnancy has taken a transcultural
approach or focused on international populations(Dragonas,
1997; Séguin, 1999; Zachariah, 1996; Aberbigbe, 1995;
Kitamura, 1996; Thorpe, 1992; Wijma, 1992), and describes
experiences which may not be applicable to the lives of
African American women. Finally, social support is a common
theme in this evolving body of work which endeavors to
address psychosocial risk factors in pregnancy. The social
support literature actually comprises a separate and
distinct area of sociological research. It is the goal of
the present work to integrate these sparse and eclectic
bodies of work in a manner which will frame relevant but as
yet unaddressed questions, and thereby advance the current

state of research on pregnancy and psychological distress.

CHAPTER 1:

RATIONALE AND CONCEPTUAL FRAMEWORK

Why study psychological distress as a variable in
pregnancy? Furthermore, why study race as a factor in
psychological well—being and pregnancy outcome? In addition
to the previously presented demographic trends, consider the
following: 1) Maternal depression has been related to low
birth weight and pre—term delivery (Dunkel—Schetter, 1998;
Steer, 1992); 2) African American women are more likely to
experience poor pregnancy outcomes than women in other
racial groups (Rowley, 1993; David, 1991); 3) the population
of women at risk for poor pregnancy outcomes tends to also
be the same population at risk for increased depression
(i.e., low-income women) (Orr, 1995); 4) depression and
excessive stress have been associated with diminished
participation in prenatal care (Sword, 1999), a factor
demonstrated to impact pregnancy outcomes (Rowley, 1995);
and 5) maternal depression has been associated with general
maternal health following childbirth (Schmeelk, 1999).
Maternal depression is also a significant factor in
important outcomes such as marital and family satisfaction,
and child development. In spite of these associations,
little attention has been given to the role of maternal
psychological distress, including depression, in studies of
pregnancy outcomes. Calls within the literature for

5

increased studies of mental health in African American women
speak to a gap in the current research and acknowledge the

limited understanding of the forces at work in the mental

health outcomes of Black women (Dunston, 1990).

The present study integrates several bodies of

literature. First, the focus on psychological distress is

informed by the general mental health research, by some

components of the clinical research on depression and

depressive symptoms, and by the sociological study of

stress. Second, the attention to race is informed by

emergent work on race and health (including mental health)

and the evolving research on the health consequences of

discrimination and racism. Lastly, the pregnancy research

which addresses pregnancy outcomes, pregnancy and stress,

and pregnancy and depression, attend to the specific

questions of pregnancy considered in this study. The
preponderance of the comparatively small number of existing
studies on pregnancy and mental health tend to be focused on

lower socioeconomic samples, a disappointing fact in light

of the trend of poor pregnancy outcomes among African

American women across all economic segments (Schoendorf,

1992). The present study draws heavily on a nascent body of

work which strongly suggests the existence of a relationship

between maternal depression and poor pregnancy outcomes.

This body of literature further suggests that depression is

a likely part of the pregnancy experience of women (Orr,

1995; Hobfoll, 1995; O’Hara, 1986).

Another important theoretical orientation of this study

is derived from the sociology of mental health (Pearlin,

1981; Pearlin, 1999; Aneshensel, 1999; Aneshensel, 1992).

Notwithstanding the diverse disciplinary interests in mental

illness and mental health, the sociology of mental health

provides an important contribution in this field of study.

Social characteristics have been strongly correlated with

the causes and consequences of mental illness (Aneshensel,

1999). As a result, sociological inquiry in this area has

been distinguished by investigations of social

characteristics which delineate social status and social

roles. Aneshensel (1999) describes two goals of the

sociological study of mental health which are particularly

applicable to the sociological study of pregnancy and

pregnancy outcomes: 1) the identification of social strata

who are at—risk; and 2) the explanation of why differentials

exist. Social strata refer to those systems and

characteristics which stratify and divide individuals in

this society such as socioeconomic status, gender, age, and

The second major characterization of individuals in

race.
sociology addresses the social roles they occupy, including
marriage, parenthood, and employment (Pearlin, 1989).

the sub—

Furthermore, within the sociology of mental health,

speciality of stress research also informs the present work.

7

Aneshensel (1999) further posits that the unique

sociological contribution to the explanation of which

differentials exist and why, lies in the articulation of

issues of context. Contextual analysis facilitates the

meaningful incorporation of multiple structural and

descriptive factors. Simon (1999) defines context as the

social circumstances in which events occur. This is not to

imply that context revolves only around episodic events,

rather it envelopes multiple aspects of one’s life
experiences and the manner with which they interrelate.

Context shapes the personal and social meaning attributed to

events and relationships. Context facilitates the

consideration of phenomenon such as racism and

discrimination (as o osed to ‘race’). Context reco nizes
P

the dynamics poverty and wealth (as opposed to socioeconomic

status); it validates the influence of roles and

socialization (as opposed to gender and age). Context views

the insidiousness of identity within ethnicity, of

inequality within class, and of oppression within nation.

It is the integrated nature of context and its
representation of subtleties which yields its greatest

strengths. Context shapes the manner in which variables of

interest are measured; however, its greatest influence

emanates from the interpretation of findings from those

measures. The sociologically imaginative perspective of

context will be applied in this study and its analyses.

8

The conceptual model guiding the analysis in this study

is illustrated in Figure 1. The model consists of the

following major components: 1) Maternal Psychological

Distress, which serves as the dependent variable of the

study; 2) Pregnancy status, which frames the fundamental

focus of the study; and 3) Race, which defines the research

questions posed. The framework of the model presented also

incorporates social domains in order to contextually situate

the lives of the women surveyed. The conceptual model of

this study is purposefully limited to three social factors

(education, income, and marital status) which have been

empirically shown to affect mental health. Many of the

models presented in the literature are multi—factorial and

superficially review numerous potential influences (Lobel,

1992; Paarlberg, 1995). Others have acknowledged

relationships among the variables of interest here, but lack

a thorough interpretation of those relationships. Narrowing

the focus in this study to measures of socioeconomic status

and marital status facilitates a more thorough analysis of

these particular measures of interest.

 

FIGURE 1
Proposed Model of Mental Health, Pregnancy, and Race
MATERNAL
PREGNANCY +
PSYCHOLOGICAL
STATUS
///, DISTRESS
RACE

Socioeconomic status (measured by education and

household income) and marital status are hypothesized as
external factors which characterize the contextual effects

at work in the pregnancy experience and provide a key

orientation for the study. Fueled by the discourse of

outcome disparities previously noted, considerations of

intra—racial and inter-racial variation will be integral

parts of the analyses of this study. Pregnancy is viewed as

an extrinsic variable in that it is not affected by the

model proposed, but rather, as with race, represents a more

immutable characteristic of the sample pOpulation.

The relationships between socioeconomic status and
mental health, as well as between marital status and mental
health, have been fairly well established. However, the
frequency with which contradictory findings result in these

works also found suggest the need for further deliberation

(Steinberg, 1999; Orr, 1984; Gotlib, 1989). Furthermore,

the health benefits of marital status tend to accrue more

readily to men (Umberson, 1992), a point of critical

consideration in light of the gender orientation of

pregnancy.
The debate over disparate health and pregnancy outcomes
has recently polarized along race and wealth perspectives.

The discourse of the socioeconomic influences on health have

framed a dialectic which has tended to de-emphasize the role

of race in favor of economic explanations (Pappa, 1994;

10

Syme, 1998). Notwithstanding the demographic trends which
place African Americans disproportionately among the poor,
the persistence of racial disparities even when controlling
for education or income is documented (Krieger, 1993) and
sobering. Similarly, marital status is an important
consideration not merely because of its demographic
categorization, nor even because of its great influence upon
the financial well—being of women (Christensen, 1992), but
moreover because of its correlation to physical and mental
well—being (Hu, 1990; Kitson, 1990; Broman, 1993). The
relationship between marital status and health has also been
found to be gender dependent (Bruce, 1992), a
characterization which holds potential relevance in
considerations of pregnancy, given its gender—defined
occurrence. Marital status is also noted as a key factor in
post-partum depression research as well as transition to
parenthood literature (Steinberg, 1999), thus further
justifying its consideration in this case. In addition,
this study is distinguished by including both married and
Single women. Many studies to date have focused exclusively
on either singled or married women, thus minimizing the
ability to consider the role marriage plays in psychological
well—being (Dragonas, 1997; Kitamura, 1996; Turner, 1990).
Acknowledging the clinical and biomedical nature of the
current literature, an additional critical goal of this

study is the introduction of a sociological perspective into

11

the current discourse of pregnancy and pregnancy outcomes.
Consequently, in an effort to avoid the weakness of the
current literature in the over—inclusion of variables, this
study addresses the relationships among two frequently
described factors, in order to facilitate a more thorough
analysis of the dynamics and relationships at work.

The linear causal model presented in figure one
describes three relationships which affect maternal
psychological distress, two of which are direct
relationships and one indirect relationship. In the first
direct relationship, it is believed that pregnancy will
affect psychological distress. Pregnancy, having been
categorized as both a well and sick state (Oakley, 1980;
Myers, 1990), defines a woman’s status, influences her
perceptions of self, her relationships with others and
likely, her psychological well-being. The second direct
relationship suggests that race, as socially defined,
directly affects psychological distress. Substantiated by
the growing evidence of a relationship between race and
physical health, the projection that a comparable
relationship exists between race and mental health is
justified. The indirect relationship portrayed in the
model, implies that race has a likely influence on the
pregnancy experience and as a result, indirectly impacts

psychological well-being.

12

The current state of the literature warrants the
investigation of fundamental relationships and the
contextual analysis which will be applied in this study adds
an additional richness and depth. Consequently, this study
is purposefully exploratory and descriptive in nature. As
described earlier, race and pregnancy remain an encompassing
framework within each of the hypotheses. Thus differential
outcomes in each research question between women who are or
are not pregnant and women who are Black versus those of
other races are fundamental to all analyses. The study will
test the following hypotheses: 1) Measures of psychological
distress in women who are pregnant will significantly differ
from psychological distress in women who are not pregnant
(Hypothesis 1); 2) Maternal psychological distress will
vary by race, socioeconomic, and marital status (Hypothesis
2); and 3) Current psychological distress will be
significantly related to previous pregnancy outcomes

(Hypothesis 3).

13

CHAPTER 2:

LITERATURE REVIEW

Mental Health and Pregnancy

Pregnancy is a gendered experienced (Umberson, 1996)
and as such it can not be divorced from the general
experiences of women. In many ways, the pregnant woman is
objectified in much of the pregnancy literature. The woman
either becomes defined by her internal organs (uterus) and
physiological processes (labor) (Martin, 1997), or is
minimized by the unbalanced emphasis on the infant’s health
(Goldenberg, 1998). In many cases the pregnancy becomes
reified, serving as an entity in and of itself (Mirowsky,
1989). Much of the impetus for this study emanates from the
research findings which describe poor pregnancy outcomes.
However, the focus of this field of research would be
strengthened by the re-conceptualization of pregnancy
outcomes as measures more closely allied with the mother as
opposed to her child, born or unborn. For example, concern
for the woman's mental health and depressive symptoms
(Kennerly, 1989; Aneshensel, 1981), issues of maternal self—
esteem and self-illusion (Smith, 1997; Mamelle, 1989), or
attention to the principal relationships of a woman who is

Pregnant (Henly, 1997; Tiedje, 1990; Aneshensel, 1991;

Tietjen, 1985) are additional constructs which can be
investigated as “pregnancy outcomes” and which portend

14

greater sociological relevance. The pregnancy experience as
characterized by the above features, initiates the
transition to parenthood and correlates not only to
psychological distress in women but subsequently to the
well—being of the infant and family (Lederman, 1998). This
shift in focus to maternal psychological distress as the
pregnancy outcome of interest frames the distinct
contribution of this study in particular and of sociology in
general to the current field of research. The investigation
of psychological distress in women who are pregnant
formulates a model which gives voice to the women as
subjects in and not merely objects of the pregnancy.

The fact that pregnancy is such an expected and usually
routine experience within the human experience may serve as
both a benefit and a liability in its appeal as a research
topic. Pregnancy has been problematized, medicalized, and
even minimized as an investigative subject. The
investigation of a specialized topic such as the one posed
here, i.e., psychological distress in pregnant women, with
particular emphasis on Black women, presents clear
challenges in establishing a substantive research
foundation. As a consequence, some of the literature cited

throughout this chapter may be more dated than would

otherwise be ideal. Nevertheless, this fact reiterates the

need for additional study and highlights the timeliness of

the topic.

15

Maternal Mental Health: The Stress and Social Support
Connection

The dynamics of mental health and pregnancy have most
commonly been investigated as studies of stress and social
support, as investigations of pregnancy and depression along
the pregnancy continuum, including antenatal and postnatal
stages, or in ancillary matters such as maternal adjustment
(Beck, 1999). This section will ensue with a summary of the
stress and social support literature as applicable to
pregnancy and conclude with a more specific critique of the
pregnancy and depression literature.

Kitamura (1996) defines antenatal depression as “a type
of depressive illness which appears through reaction to
stressful situations, in this case pregnancy” (1996:737).
Thus pregnancy is both stressful and a stressor. Within the
stress and social support literature, particularly that
which has addresses considerations of pregnancy, findings
are either inconclusive or contradictory (Oakley, 1990;
Norbeck, 1989a). The most widely accepted model of stress
and social support presents social support as a buffer or
mediator of stress (Cobb, 1990; Cohen, 1990). It is clear
that stress and pregnancy have a relationship but
contradictions in work which has attempted to clarify that
relationship are obvious (Coyne, 1991; Crnic, 1983; Norbeck,
1989b). Within the biomedical literature various measures

Of stress and psychological distress, such as trait anxiety

l6

are investigated as correlates to pregnancy outcome.
However, these measures have shown no relationship to birth
outcomes in some studies (Lobel, 1994) while other studies
have demonstrated positive correlations to abnormal birth
outcomes (Paarlberg, 1995; Cooper, 1996). Unfortunately,
these findings are somewhat clouded by the inconsistent
application and definition of measures, including the
alternate use of trait anxiety as a measure of depression,
as opposed to stress (Steer, 1992). Stress has been applied
both as a measure of depressive symptomatology and as a
cause of depression.

Life events are also applied as a common social measure
of stress in the pregnancy literature (Sheehan, 1998;
O’Hara, 1986; Ramsey, 1986). Summary findings on the affect
of life events, in many cases show a positive association
with low birth weight, pre—term delivery and gestational age
(Lobel, 1994). Women face numerous stressors (e.g.,
structural, role-related, etc.) and the addition of role
strain, another social construct of stress, which can be
caused by pregnancy and motherhood may constitute an
additional stressor in their life experience (Pritchard,
1994). For example, Brett et al. (1997) investigated
occupation as a role strain, finding that Black women are at
greater risk for job strain than White women. One might
then extrapolate and conclude that pregnant African American

women who hold high strain jobs face an even greater risk of

17

mental distress, a potential underlying cause of their
disproportionate experience with poor pregnancy outcomes.

The distinctions by race are particularly important
given that considerations of race in the sociological study
of stress yield mixed results (Brown, 2000; Broman, 1995;
Aneshensel, 1981; Dunston, 1990; Henly, 1997). Some studies
reveal no racial differences in depression which results
from stress and others demonstrate an attenuation of
differences once race is controlled (Ulbrich, 1989). These
broader trends of racial variation in stress among women are
likely to be exacerbated by the inclusion of considerations
of pregnancy status. Interestingly, the stress literature
suggests that the role of social support as a mediator among
pregnant women may vary by race (Lobel, 1994; Lobel, 1992;
Bryce, 1988). Social support varies not only in importance
but by source among racially and socially distinct
populations (Sagrestano, 1999). The relationship between
social support and psychological distress suggest the
potential for similarly unique findings in psychological
well—being among these diverse populations as well. In
addition, the positive finding among these important
variables in the sociology of mental health give rise to
questions about the potential underlying effects at work in
psychological distress among racially distinct women. The
investigation of social correlates of psychological distress

in the present study is likely to broaden the understanding

18

of other related concerns such as stress and social support.
Moreover, because much of the work to date is inconclusive
and small in scope, the establishment of empirical
validation of current findings is a pressing research need.
Maternal Mental Health: The Connection to the Sociology,of
Mental Health

Little empirical work investigates the impact, causes,
or consequences of depression in pregnant women, a point
broadly acknowledged in the literature (DaCosta, 2000;
Hobfoll, 1995; Séguin, 1999; Powell, 1992; Kitamura, 1996;
Elliott, 1999; Zuckerman, 1989). Distinct from the clinical
research on post-partum and perinatal depression, there
presently exists only a small body of research which
investigates psychological well-being during pregnancy.
Significant contributions are found in the discipline of
nursing, which has formulated studies germane to sociology
such as pregnancy wantedness, maternal attachment and their
roles on family and marital satisfaction (Beck, 1999;
Hummer, 1995; Barber, 1999). Sociological work has most
commonly concentrated on the experiences of adolescents
(Turner, 2000; Turner, 1990) and historically viewed
pregnancy a potential crises for both marriage and the
family in general (Lips, 1985; Hobbs, 1976). Sociology has
also been interested in pregnancy as the transition to
parenthood, not purely in the pregnancy experience itself

(Simon, 1999; Tietjen, 1985)-

19

Regardless of the theoretical or disciplinary bent, few
of the current investigations in this area have looked
broadly at social antecedents to psychological distress
during pregnancy. Still fewer are represented in the body
of work which investigates psychological distress among
women of color throughout the duration of their pregnancy.
This study purposefully employs the concept of psychological
distress as opposed to clinical depression, a decision which
presents challenges in applying the current pregnancy and
mental health literature. While there are studies of
pregnancy and emotional or psychological well—being which
can be found (Dragonas, 1997; Thorpe, 1992), most are
focused on clinical diagnoses of mental health conditions
(Aderibigbe, 1993; Gotlib, 1989; Hedegaard, 1993; Kumar,
1984). Furthermore, many of the studies use diagnostic
tools for the measurement of mental state instead of the
broader screening tool used in this study. Despite the fact
that only 10% of pregnant women in the United States will
meet diagnostic criteria for major depression, the
preponderance of the literature speaks in terms of
diagnosable clinical conditions (Weissman, 1995). Pregnancy
studies have found higher rates of clinical depression.
DaCosta et al. (2000) found that 25% of their sample
experienced clinical depression, however, the depression

typically resolved after the delivery.

20

The transitional nature of pregnancy has long gone
uncontested from both a role and a status perspective (Lips,
1985; Hobbs, 1976) and an application of a broader mental
health measure, i.e. psychological distress as opposed to
clinical depression, suggests that thirty to 38% of pregnant
women experience some level of depressive symptoms
(Zuckerman, 1989) during this transitional experience. The
use of the broader measure of psychological distress
facilitates the evaluation of a greater number of women, all
of whom are sure to have some emotional and social response
to the pregnancy experience. Even clinical researchers have
suggested that limiting one’s focus to clinically defined
post—partum depression in women increases the risk of
underestimating the psychological distress associated with
this phase of life (Raskin, 1990).

The commitment to a sociological orientation in the
construct of psychological distress requires the
consideration of other related areas of investigation within
the field of sociology. One such relevant body of work is
the sociology of stress and mental health. Investigations
of stress, as a nondiagnostic construct, inform this study;
however, its omission of comprehensive considerations of
race is especially problematic for the present study.
Considerations of both race and socioeconomic status (SES)
could further inform the contextualization of both stress

and mental health and provide relevance to the study of

21

mental health during pregnancy. For example, it is
generally accepted that race and SES affect stress and
consequently, psychological well—being. For low income
women, the experience of poverty induced stressors have been
shown to be stronger predictors of depression than acute
stressors (Séguin, 1999). Ulbrich’s (1989) finding that SES
interacts with race to increase the symptoms of distress,
demonstrates the contextual influences at work and holds
intriguing relevance to the stress and pregnancy research.
The deficiencies in the literature are a telling
assessment of the current state of work on social factors
and pregnancy outcomes, and they suggest future directions
for research. While many disciplines hold a vested interest
in this field, including psychology, anthropology,
epidemiology and others, there is indeed a great need for
the renewed contributions of sociology to this critical area
of study. Much of the research on pregnancy outcomes uses
approaches which emphasize interpretation at the individual
level, the very essence of the biomedical approach, and
neglect the structural and collective influences which frame
even micro-level phenomenon. In addition, the simplistic
assembling of social variables found in most of the work on
social factors, does not equate to sociological scrutiny.
Sociological stress research looks at how the organization
of people’s lives can be disrupted in the stress process.

Pregnancy as a stressor is experienced differently and may

22

be predictably influenced by issues of social stratification
and social domains (Pearlin, 1989). Collins (1993) sums up
the current state of affairs cogently and succinctly:
“Pregnancy is not uniformly stressful for all women"
(1993:1245). Attention must be focused on identifying those
at risk for stress or for the purposes of the present study,
psychological distress, and understanding what differentials
exist and why.

Role strain theory defines role strain as those
stressors that arise within the context of institutionalized
social roles (Aneshensel, 1999). This study endeavors to
describe and analyze the general mental health of pregnant
women in light of the socially defined roles they hold and
the societal stratifications which categorize their
experiences. The social circumstances in which events occur
or in which strains persist constitute the social context
and ultimately shape the personal meaning and emotional
significance of the stressors (Simon, 1999). Women
frequently find themselves in multiple roles, including
those defined by marital status, parental status, and
employment status (Tiedje, 1990). The literature suggests
that women who experience depressive symptoms are likely to
be managing a number of stressors in their lives (Warren,
1997; Neumann, 1986; Weissman, 1977). Sociological findings
in these areas hold important implications for pregnancy

outcomes and serve as key considerations for this work.

23

Maternal Mental Health: The Applicability of Depression in

Pregnancy Investigations

The average age of diagnosis of clinical depression in
women ranges from 20 to 30 years of age, a span which
parallels the years of child-bearing age for women
(Zuckerman, 1989). Many of the symptoms associated with
depression, such as fatigue and loss of appetite, are also
common symptoms of the first trimester of pregnancy (Orr,
1995). Thus, causal relationships between pregnancy and
mental health may become clouded. Although this study
avoids the application of diagnostic indices of
psychological distress, the independent variable is measured
with a construct which has been strongly correlated to
clinical depression (Radloff, 1977). Thus, the literature
which describes those pregnancy experiences which are
characterized by depression and depressive episodes are
important to note.

Despite the potential overlap in measuring the outcomes
of pregnancy related symptoms and depressive symptoms, the
potentially negative role of depression in pregnant women
(Zuckerman, 1989) and the growing evidence for the
association of depressive symptoms with pregnancy outcomes
(Dunkel—Schetter, 1998; Steer, 1992; Orr, 1995; Orr, 1989;
Levin, 1988) necessitate continued investigation in this

area. Depression and pregnancy studies often present

24

findings which are ambiguous, and some studies have failed
to find an association between depression and pregnancy
(DaCosta, 2000), while others present evidence for positive
associations or still, varied associations among different
groups of women (Gotlib, 1989). The need for further study
is clear.

As with the other work reviewed to this point, the
studies of pregnancy and depression form a helpful yet
enigmatic body of literature when applied to this study.
Notwithstanding studies of post-partum depression, studies
which have investigated depression and psychological
distress during the pregnancy experience are in the minority
and many suffer from methodological and theoretical
challenges. More commonly, the work on depression and
pregnancy defines the pregnancy experience broadly and
beyond the date of delivery. Thus post—partum
considerations often cloud the examination of pregnancy
experience itself on the mother (Hobfoll, 1995; Gotlib,
1989; Steinberg, 1999). Even investigators of post-partum
depression acknowledge the need to expand the investigation
of many of the questions raised by post—partum depression to
the perinatal time frame, including the pregnancy itself
(Powell, 1992). Interestingly, depression during pregnancy
is often used as a control variable or as a predictor of

post—partum depression (Aberibigbe, 1993; DaCosta, 2000;

25

Pfost, 1990; Seguin, 1999). The preponderance of work views
the continuum of pregnancy through the post—partum stage and
fails to consider etiological causes of depression in
pregnancy at the start of that continuum.

Only a few works were found which actually address
studies of depression or psychological distress during the
pregnancy (Kitamura, 1996; Thorpe, 1992). It is intriguing
that both of these studies were completed in international
contexts (Greece and Japan). Many others have investigated
treatments of depression during pregnancy (Steinberg, 1999)
but fail to investigate causes and correlates of depression.
Nevertheless, much of the work on post-partum depression
provides insight to potential mechanisms at work during the
pregnancy. DaCosta’s work describes a unique model for the
relationships between depression and pregnancy. Her work on
post—partum depression suggests that biological factors
provide a greater influence than psychosocial factors for
maternal depression which begins following delivery.
However, for post—partum depression which represents the
continuation of maternal psychological distress initiated
during the pregnancy, psychosocial factors are likely of
greater importance than biological risk factors (DaCosta,
1999). Furthermore, while post-partum depression has held
much of the attention of mental health researchers, there

appear to be distinct correlates for depression during

26

pregnancy as compared to depression after delivery (Gotlib,
1989). Most frequently, psychological distress,
particularly as measured by non-diagnostic means, during the
pregnancy is precipitated by psychosocial factors
(Aderibigbe, 1993; Gotlib, 1989).

Additional review of the literature endeavored to
identify current investigations which focused solely on the
mental health of pregnant women as measured within the
duration of the pregnancy experience. Much of the work in
this area has approached the depression question within the
continuum of pregnancy, including antenatal, post natal, and
perinatal experiences. Most frequently, the studies
identified addressed the concerns of both prenatal and post—
partum observations (Raskin, 1990; DaCosta, 2000; Hobfoll,
1995; Steinberg, 1999; Gotlib, 1989; Lips, 1985). As a
matter of fact, much of the work on depression during
pregnancy approaches the question primarily as it informs
the post—partum experience (Elliott, 1996; Najman, 2000), a
perspective which is different from the interests of this
investigation. Exhaustive searches of both social and
clinical data bases reiterate the dearth of work which
investigates psychological distress exclusively within the
time frame of the woman’s pregnancy.

Upon review of the small contingent of work which

investigates this mental health during pregnancy, one is

2'7

struck by the dependency of these studies on relatively
dated work to validate their own studies. It is notable
that even among the few studies completed in the 1990’s,
most rely on and cite investigations from the 1980’s to
frame the premise of their study (Stancil, 2000; Zachariah,
1996; Kitamura, 1996; Thorpe, 1992). Eighty—four percent of
the articles cited in four of the primary works which
investigate psychological distress during pregnancy were
published before 1989, a trend which again underscores the
need for more immediate work in this area. Furthermore,
once identified, it was still a greater challenge to isolate
studies with comparable design and purpose to ours. Many
addressed very psychological questions (Zachariah, 1996) or
focused on other ‘social factors' such as social support,
stress and coping (Thorpe, 1992) within the framework of the
investigation. Nevertheless, several intriguing trends are
noted in the literature.

Kitamura et al. (1996) and their study of depression in
early pregnancy is perhaps the most methodologically
consistent study identified. Unlike other works (Lips,
1985; Seguin, 1999), Kitamura's work more particularly
specifies the experience of depression during pregnancy,
suggesting that such symptoms occur within the first
trimester. Kitamura, et al. are most interested in

diagnostic phenomenon; however, their decision to utilize a

28

nondiagnostic measure in their study (Self-rating Depression
Scale (SDS) bodes well for the present work. Kitamura
(1996) found no significant difference in depression
measured across educational level, age (at marriage), or
income among the women with high or low SDS scores. The
counterintuitive nature of this finding and its
contradiction of the broader literature (Gotlib, 1985),
bring into question the comparability of their sample.
Despite their methodological rigor, social and cultural
factors distinct to their sample of Japanese women, may be
the source of this unexpected outcomes and as a result,
minimizes it's applicability in an American sample.

Nonetheless, the positive correlations found in
Kitamura's study are worth noting. Depression during the
early stage of the pregnancy is associated with the first
pregnancy and with there being no prior delivery (i.e.,
termination or miscarriage). In addition, it is related to
low partner intimacy, being remarried, being younger, and
with some measure of pregnancy wantedness. Finally, a
higher prOportion of cases (depressed women) had been
working outside the home at the time of the questionnaire,
but this was not a significant predictor of depression in
this study.

Additional psychosocial factors have been shown to

contribute to variations in the individual pregnancy

29

experiences of women (Thorpe, 1992). Thorpe et al. (1992)
present what is ostensibly a study of stress (particularly
of life events) and social support in pregnancy, and exclude
considerations of depression. Notwithstanding, the Thorpe
study contributes to the understanding of psychological
well-being in pregnancy and relates to other investigations
which have correlated depression in pregnancy with adverse
life events (Aderibigbe, 1993). Kitamura’s (1996) evidence
that depression is predicted by having no prior deliveries
also relates to the findings of Thorpe et a1. (1992). Given
the dearth of race based considerations in this field, the
Thorpe study is also of importance in its contextualization
of the experiences of pregnant women. The study compares
the psychological status of Greek women to British women; a
model which may provide insight into the experiences of
Black and White women in America by reconceptualizing
findings as representative of the experiences of majority
versus minority populations.

Only 18% of the total sample investigated by Thorpe
(1992) was represented among the highest depression scores,
thus psychological distress in this population was
relatively low. British pregnant women also had lower
depression scores than the general population, reiterating
the strength of their psychological well-being. However,

marked differences were found between Greek and British

30

women. Clear associations between psychological well-being
and life events were evident. No association was found
between social support and psychological well-being in the
Greek sample but there was a negative association between
poor psychological well-being and social support in British
sample. Differences by culture were seen between British
and Greek women with social support more predictive in
British women and life events more predictive for Greek
women.

The absence of studies which exclusively investigated
depressive symptoms during pregnancy, forced a closer
scrutiny of the prenatal components of studies of depression
along the pregnancy continuum as well as examining the
prenatal considerations in studies focused on post—partum
depression and distress. Most investigators expect that
pregnancy will impact the emotional or psychological state
in some way (Lips, 1985; Hobfoll, 1995; Gotlib, 1989).
Explanations for psychological changes during pregnancy
appear to be characterized by disciplinary inclinations and
offer different rationale within varying disciplines (Lips,
1985). Substantiations of depression in pregnancy also
appear to fluctuate with the timing of measurement during
the pregnancy. Despite Kitamura’s (1996) evidence of
psychological distress during early pregnancy, most studies

suggest a decline in depression over the course of the

31

pregnancy (Hobfoll, 1995). More recent measurements of
psychological distress during the third trimester remained
consistently above previously published estimates of
depression during pregnancy, with upwards of 23% of women
demonstrating some level of distress (Hobfoll, 1995).
DaCosta (1995) provides intriguing and rich findings which
demonstrate that 25% of women experienced depression during
the pregnancy, a percent which declined after delivery.
However, the study is comprised of mostly a White, well—
educated sample of very small size (n=100). The validity of
applying these findings to in a study such as ours which
investigated race based differences, is questionable.

Lips (1985), however, found that types of emotional
symptoms manifested during pregnancy were not consistent
across the three stages of pregnancy measured during their
investigation. Most variability was found in feelings
measured early in the pregnancy which were related to
physical factors (felt fat, felt overweight, felt sick,
etc.). However, measurements later in the pregnancy and
immediately post-partum, suggest women were more likely to
express increased negative emotions. Lips (1985) summarizes
her findings by stating that pregnancy is distinctly
associated with the reporting of particular symptoms (albeit
physical and not emotional symptoms). The generalizability

of her findings beyond the middle class, White Canadian

32

population are questionably acknowledged by Lips and her
call for further study in other populations and cultures is
heeded in the present study.

Gotlib et al. (1985) found depressed pregnant women to
be younger, less educated, and to have had more children
than their non—depressed peers. However, other studies have
demonstrated no relationship between age and education.
Furthermore, ethnicity, family income, and employment were
not significant predictors of psychological distress in the
literature (Hobfoll, 1995; DaCosta, 2000). DaCosta (2000)
also found no relationship between the number of children
and depressive symptoms. Thus, the pattern of contradiction
in the body of research persists and the need for additional
work to elucidate the dynamics at work are a necessity.

The role of marital status is critically discussed in
the work of Hobfoll (1995) who argues for more investigation
of the effects of marital status in the context of
pregnancy. Steinberg’s (1999) study of clinical treatment
issues in a sample of depressed pregnant women found that
most women in his study were in stable relationships
(married or remarried), calling into question the health
benefits of marital status. In addition, Steinberg asserts
that poverty, and not ethnicity, is the crucial factor in
depression among pregnant women. In light of the scarcity

of studies which consider race and ethnicity, this

33

investigator is critical of Steinberg’s failure to
substantiate his rationale of a declining significance of
race. Allowing for the epidemiological trends of poor
pregnancy outcomes after controlling for education and
income, this researcher remains unconvinced by Steinberg’s
position.

In summary, the depression and pregnancy literature is
a unique blend of varied methods, theories and disciplines.
Its effective application to this study is challenged by
several inherent characteristics. First, the measurement of
mental state is diverse, ranging from self-administered
tools to professional diagnostic interviews. Secondly,
considerations of pregnancy across the continuum of pre—
partum to post-partum complicate the applicability of
findings. While numerous publications hold titles which
encompass both pregnancy and after pregnancy orientations,
it is clear that the most important perspective in most of
the dually titled pieces is post-partum analyses. Finally,
appallingly few works have incorporated considerations of
race. Among those including ethnicity and culture,
international samples were utilized (Greek, British,
Japanese, and Canadian). Situating the present study within
the broader infant mortality puzzle of the United States is
an ultimate goal, thus the international perspectives of

these parallel works may be a disadvantage.

34

The Intersection of Mental Health, Pregnancy and Race

The intersection of mental health, pregnancy, and race
culminates at a point which endeavors to integrate two
rather incongruent findings. Intriguingly, the relationship
between race and pregnancy is the inverse of the
relationship between race and mental health. Seminal
epidemiological studies suggest that race serves as a
protective factor against depression, i.e., African American
populations had better mental health in many measures of
psychological well-being than their White peers (Robins,
1991; Sullivan, 1998). Sociological work has also
documented that unlike trends found among other measures of
morbidity and mortality (Villarosa, 1994; Pappas, 1993;
Goldberg, 1992; Williams, 1992a; Cowie, 1992), no excess
prevalence of depressive disorders exists among communities
of color (Takeuchi, 1998; Somervell, 1989).

Despite its favorable contribution to mental health,
being Black remains a liability in pregnancy outcomes
(David, 1991; Rowley, 1993). The present study strives to
merge considerations of mental health and pregnancy in light
of their contradictory relationships to race, a challenge
infrequently undertaken in the current literature. It is
increasingly apparent that racial disparities in illness in
general and in pregnancy outcomes in particular, will not be

effectively mediated without evaluating fundamental features

35

of this intersection, namely, the underlying problems of
status and social structure and their contributions to the
health and illness experience. It is also this point of
intersection which characterizes the contextual nature of
sociological work on this subject. The theoretical
intersection of race, mental health and pregnancy will be
initially addressed before moving to an elucidation of the
domain of context.

A complementary body of work on race and health, based
in the social sciences, endeavors to move beyond the
biological and slightly eugenical orientation of clinical
research and holds great promise for enhancing understanding
and expanding the perspective in the health care disparity
discourse (Williams, 1996; Krieger, 1996; David, 1991;
Broman, 1996). This shift receives support even among
medical researchers (Fiscella, 1996). A Myrdalian dilemma
plagues the African American pursuit of health, happiness
and the American dream as evidenced by the morbidity and
mortality rates of that community. Race again becomes a
“moral problem” representing the inequities in outcome and
opportunity for good health and well-being among Blacks in
America as distinct from those of White Americans (Myrdal,
1944). These inequities beg for constant vigilance.
“Health research that will advance an understanding of the

role of race as a social category must seek to identify the

36

ways in which economic, ideological, political, and cultural
forces as well as racial discrimination, shape the daily
experiences of people in ways that promote illness or hinder
wellness" (Jackson, 1996:132). The present study
investigates the question of maternal psychological distress
in light of this charge, endeavoring to present analyses
which have been interpreted from a sociological orientation
and which applies an understanding of the dynamics of
numerous social forces.

There is little research which empirically investigates
the social production of health and illness in general and
even less which investigates the social etiology of
pregnancy related conditions and diseases. Epidemiologic
work on infant health has documented the increased risk of
pre-term delivery and low birth weight among women of color
(Carmichael, 1998; Hogue, 1993). However, the broader
literature has tended to extrapolate findings from other
general studies often applying many conclusions to pregnancy
erroneously. It is presumed that if Black, Latina, and poor
women are generally at—risk for poor health outcomes, then
they, in all likelihood will also be at risk for poor
pregnancy outcomes. These conclusions are made in spite of
the dearth of definitive studies which have investigated the
pregnancy experience of women in general, and African

American women in particular.

37

Unfortunately the current body of literature most
commonly utilizes race and class as control variables,
rather than as definitive or predictive of pregnancy
experience findings. Little attention is directed at
interpreting the meaning of race and social class in
pregnancy, notwithstanding the efforts of numerous authors
to describe study populations ethnically, racially, and
economically (Lobel, 1994). One of the few documented
associations is one of increased depressive symptoms
correlating to increased pre—term delivery and negative
health behaviors among African American and Latina women
(Hoffman, 1996; Orr, 1995; Orr, 1989). A second finding
which holds relevance for women of color suggests that
social support benefits appear to more readily accrue to
women who are not categorized as high risk by obstetric
criteria (Hoffman, 1996). Ironically, Black women are more
frequently categorized as “high risk” and would by this
finding, appear less likely to benefit from social support,
a conclusion again, contradicted in other work. Additional
studies which attempt to extend the discourse by focusing
exclusively on the relationship between maternal
psychological distress and pregnancy outcomes suggest that
psychological distress is deleterious to pregnancy outcomes
for African American women only and not for White women

(Orr, 1995). A thorough analysis of the underlying

38

mechanisms of action in the experiences of African American
women if done well, further enhances the understanding of
the pregnancy experiences of all pOpulations and will
facilitate more completely appreciation of the macro- and
micro—level consequences for pregnant women.

The current literature is also void of a recognition of
the breadth and scope of social factors, and little work has
clarified the relationships between social factors and
general health or the variations in relationships among
racial and ethnic populations. While it appears that Black
women may be affected by different factors than White women
(Hoffman, 1996), the scOpe of that variability is unclear.
For example, the Black—White gap is smallest for Black women
who did not complete high school and greatest for Black
women with college degrees (Krieger, 1993), another counter-
intuitive finding given the socioeconomic paradigms which
are commonly presented as explanatory factors (Syme, 1998).
The complicated nature of the relationships at work are
illustrated by such paradoxical findings.

Dr. Diane Rowley, in her opening remarks at the Centers
for Disease Control benchmark conference, “The Social
Context of Pregnancy among African American Women” (Atlanta,
GA, February 18-19, 1999) encouraged the group of invited
scholars to search not only for the causes, but for the

“causes of causes” of poor pregnancy outcomes. Whether

39

social factors have a direct effect (cause) or indirect
effect (cause of causes) on health outcomes, her insistence
that this predominately biomedical audience begin to
consider the role of social factors is monumental. The
lived experiences of many women, in concert with a expanding
body of literature, suggest that not only may social factors
be causes of causes, they may indeed be direct causes
themselves in this bewildering trend.

Much attention has recently been directed toward the
controversy over the use of race in medical and other
research (Bhopal, 1998; Fullilove, 1998; Krieger, 1999).
Pregnancy, as a condition plagued by race-based inequities
and racially disproportionate trends, presents an important
circumstance in which to consider the role of racism as well
as gender discrimination. Williams (1993) describes race as
a “societally constructed taxonomy that reflects the
intersection of biology, culture, socio-economic, political,
and legal determinants as well as racism (Williams,

1993:10). Racism and discrimination become the reasons why
different groups of shared race or ethnicity may be more
closely identified with certain risk factors and poor
outcomes. This orientation shifts racism and discrimination
into an explanatory model and away from a causal model. It
also elevates the discourse above the traditional

orientation of investigating racial differences in disease

40

or the social patterning of disease by race (Krieger, 1993).
While beyond the scope of this study, a commitment to the
study of race as racism will also continue to battle the
residual beliefs of genetic inferiority that persist in the
interpretations of disparate health outcomes (LaVeist,
1996).

Nevertheless, considerations of racism and
discrimination are critical in the comprehension of dynamics
at the intersection of mental health, pregnancy and race.
Investigations of the relationship between racism and health
are particularly germane to understanding the
disproportionate number of negative pregnancy outcomes in
women of color (Williams, 1996; David, 1991; Murrell, 1996a;
Krieger, 1996). Landrine et al. (1995) found that the
relationship between sexist discrimination and symptoms was
contingent upon the age and ethnicity of the woman. Black
women experienced higher levels of sexism in close
relationships (Landrine, 1995) which has implications for
another finding: the greater susceptibility of Black women
to negative life events among members of their immediate
support network (Krieger, 1993). There is little empirical
validation of the stress experiences of Black women to date
(Dunston, 1990) and the need for further work in this

population is apparent.

41

Additional support for study in this area arises from
the work of Murrell et al. (1996b). Her qualitative study
of childbearing women identifies two relevant issues: 1) the
pervasiveness of the stereotypes of pregnant African
American women; and 2) the totality of the experience of
racism (Murrell, 1996b). Furthermore, Thompson’s (1996)
work finds that those who perceive of experiences of racism
are more susceptible to measurable psychological distress.
Considering findings such as these is important in the
interpretation of any study of mental health and pregnancy,
especially those such as my own, which strive to understand
the particular experiences of women of color.

Within the bounds of race and pregnancy, the role of
discrimination in pregnancy outcome has received even less
attention to date. One of the first, if not the only
published study of discrimination and pregnancy outcome
concludes that maternal perception of exposure to
discrimination during pregnancy may be associated with low
birth weight in their infants (Collins, 2000). This study
is an important start; however its involvement of only 94
women, and its exclusive concentration on low—income women
in Chicago, limit its strength. The use of low—income,
urban populations (Orr, 1995; Collins, 2000) has
predominated the literature, and still few have presented

meaningful conceptual models which inform the experiences of

42

African American women who are pregnant. This is a critical
oversight in light of the afore-mentioned disproportionate
representation of women of color in the pregnancy morbidity
and mortality statistics.

Interestingly, Orr and James (1984) found that two
measures on the CES-D scale tended to have more extreme
measurements. The measures in questions were, “Do you feel
you are just as good as other people” and “Do you have hOpe
for the future.” While these researchers did no analysis of
responses to these questions by race, these measures may be
applicable proxies to common measures of discrimination
(Krieger, 1990). Because this atypical application of the
CES-D tool has not been validated in the literature, this
technique will not be applied in the present study.

However, their intriguing suggestion, that women of
different races may hold concerns about different issues, is
an important consideration. Furthermore, Lips (1985) found
that the emotional dynamics of pregnancy tend to be
represented by more individual measurements. Thus, common
responses to questions such as those mentioned may indicate
a shared experience which defines both the pregnancy and the
mental health status of these women. Depression may serve
as a risk marker of pregnancy outcomes as opposed to a risk
factor, and these relationships may be more salient among

specific groups of women, e.g., African American women.

43

The failure of the pregnancy and mental health work,
especially the stress and pregnancy research, to consider
issues of race is indefensible. Race is addressed
demographically and descriptively, however, few
investigators have addressed race analytically in their
work. “There are few studies in which instruments are used
in ethnic, cultural, or age groups dissimilar from those in
which the instruments were originally developed” (Lobel,
1994:233). Similarly, the sociological work on stress and
race is in need of further amplification. The opportunity
to merge studies of race and psychological well—being in the
context of pregnancy holds great promise for both bodies of
literature.

The Issue of Context

In addition to the conceptual contributions of the
sociology of race and the sociology of mental health, the
sociological elucidation of issues of context serves to
enrich the growing body of pregnancy research. The National
Institutes of Mental Health (NIMH) through its Behavior
Science Workgroup of the National Advisory Mental Health
Council (1999) lists “understanding how social or other
environmental contexts influence etiology and prevention of
mental illness” as one of its three priority behavioral
research areas. For the purposes of this study, the

contextual and conceptual interpretations are not to be

44

confused with environmental assessments or even evaluations
of community level measurements, as is aptly done in
considerations of social capital and health (Lomas, 1998;
Kawachi, 1997). Rather, social context analysis addresses
1) social stratifications, including those contained within
status, 2) social institutions, including associated roles,
and interpersonal relationships, especially those which
evolve from roles and institutions (Pearlin, 1989).

Most contextual studies have only addressed structural
aspects of context, neglecting the sociocultural aspects
(Simon, 1999). A thorough sociological assessment of
context is not simply, the macro— versus the micro—, but
instead speaks to the interrelationship and overlap of
variables and the creation of new domains of influence
within the lives of these women.

Context is defined as those social circumstances in
which events occur (Simon, 1999). Sociologists Upchurch and
Aneshensel (1999) more succinctly describe context as the
social conditions that covary within structural attributes.
This is not to suggest that context evolves only around
acute episodes, but rather, it includes multiple aspects of
life's experiences. While it is clear that contexts are
composed of the settings of one’s lives, sociologists must
at all costs, avoid the temptation for narrow orientations

which interpret context only at the environmental level.

45

Race frequently characterizes context and thus shapes
personal and social meaning. Context moderates, or
exacerbates, the impact of life experiences, a point
demonstrated in considerations of race as context. This
framework, thus challenges one to look beyond race as the
cause of illness, but rather to view race as the medium
which constructs illness. The context of race represents a
social location, as opposed to a structural location.

Aneshensel further argues that “social arrangements and
processes are fundamental to understanding the causes of
mental illness and its consequences” (1999z4). Social
context (Thorpe, 1992) has been implicated in explanations
of emotional difficulties and individual differences in
emotional reactions to pregnancy. Mothers' attitudes toward
pregnancy and motivations for becoming pregnant have also
been shown to relate to psychological well-being during
pregnancy and may very well be socially determined. Social
context of pregnancy determines whether motherhood is viewed
as a loss or gain, a positive or negative event (e.g., the
view of pregnancy as a hindrance to productive
social/economic contributions to society and other cases
where relatively low value is accorded pregnancy in many
modern western societies).

Contextual factors have included frameworks such as

marital status and socioeconomic status (Sagrestano, 1999)

46

and will be applied similarly in the present study. Each
level of progressive adjustment in this study controls for
a contextual variable in association with a demographic
characteristic (e.g., income and education paired to equate
to socioeconomic status (SES) and the pairing of marital
status with age). Sagrestano (1999) calls for the inclusion
of more contextually relevant demographic factors in the
study of psychosocial factors and birth outcomes.

The literature has failed to clearly investigate the
contextual influence of the woman’s environment.
Identifying those potentially influential social factors
which define the context for pregnancy is critical since
stressors within the woman’s social environment are likely
to increase depression, (Zuckerman, 1989). Much of the
previous work has included only married women, neglecting
the important contribution of the experiences of single, and
perhaps unemployed or under—employed women who are pregnant
(Sanderson, 1991). Furthermore, pregnancy itself may not
categorically influence the mental health state of women but
rather may have varying influence over the duration of the
pregnancy. For example, relationships between depression
and pregnancy outcome have been identified for some stages
of pregnancy and not for others (Raskin, 1990; Orr, 1995;

O’Hara, 1984; Hedegaard, 1993). The critical time of

47

pregnancy has most consistently been reported as increasing
toward the end of pregnancy (O'Hara, 1984; Hedegaard, 1993).

In addition, external variables appear to have varying
influences at different stages of the pregnancy (Lips,
1985). The relationship between maternal stress and
pregnancy outcomes has differed when stress exposure
occurred late in pregnancy as opposed to earlier exposures
(Hedegaard, 1993). Still other researchers speak to the
evidence that women with high stress before they conceive
tend to be at greater risk for poor outcomes (Collins,
1993). Researchers are cautioned against erroneously
attributing changes in maternal psychological measures to
the woman’s pregnancy without critically considering other
environmental and social factors which may play a greater
role in maternal depression and psychological distress
(Lips, 1985).

Sociologist Ann Oakley has long argued for the
consideration of socio—structural and contextual influences
on pregnancy (Oakley, 1980). The strength of the context
model is found in its representation of the subtleties.
Context allows the inclusion of phenomenon like racism and
discrimination (as opposed to ‘race’), it identifies
poverty, wealth, and status (as opposed to more traditional
measures of SES), it validates role socialization (as

Opposed to gender and age). Context sees the insidiousness

48

of identity within ethnicity, of inequality within class,
and of oppression within nation. It is this integrated
nature of context which yields its greatest strengths. The
present study demonstrates somewhat of a methodological
bridge. Limited by constraints of secondary data analysis,
traditional measures are used in this investigation,
however, the sociological imagination and contextual
interpretation of findings promote it contribution to this
broader discourse. Thus, the on-going deliberation of
issues of context are not only theoretically warranted, but
are encouraged by numerous research and public policy

domains, including those suggested by the NIMH.

49

CHAPTER 3:

METHODOLOGY

The National Maternal Infant Health Survey

The present study utilizes data from the National
Maternal and Infant Health Survey (NMIHS) conducted by the
National Center for Health Statistics (Sanderson, 1998).
The NMIHS, a nationally representative study, was conducted
to investigate factors related to poor pregnancy outcome and
holds a wealth of supplementary data which have yet to be
analyzed. Given the breadth and magnitude of this federal
survey, the present study is one of a surprisingly few which
have utilized this rich data set (Covington, 1993; Deal,
1998; Kost, 1995; Visness, 1995; Yu, 1996; Lindberg, 1997;
Kogan, 1994). The publication topics range from breast—
feeding to maternal employment to prenatal care, thus the
present study further expands the utility of this broad
national study beyond its initial stated goals.

The final report for this federal survey and additional
published reports of the study design and methods serve as
the primary source of information about the research methods
and design employed for this data source (Sanderson, 1998;
Sanderson, 1991). The NMIHS is a collaborative initiative
of sixteen federal agencies including the Office of Minority

Health and the National Institutes of Mental Health.

50

Numerous other professional organizations reviewed and
participated in the development of the survey. The NMIHS
consists of a two-wave longitudinal survey of 21,000 women;
11,000 who experienced live births, 4,000 who experienced
fetal deaths and 6,000 who experienced infant deaths. In
addition, the health care providers and health institutions
which delivered care to these women were also surveyed.
Data were initially collected through four sources; maternal
self-administered surveys, vital records reviews, hOSpital
questionnaires, and prenatal care provider questionnaires.
Information from these sources was then linked to vital
records, resulting in an extensive and comprehensive source
of data on maternal and infant health. Birth and death
certificates from the specified time period in 1988 serve as
the initial source of identifying cases. These data were
then linked to other vital records, including hospital
records. The combination of vital records analysis with
separate questionnaires of women and their health care
providers make this one of the more exhaustive sources of
information on maternal and infant health to date.

The NMIHS is the recent most iteration in a series of
previous surveys conducted by the federal government which
addressed issues of natality, infant mortality and maternal
health. The NMIHS is unique, however, from its predecessors

in two important characteristics which are instrumental for

51

the interests of the present study: 1) unmarried women were
included (prior investigations were of married women only);
2) Black infants were over-sampled for all components of the
survey, an important feature because Black infants have
rates of infant mortality twice that of Whites. The first
wave of the survey was completed in 1988. A longitudinal
follow-up of mothers initially surveyed occurred in 1991.

The 1988 survey was implemented within stratified
random samples from the 1988 vital records of 48 states, the
District of Columbia and New York City. The systematic
samples were obtained from a sampling frame of 3,898,922
live births, 38,917 fetal death, and 15,259 infant deaths;
resulting in questionnaires mailed to 13,417 mothers of live
births, 4773 mothers who had late fetal deaths, and 8166
mothers who had infant deaths. Fetal deaths are defined as
those occurring at 28 weeks gestation or more. Data were
weighted to represent national counts as well as to account
for non-response and post-stratification adjustments of
particular characteristics. Response rates varied by birth
category and included rates of 74.2% for the live birth
category, 69.3% for the fetal death category, and 65.3% of
the infant death. Sanderson et al. (1998) report that
mothers were more likely to respond if they were 20-39 years
of age, were White, were married, had fewer children, and

had more years of education.

52

The NMIHS was carried out between January, 1989 and
August, 1990 and assessed 1988 pregnancy experiences. The
1988 baseline survey alone serves as the source of data
analyzed for the present study. The mean interval between
delivery and interview was 16 to 19 months, depending upon
birth category, with the infant death component having the
longest interval and fetal death having the shortest
interval. Questionnaires were mailed as soon as possible
after the vital event in order to minimize recall problems
(Sanderson, 1998). Most measures are worded to account for
a retrospective assessment of experiences occurring
immediately before and within 12 months prior to the
delivery.

A 35-page self-administered questionnaire was mailed to
selected mothers over the age of 15 years (with the
exception of the infant death component which also included
women under 15 years of age). In addition to the survey,
mothers received information brochures, prepaid return
envelopes. Mothers were also offered various participation
incentives. Follow-up to non—respondents included mailing a
second questionnaire, post-card reminders and telephone or
personal interviews. The NMIHS interviews were completed by
trained U.S. Bureau of the Census interviewers under
contract with the National Center for Health Statistics, who

then scheduled a phone or personal interview. The content

53

of the maternal survey was extensive and addressed prenatal
care and health behaviors, hospitalization and delivery
experiences, pregnancy history and experiences, maternal and
paternal characteristics, family income and the babies
health. As mentioned, this extensive survey is unique in
its inclusion of unmarried mothers and also distinct in its
attention collecting information specific to the father of
the infant.

The Present Study: Pregnancy and Psychological Distress

The present study is driven by an interest in

investigating the disparate experiences of two categories of
women: 1) those who are defined by current pregnancy status;
and 2) those who are defined by race, with specific
attention on the experiences of African American women.
This study is further driven by the investigation of a very
fundamental question: What is the psychological well—being
of these women and how do these above mentioned defining
characteristics, as well as additional social factors,
influence their psychological well-being? The National
Maternal Infant Health Survey, while initially conceived for

other purposes, serves as a rich source of data aptly

related to the particular questions which are raised in my

investigation.

Because vital records were used to identify the

original sampling frame for the NMIHS, the present study may

54

more precisely be described as a study of births (MMWR,

1995) and then extrapolated more specifically to a study of
mothers, as Opposed to women in general. The maternal
questionnaire for the NMIHS was an extensive survey covering
a number of topics including the health of the mother and
infant. The present study exclusively uses information
obtained from the maternal self-administered survey and
focuses only upon questions specific to the mother’s
characteristics and experiences for analysis. Data from the
provider and hospital components of the source study are
omitted. As an investigation of maternal psychological
distress during pregnancy, the present study utilizes a sub-
sample of participants in the NMIHS. The following
foundational questions guided case identification: 1) A
specific interest in comparing the experiences of women who
are and are not pregnant, thus cases from the NMIHS who
failed to respond to the question which measured

current pregnancy status were deleted; and 2) A particular
interest in the experiences of women who are beyond
adolescence. Disparate pregnancies outcomes between Black
and White women are not limited to adolescents. On the
contrary, this disparity persists even among college

educated women (Schoendorf, 1992). It is the experience of
these women which frame the interests of this researcher and

as such, are addressed in the present study. Furthermore,

55

the pregnancy experiences of African American adolescents is
more extensively addressed in the literature (Henly, 1997;
Turner, 2000; Jacobs, 1994; Steer, 1990), thus attention to
other components of the population of pregnant women will
enrich the current literature base.

The NMIHS uses a prospective study design, comprised
of two waves of data collection; however, the survey
questions collect information on a past event. The present
study utilizes a retrospective design and links a present
phenomenon (psychological distress) to other events
occurring prior to the development of this outcome of
interest. The dependent variable for the present study, as
measured by CES—D, asks about emotional experiences within
the past week, emphasizing the currency of the measure.
While CES-D is a current measure, the independent variables
are comprised of both retrospective and current measures.
The survey directly assesses the pregnancy experiences of
women in relation to an immediate past pregnancy, thus, many
questions use the suffix “in the 12 months prior to your
delivery.” However, because the dependent variable measures
the mental health state of women who happen to be pregnant
(again) the currency to the primary measure of interest in
the present study is maintained.

A study such as this would certainly have been enriched

by taking advantage of the longitudinal nature of the source

56

study. However, the demographic distributions of the source
study and the particular research questions of interest
preclude this choice. As is expected in any longitudinal
study design, there are notable losses to follow-up between
the first and second waves of the NMIHS, with respondents
decreasing from over 27,000 to just under 10,000 in the
second wave. Despite the loss of cases, the demographic
distribution of remaining cases is relatively consistent
with those from Wave 1. Respondents to Wave II had a mean
age of 25.7 years had an average years of completion of
school of 12.6 years, both measures which are consistent
with Wave 1. In addition, the Wave II respondents had a
slightly higher household income $2000 above the mean in
Wave 1, 58% were married and 48.5% were African American.
The loss in absolute numbers of participants from Wave
I to Wave II poses no particular difficulty for the present
exploratory study. However, additional loss in sample size
is experienced when evaluating the cases as related to the
particular research interest in pregnancy and race. The
number of women who are currently pregnant at the time of
the survey is much smaller in the longitudinal follow-up
data from 1991 (n=722). In addition, the African American
respondents who are pregnant becomes even more smaller, with
only 358 cases identified who meet those two criteria. This

decrease in numbers not only prevents the effective

57

utilization of a prospective panel study design, but it also
reverses the initial preference to use the more recent data
from 1991 for the present study. In order to preserve as
large a sample as possible, the 1988 component of the NMIHS
was used. Thus, the larger number of pregnant respondents
and their racial distributions, in the 1988 as compared to
the 1991 follow-up survey provide the primary motivation in
the decision to use only 1988 responses.
Sample

As mentioned, the NMIHS collected random samples and
over—sampled for Blacks in all components of the study
(Sanderson, 1991). The sub-sample obtained for this
investigation retains the representativeness of the source
study. Approximately seventy-five percent (75.3%) of the
participants in the source NMIHS did not respond to the
question which measured current pregnancy status, leaving
6432 cases available for the present study. Specifically,
women were asked “are you pregnant now” and given the
response categories of 1) Yes, 2) No, 3) Not sure. Many of
the reference studies reviewed for this present work use an
age range of 19—41 for adult studies (Hobfoll, 1995). An
additional 12% of the sample were either younger than 19
years of age or older than 43 years of age, again limiting
the sample by age and further decreasing the number of cases

used in the present study. In the present study, the

58

percent contributed by respondents aged 42 and 43 was
negligible (.OO%) and in the interest of preserving sample
size, these respondents were included in the study sample.
Thus, the 5400 women between the ages of 19 and 43 years who
responded to the survey question which assessed their
pregnancy status during the time of the 1988 baseline survey
of the NMIHS comprise the total sample for the present
study. Selecting for cases which had no missing data for
the CES—D measure, resulted in a final sample size of 5281
which was then used in the analyses of the present study.
The strength of this investigation emerges from the
nature of the data itself. The current literature is
plagued by studies which utilize small, geographically
isolated samples; primarily urban, young, and poor.
Demographic characteristics and distributions of the
respondents to the pregnancy question were very consistent
with the source study (see Table 1). Respondents were
slightly more educated and appear more likely to have
experienced a poor pregnancy outcomes (fetal or infant
death) in their prior pregnancy than the total participants
in the NMIHS. A small percent of African American women
were lost in the sub-sample, however the numbers of women of
color remained robust. Women sampled for the current study
were also less likely to be married than women in the NMIHS,

a point which strengthens the current analysis given the

59

Table 1.

Sample Dempgraphics Comparison.

 

 

 

DESCRIPTIVES NMIHS CANADY STUDY

Race (%African American) 49% 43%

Age 25.6 years 25.9 years

Pregnancy Hx (Live Births) 50.9% 34.0%

Maternal Education 12.9 years 13.7 years

completed completed

Household Income 11.3 11.7
($14,000-15,999) ($14,000-15,999)

Never Married 34.0% 27.8%

Currently Married 56.1% 63.6%

CES-D (total score) 14.82 14.89

N 26355 5281

utilization of this measure as an independent variable. The

strength of the present study is also enhanced by its large
sample size and by the value accrued from its nationally
representative sample source. Analysis of these data will
contribute greatly to the current pregnancy literature as
well as the sociology of mental health literature.

Women in the NMIHS study are differentiated by prior
pregnancy outcome (live birth, fetal death and infant
death). The present study retains respondents from all
pregnancy categories, an important decision in the effort

60

 

to preserve the broad generalizability which results from
this large, national sample. The inclusion of all pregnancy
categories is conceptually justified by the findings of the
work of Covington (1993) which utilized data from this
federal study to evaluate perinatal loss experiences. While
one might expect women in the perinatal loss group to
disproportionately experience negative emotional responses,
the qualitative analysis of respondent remarks by Covington
(1993) found a broad range of both positive and negative
emotional responses to pregnancy as well as to the survey
experience among women who had fetal and infant deaths. The
distribution of responses as described in Covington's
results suggest that inclusion of these groups is unlikely
to skew the responses in the total study sample. Further,
the qualitative techniques employed by Covington are
particularly well suited for assessment of this sensitive
question among women experiencing perinatal loss, thus
adding additional confidence in the decision to include this
population in the study. In addition to increasing the
sample size, inclusion of all pregnancy outcome groups
facilitates retrospective analyses of social factors upon
which pregnancy outcomes are stratified. Finally, it is
worth noting that the present study is comprised of women
who have had a prior pregnancy experience. Within the small

body of work which has endeavored to identify social

61

correlates of psychological distress during pregnancy, is
the intriguing finding that women expecting their first
child are more susceptible to depression (Kitamura, 1996).
Despite the lack of corroborating studies, the nature of the
present study offers control for this possibly confounding
characteristic.
Measures

The Dependent Variable for the present study is
Maternal Psychological Distress. The dependent variable is
measured using the Center for Epidemiological Studies —
Depression Scale (CES-D), a validated scale which measures
depressive symptoms (Radloff, 1977). Measured on an
interval scale, the CES-D score is a composite measure
obtained from the following questions:
“I’m now going to read some statements about the way people
sometimes act and feel. Please tell me the number of days
you felt or behaved this way during the past week, that is,
since the last (Day of the Week),”

1) I was bothered by things that usually don’t bother

me.

N
H

had crying spells.

3) I felt hopeful about the future.
4) I felt depressed.

5) I enjoyed life.

6) My sleep was restless.

62

7) I could not get going.

8) I felt lonely.

9) I felt sad.

10) I felt fearful.

11) I felt that everything I did was an effort.

12) I felt I could not shake off the blues, even with

help from family or friends.

13) I talked less that usual.

14) People were unfriendly.

15) I did not feel like eating; my appetite was poor

16) I thought my life had been a failure.

17) I was happy.

18) I felt that people disliked me.

19) I felt I was just as good as other people.

20) I had trouble keeping my mind on what I was doing.
The response categories for this measure were: 1) Less than
one day, 2) 1—2 days, 3) 3-4 days, and 4) 5-7 days. The
summative variable is coded such that higher scores
correlate to greater depressive symptomatology or
psychological distress. Among the population of the present
study, the use of the CES—D scale presents relatively
consistent results as evidenced by an alpha reliability of
.7971 for the 20 items (see Table 2).

This dependent variable measure serves as an assessment

of psychological distress, speaking more broadly to the

63

Table 2.

Descriptives - Total Sample

 

variable Name Range Mban Std. Dev.

 

Maternal Education

 

(years) 0-18 13.07 1.81
Household Income 1—20 11.71 5.53
Marital Status

(1=Married) 0-1 .636 .481
Race (1=Black) 0-1 .430 .495
Pregnancy Status

(1=Pregnant) 0—1 .412 .492
Maternal Age (years) 19—43 25.9 4.95
Prior Pregnancy

(1=Live Birth) 0-1 .342 .474
CES-D (total score)* 0—60 14.89 12.47*
a = .7871

psychological well-being of the women in this study. The
term “depression” is used categorically and not clinically
and is intended to refer more particularly to the depressive
symptoms experienced by women during pregnancy. As a non—
clinical study, use of the CES-D measure is particularly
appropriate. It is not the goal of the present study to
measure rates of clinical depression or depressive illness,

and the avoidance of such misinterpretation of the findings

of this investigation is a goal. Although particular scores

on the CES—D tend to have a strong correlation to organic

64

mental illness (Radloff, 1977), diagnosis is not the intent
of this study, rather it is the categorization of the
emotional and psychological distress in the women under
investigation. Thus, maternal psychological distress as
measured using the CES—D in this study provides an
assessment provides a broader, nondiagnostic
characterization, with higher CES-D scores suggestive of
greater depressive symptoms and psychological distress.
Given the broader interpretation of mental health in the
present study and its implementation as a self—administered
survey, use of the CES—D scale is theoretically and
methodologically justified (Radloff, 1977).

As a sociological work, the dependent variable of the
present study is intentionally conceptualized generally as
psychological distress. Epidemiological and medical
research has focused on clinical diagnoses of mental
illness, e.g., depression, dysthymia, or manic depression,
while sociological work has addressed mental health more
broadly as psychological distress, depressive symptoms, or
stress. Numerous tools have been developed to measure
mental illness and psychological distress, a point which
complicates the comparison of studies in this field. Much
of the reference data for the present study, including those
from landmark mental health studies, apply clinical

definitions of mental health measured by diagnostic tools.

65

 

The Composite International Diagnostic Interview (CIDI) used
in the National Comorbidity Survey (Robins, 1988) and the
Diagnostic Interview Schedule (DIS) used in the
Epidemiological Catchment Area Survey (Reiger, 1984) both
emphasize diagnostic measures of mental illness. Reviews of
the pregnancy literature reveal works which have used the
above tools, as well as the CES-D, a trend which again
aggravates the ease of application and comparisons of
findings across the literature. In addition, the
predominant focus on clinical measures often omits women who
experience negative symptoms below the diagnostic threshold,
yet who still have their daily lives and relationships
affected (Mirowsky, 1989). The literature recommends that
studies utilizing non-diagnostic tools mediate any potential
methodological problems by applying large sample sizes
(DaCosta, 2000), a criterion which is met in this work.

An additional measurement concern raised in the
literature addresses the appropriateness of symptomatic
levels as defined in the standard applications of these
mental health tools (Radloff, 1977) when applied in studies
of pregnant women. Although most of this discourse takes
place in relation to studies which have addressed diagnostic
measures of mental illness, the concerns are worth

considering here. In the present study, 39% of the

66

 

respondents fell at or above the total CES-D score of 16
which is suggestive of more serious psychological distress
(Radloff, 1977). Given the similarity of symptoms between
pregnancy and depression, many have raised the question of
whether the standard case definition level should be
adjusted up in pregnant women (Hobfoll, 1995). Hobfoll et
al. (1995) raise valid conceptual rationalizations for
adjusting the depressive cut-point in tools when used with
samples of pregnant women, however only a few researchers
implement their recommendation (Raskin, 1990; Orr, 1989).
Others have considered the potentially confounding effects
of pregnancy and depression, but chose not to amend their
analysis beyond acknowledging this methodological concern
(Zuckerman, 1989; Orr, 1995). In addition to adjusting the
total CES-D score used during pregnancy, a second strategy
for compensating for the overlap in pregnancy and depressive
symptoms, e.g., loss of appetite and changes in sleep, is to
use mean CES—D scores for analyses rather than focusing on
the proportion of women above a certain score (Orr, 1995).

Closer review of the literature reveals that adjusting
mental health measures appears most common in studies which
used diagnostic tools (Hobfoll, 1995; DaCosta, 2000) whereas
those using self-assessment measures such as CES-D tend to
omit score adjustment in their efforts to control for

confounding between pregnancy and depressive symptoms

67

 

 

(Raskin, 1990; Orr, 1989). In one study of note,
utilization of the CES-D cut point of 16 was supplemented by
additional analysis at a higher level and included a higher
score of 30 (Benedict, 2000). This decision was based upon
their interest to identify those more closely correlated
with major depression, a clinical goal which is not shared
in this present study. In addition, while there is a
obvious sharing of symptoms during pregnancy with those
characteristic of depression, the individual items of the
CES—D scale do not strongly emphasize those shared symptoms.
For example, of the 20 items, only four appear likely to
represent symptom overlap (I had crying spells, My sleep was
restless, I could not get going, and I did not feel like
eating). Furthermore, these symptoms are also likely to
vary depending upon the stage of gestation. One might also
conclude from the literature that other factors, such as
depressive state prior to conception, may be a greater
factor in measuring maternal depression, than the overlap of
classic depressive and pregnancy symptoms (Hobfoll, 1995;
DaCosta, 2000; Steinberg, 1999).

This researcher is convinced that given the mixed
findings in the literature on the question of CES—D
adjustment when used in samples of pregnant women and in the
absence of methodological works which have specifically

addressed this question, the need for adjustment is not

68

justified in this sample and as such will not be applied in
the present study. The CES~D measure and the data analysis
technique chosen for the current study are broadly validated
and exceptionally robust, thereby reinforcing the confidence
in the decision to withhold adjustment of the CES-D scale
for pregnant women in the present study. As such, the
indicator level for psychological distress in will remain
consistent with the recommendations of the developers of the
tool (Radloff, 1977) and stand at scores of 16 or above.

The sociology of mental health facilitates a more sensitive
assessment of the circumstances of women’s lives in an
attempt to account for more of the variance in mental health
seen between diverse groups of women. This conceptual goal
remains valid despite methodological questions of
interpretation in mental health measures within this
population of pregnant women, further justifying this cut
point decision.

Sociodemographic measures serve as the independent
variables in the present study. The literature shows
support for the contribution of various sociodemographic
factors to the mental health and pregnancy experiences of
women (DaCosta, 2000; Séguin, 1999; Dragonas, 1997; Avison,
1995).

The Independent Variables for the present study are

race, maternal education, household income, and marital

69

status. Descriptive data on all variables considered are
presented in Race frames one of the fundamental premises
of the present study because of the persistent trend of
disparate pregnancy outcomes which are most exaggerated for
African American women. For the present study, race was
dummy coded 1=Black and 0=all other races, emphasizing the
focus on the experiences of African American women. The
sample is relatively balanced in racial distribution, with
43% of the sample comprised of Black women. Household
income and maternal education are used to assess
socioeconomic status in the present study. As is the
standard in the literature, total household income was used
as the measure of income and ranged from less than $1000 per
year to more $60000 per year. Twenty response categories
were provided for household income in the source survey,
with the first category representing income less than $1000
and the last category representing income greater than
$60,000. Intermediate categories increased in $1,000
increments for total incomes less than $10,000, in $2,000
increments for total incomes from $10,000-$20,000 and in
increments of $5,000 for total incomes above $20,000.
Interestingly and rather inconsistent with standard
measurements of income within the literature, is the
decision to specify total household income during the 12

months before delivery. It should be noted that for the

70

purposes of the present study, occupational prestige has
been omitted from the assessment of socioeconomic status.
Within the source data set (NMIHS), the coding for
occupations utilized federal guidelines, however, the large
number of varying categories, in addition to a large number
of missing cases contribute, prompted the decision to
exclude occupation in the assessment of socioeconomic
status. The mean income for this sample was $16,000 to
$17,999 with 49.3% of the sample falling above that
threshold.

The fact that some of the women were currently pregnant
was of little consequence or focus in the NMIHS purpose of
assessing the experiences of the prior pregnancy.
Nevertheless, the atypical definition of “12 months prior to
delivery” is rather astute. The measure then provides
insight to the context of the pregnancy, given that the
measure then covers a time period which encompasses the
gestational period. While women were asked to reflect on
the context of their previous pregnancy, it is plausible
that the experiences of the prior pregnancy, which took
place from 16 to 19 months previously, may be duplicated
during the current pregnancy. Women who were likely to work
more or less, as a result of the needs predicated by
pregnancy are likely to repeat similar behaviors during the

current pregnancy.

71

 

Education is measured in years and ranges from no
education or kindergarten only, to completion of some years
of graduate school. Education was presented categorically.
The original education variable was recoded into the
following groups: 1) 0—8 years of school completed, 2) 9-11
years of school completed, 3) 12 years of school completed
(presumed equivalency to high school diploma completion), 4)
13 to 15 years of school completed (presumed some college
completed), 5) 16 years of school completed (presumed
equivalency to bachelors degree), and 6) 17 to 18 years
completed (presumed equivalency to some graduate school or
graduate degree earned). This recoding is consistent with
the source data with the exception of this investigator's
decision to separate the category of graduate school from
bachelor’s degree completion. Because the literature
suggests poor pregnancy outcomes persist among African
American women (Schoendorf, 1992), despite educational
attainment, the close evaluation of the influences of
education at a variety of post—secondary levels of education
seems instrumental. The study sample is comprised of a
relatively educated group of women with 60% of participants
having completed high school. The mean education level was
represented by those having completed approximately half a

year of education beyond high school completion.

72

Marital status is the third independent variable
applied in the present study. Numerous studies of pregnancy
have controlled for marital status or included only one
marital category by virtue of the study design (Turner,
1990; Kitamura, 1999). This choice is questionable in light
of two solidly established research bases; one that
investigates the roles of health and marital status (Kitson,
1990; Goldman, 1993; Marks, 1996), and one that investigates
social support and pregnancy outcome (Hoffman, 1996;
Sagrestano, 1999; Collins, 1993). The existence of gender-
based variability in marital status and health findings
(Hemstrom, 1996; Bruce, 1992) as well as suggestions of
race—based variability in pregnancy and social support
findings, warrant further consideration of these factors as
put forth in the present study. My study thus evaluates the
likely contextual parameters defined by the presence or
absence of a spouse and the subsequent psychologically
mediated roles facilitated by them.

For the purposes of the present study, the marriage
variable was created from two variables (“Are you married
now?” and “Are you now or have you ever been married”). In
the primary survey, participants were asked “Are you now or
have you ever been married?” (with response options of yes
or no). Respondents answering affirmatively to this

question were then asked “What is your current marital

73

 

status” (with response options of married, widowed,

\\

divorced, and separated). Negative respondents to the are
you now or have you ever been married” comprise the never
married category for purposes of the present study. A new
variable was created which combined the five marital
categories as obtained from these two variables. For
regression analysis, the recoded variable was dummy coded,
with 1=married and O=not married (i.e., either widowed,
separated, divorced or never married). 63.5% of the sample
were married, a distribution which is somewhat unexpected
given the nationally representative sample but suggesting
that as evidenced by this sample, pregnancies still occur
predominately within the confines of marriage. Of the
married respondents 24.4% were Black.

Other important variables for the present study include
current pregnancy status, age, and prior pregnancy outcome.
A natural consequence of the time frame for implementing the
NMIHS, following identification of the cases from vital
records, and the large size of this national sample of women
in child bearing years, is that a large percent of the women
were again pregnant at the time of the study. As a result,
while the primary purpose of the source study is the
evaluation of a pregnancy which has been concluded, the size
and scope of the initial investigation resulted in a cohort

of pregnant women and thus lends itself aptly to the goals

74

 

 

of the present study. Forty—one percent (41%) of the study
sample were pregnant at the time of the survey. The
remaining respondents were either not pregnant or were
unsure of their pregnancy status. As with race and marital
status, the current pregnancy status variable is dummy
coded, with 1=pregnant and 0=not pregnant. Of the
respondents who were pregnant, 38.4% are African American
(n=839), a racial distribution only slightly
disproportionate to the racial representation of 43% African
American in the total study population.

Age serves as a control variable for the present study,
with women less than 19 years of age and older than 40 years
of age excluded from this study. Adolescents or women who
have delayed conception are often at high risk for poor
pregnancy outcomes or for pre—existing clinical diagnoses
(Hobfoll, 1995). Additional supporting research justifies
the selection of this age range for studies of pregnancy in
non—adolescent samples (DaCosta, 2000; Hobfoll, 1995). Age
in the source survey was reported as “age at time of
delivery” and as such is measured similarly in the present
study. The prompt implementation time lines of the NMIHS
survey, discussed previously, assure that this reported age
lies within 12-15 months of the current age of respondents.
The mean age in this sample was 25.9 years with a standard

deviation of 4.95.

75

 

 

The inclusion of the measure Prior Pregnancy Outcome
indirectly situates the present study within the body of
work on psychological distress and pregnancy outcomes.
Participants in the National Maternal Infant Health Survey
are categorized by an event history which was obtained from
the vital records which identified the cases and which
describes the outcome of the index pregnancy as defined by
the NMIHS. This index pregnancy for NMIHS serves as the
pregnancy outcome measure and is defined as the outcome of
the pregnancy immediately prior to the current pregnancy.
Women were grouped into the following categories, based upon
information from vital records: 1)Live birth, 2) Fetal
Death, and 3) Infant Death. 34.2% of the sample were from
the live birth category and 65.8% were either fetal or
infant death cases. This variable was also dummy coded for
regression analysis, with 1=live birth and 0=other pregnancy
outcomes.

It should also be noted that while the original data
set held a wealth of socially defined variables which could
have been considered, there were additional methodological
constraints, in addition to the previously mentioned
conceptual rationale, which limited inclusion of potentially
useful measures. Missing data is a problem which plagued
the NMIHS in general and specifically many of the other

social variables which may have been included and which

76

would have contributed greatly to framing the context of the
women’s lives. Missing data is a particularly critical
issue in light of the utilization of multiple regression for
data analysis in this study. For example, the following
variables may have offered important insight but were not
feasible for inclusion because of missing data: number of
own children living with the mother (presently and during
the index pregnancy), employment patterns (including were
you employed 12 months before the delivery), and employment
categories (professional, skills trades, etc).
Data Analysis Plan

The choice of data analysis used in the present study
further defines its sociological orientation. Multiple
regression is commonly applied in sociological research and
is particularly appropriate for the analysis of associations
between mental health measures and social circumstances
(Mirowsky, 1999). Multiple regression serves as the
analytical technique of choice in this study primarily
because of the goal of estimating the effects of more than
two independent variables in a causal model (McClendon,
1994). The utilization of this particularly robust
technique, in combination with the strength of a nationally
representative and large sample, heightens the credibility

of these findings.

77

The statistical testing of research questions posed in
the present study utilize three levels of analysis: First,
descriptive analysis of fundamental relationships, using
variable descriptives, correlations, and cross—tabulations
of relationships between various independent variables.
Preliminary analysis of zero—order correlations suggests
high inter-variable correlation, another characteristic that
supports the appropriateness of the use of multiple
regression in this study. Hierarchical multiple regression
analysis is performed to determine the role of the chosen
predictors on maternal psychological distress. Secondly,
multiple regression using progressive adjustment to test for
possible mediators is applied. Given the methodological
goal of predicting the level of a dependent variable,
maternal psychological distress, from more than one
independent variable, the choice of multiple regression is
again validated. Lastly, multiple regression using
interaction modeling to test for moderators of the
relationships previously defined is used. Multiple
regression facilitates the assessment of which variables
predict psychological distress and how predictors relate to
each other in precipitating various effects on psychological
distress.

The initial testing of this model regresses

psychological distress in three steps, beginning with the

78

variables which define the fundamental assumptions of the
study. The first regression included only race and
pregnancy status in testing their roles as predictors of
psychological distress. The contribution of socioeconomic
factors was evaluated in the next stage of the analysis
through the addition of maternal education attainment and
household income measures. Finally, the role of marital
status was introduced into the model as the final social
predictor of psychological distress. Following the
assessment of direct effects facilitated by correlational
and descriptive analyses, a closer investigation of inter—
variable effects, using interaction modeling was applied.
The scrutiny of findings from the more simple correlation
analysis to the in-depth levels of regression analysis,
including confirmatory sub—set analyses by race and
pregnancy status, provide interesting statistical
explanations for the theoretical relationships which are
posed.
Considerations of Interaction Effects

It is theoretically plausible that any results obtained
through the above processes may be the result of interaction
effects between race and pregnancy or between other
independent variables in the present study. An interaction
is present when the effects of one independent variable on

psychological distress, in this case, change at differing

79

 

levels of a second independent variable (Kerlinger, 1973;
Keppel, 1991). In other words, interaction models specify
those variables that regulate the size of an association
(Mirowsky, 1999). Tests for interaction effects between
race and pregnancy are used to confirm regression
predictions.

The focus of the present study strives to further
expand the understanding of other theoretically sound
relationships which involve the primary independent
variables of this study, race and pregnancy status.
Interaction variables were defined based upon statistical
and conceptual rationale. Thus, those relationships
identified by strong zero—order correlations were viewed as
viable candidates for interaction inquiry. However, from a
more conceptual framework, Mirowsky (1999) explains that “a
moderator could be completely uncorrelated with the
hypothetical cause, yet determine the extent or direction of
its effect on the hypothetical consequence” (1999:113).
This is also the case among many of the relationships of
interest which involve the measure of pregnancy status.
Interaction effects were tested across the full models as
was methodologically sound and in accordance with procedures
described for mental health research (Mirowsky, 1999).
Where appropriate, interaction effects were run on race-

based sub-samples.

80

Justification for considering interaction effects
between pregnancy and race have been previously presented.
In addition, although particular age categories have been
selected from the sample (Turner, 1990; Steer, 1990),
questions about the usefulness of age in defining the
prenatal experiences of adolescents, prompts the further
determination of whether age plays any role within this
adult population; thus test for interaction effect between
pregnancy and age are warranted. Questions concerning
potential interaction effects between pregnancy and marital
status are also justified and are theoretically sound.
Given the work on variations in marital satisfaction by race
(Broman, 1993) as well as the role of marital status in the
psychological well-being of pregnant women (Hoffman, 1996;
Collins, 1993), it is reasonable to expect that maternal
psychological distress may vary among married or unmarried
women, depending upon their pregnancy status. The present
study is one of few, particularly among those utilizing
federal databases, which have purposefully included both
married and unmarried women. In addition, the attention on
the dynamics of marital status is warranted by the varied
findings on marital status by race, as well as from the
important role that spouses and significant others play in
providing social support. Each of these factors is likely

to also influence psychological distress in pregnant women.

81

Much of the social support literature suggests the receipt
of support and value of support varies by race (Lederman,
1998; Norbeck, 1989a).

The literature is replete with evidence for suspecting
interaction effects between race measures. In addition to
the previously justified basis for race and pregnancy
interactions and for race and marital status, there is also
cause to consider interaction effects between race and
education, and race and income. One of the strongest zero—
order correlations in the current model is held between race
and income. The work of Krieger et al. (1993) on race
inequity across income and the evidence for the lack of
congruity in educational worth and salary attainment across
diverse populations, warrants closer scrutiny of these
measures for possible interactions. A similar rationale for
considering interaction effects holds for all hypotheses

tested in the present study.

82

CHAPTER.4:

RESULTS

Descriptive Findings

Bivariate analyses for the present study document
statistically significant relationships between all
sociodemographic predictors, a trend not unseen in the
literature (Kitamura, 1996). Interestingly, the only
relationships which were not significantly correlated
involved those between the pregnancy variable and various
other measures in the study. Only approximately 10% of the
variance in psychological distress is explained by the
strongest of the models presented. Furthermore, there were
no significant relationships between pregnancy and
psychological distress. Consequently, interaction analyses
and subset analyses follow the regression evaluations in
order to facilitate a fuller assessment of contextual
interrelationships. Completed analyses control for age and
for race where appropriate. Similarly, group comparisons
between women who are or are not pregnant as well as within
groups differences as defined by pregnancy status are
considered.

The study sample is comprised of a fairly well-educated
group of women who have completed a mean of one year of

education beyond high school. The educational attainment of

83

the women ranged from no formal education or kindergarten
only, up to two years of graduate school. Fifteen percent
(15%) of the sample completed the equivalent of a bachelors
degree or graduate school. Twenty—two percent (22%) of the
sample had not earned a high school diploma and as might be
expected, the correlation of income and education presented
one of the stronger correlation coefficients in this study.
The mean household income earned during the 12 months before
the delivery falls within the range of $16,000 to $17,999.
The sample is comprised of 49% African American women and
49% pregnant women. The ages of the women in the study
range from 19 to 43 years, as defined by selection
criterion, with a mean age of 25.9 years. Sixty—four
percent (64%) of the sample are currently married, 28% had
never married and the remaining small percentage were either
divorced, separated or widowed. In terms of prior pregnancy
outcomes, 34% had experienced a live birth, 26% had
experienced a fetal death, and 40% had experienced an infant
death. The mean score for the dependent variable, total
CES-D score, was 14.89. The total CES-D scores among
respondents range from O to 60, with 36.1% reporting scores
suggestive of psychological distress, i.e., above the level
of 16.

The attention to the specific concerns of African

American women and pregnant women leads to the investigation

84

of demographic characteristics within these separate groups
of women through further descriptive analysis, including
cross-tabulations (Tables 3, 4). African American women in
the study are less educated and poorer than the White
participants. Black women are also less likely to be
married and are younger. Of note as well, is their lower
likelihood to be pregnant and surprisingly, given the
literature, Black women are more likely to have experienced
a live birth in their prior pregnancy than the White women
in the study. CES—D scores suggest a much greater amount of
psychological distress among the Black women in the study,
with mean scores well above the symptomatic threshold of 16.
The demographic characteristics of pregnant women are
notably consistent with those of women who are not pregnant
with the exception of race. Only 38% of the pregnant women
were African American, while 46% of the sample of women who
were not pregnant were women of color. Particular attention
must be drawn to the similar CES-D scores of 14.8 and 14.9
among pregnant and not pregnant women, respectively. Cross—
tabulations to assess the overlap in race and pregnancy
status reveal that 38% of the pregnant women are African
American, while 27% of the African American women in the

study are pregnant.

85

Table 3.

Descriptives by Race

 

 

 

variable Name Total Sample African American women-Other
women Races

Maternal Education

(years) 13.07 12.6 13.35
Household Income 11.71 9.23 13.59
Marital Status

(1=Married) .636 .361 .842
Pregnancy Status

(1=pregnant) .412 .369 .445
Maternal Age (years) 25.9 24.7 26.7
Prior Pregnancy

(1= Live Birth) .342 .416 .286
CES-D (Total Score) 14.89 17.64 12.8
Table 4.

Descriptives by Pregnancy Status

 

 

Variable Name Total Sample Pregnant Women Not Pregnant.
WOmen

Maternal Education

(years) 13.07 13.16 13.01
Household Income

(Dollars) 16,000-17,999 11.97 11.53
Marital Status

(1=Married) .636 .685 .600
Race (1=Black) .430 .384 .462
Maternal Age 25.9 25.9 25.8
Prior Pregnancy

(1=Live Birth) .342 .346 .339
CES-D Total Score 14.89 14.85 14.92

 

Bivariate correlations between variables are presented

in Table 5. All but four of the twenty—eight correlations

86

were found to be significant, with the four exceptions each
involving pregnancy status. The large number of significant
correlations among the independent variables is expected and
common within social science research as many social
variables are related in some way. The correlations between
the measure of pregnancy status and the other variables
encompass not only those four which were not significant,
but are also marked by otherwise weak correlations.
Significant correlation coefficients in this analysis ranged

from .067 (for prior pregnancy correlated to CES-D) to .503

Table 5.

Bivariate Correlations

 

CES-D Maternal Household Marital Race Age Pregnancy

 

score Education Income Status Status
CES—D score
Maternal ——
Education .196*
Household —- .392*
Income .272*
Marital -— —- .503*
Status .229* .279*
Race .192* —- —— -_

.181* .390* .495*

Age -- .335* .371* .302* 207*

.183* .
Pregnancy -.003 .041* .039 .087* 05% .006
Status .

—— * -- .007

Pregnancy —- -— —— * .136 084*
History .067* .063* .096* .135. .

 

*p <.05 (two-tailed)

87

with the strongest correlations occurring between marital
status and income, a trend well documented in the literature
(Christensen, 1992; Kitson, 1990). Correlations between
income and other independent variables presented some of the
strongest coefficients in this analysis. Again, the only
correlations which were not significant involved pregnancy
status with CES—D, income, age, and prior pregnancy history.
Bivariate correlations among the independent variables
and the dependent variable results in relationships which
are also consistent in the literature. Psychological
distress is negatively correlated with education, income,
age, marital status, and prior pregnancy outcome. Thus,
distress is likely to be greater among women with lower
education and income, and among younger, single women.
These analyses also suggest that psychological distress
increases among women who previously experienced a poor
pregnancy outcome as opposed to a live birth. It is notable
that current pregnancy status is not significantly
correlated with psychological distress. Neither does
cantly with age or

current pregnancy status correlate signifi

Prior pregnancy outcome.

Even at this elementary level of analyses, the present

study provides additional documentation of the mental health

experiences of Black women. Within this representative

sample, Black women demonstrated poorer psychological well—

88

being than their White counterparts as evidenced by average
CES—D scores above symptomatic levels (greater than 16) for
Black women in this sample. In addition, the bivariate
correlation between current pregnancy status and
psychological distress is not statistically significant.

The similarity of mean CES—D scores among women who were
pregnant and those who were not pregnant, suggests findings
that are clearly counter-intuitive and which contradict much
of the current literature.

Additional insight into the characteristics of this
sample were obtained through a second level of descriptive
analyses using cross—tabulations and comparisons of means.
Cross tabulations of pregnancy status with race have been
presented previously and it is noteworthy that the age
distributions across race are fairly balanced at the younger
age categories (19 to 29 years of age). However, the
percentage of African American women decreases with each
increase in age category, including a mere 3 cases
represented among the oldest age category. Manipulations of
pregnancy status and marital status reveal that married
women comprise the majority of the sample, regardless of
pregnancy status. In the present study, among women who are
not married, 40% are pregnant and 32% are not. The

inclusion of unmarried and never married women in a study

89

with married women is a unique contribution within federal
studies and ultimately the broader research literature.

Comparisons of means provide contextual insights into
the lives of the women in the present study. Comparisons of
CES—D means by educational attainment and household income
were completed. Although descriptive data on the total
sample suggest low levels of psychological distress in this
sample, closer review of CES—D total scores by race,
pregnancy status and educational attainment identify only
one category of women with CES—D scores below the distress
threshold. Educated White women (attained high school
degrees or beyond) have CES—D scores less than 16 while
under-educated women (less than high school degree) tended
to demonstrate distress, regardless of race. Furthermore,
educated African American women (high school degrees or
beyond) had CESeD scores which were comparable to White
women with no high school degree. Greater still, were the
CES-D scores of educated Black women who were also pregnant
(Table 6).

Income attainment for this sample reiterate trends
previously noted in the literature (Krieger, 1993). African
American women tended to be more greatly represented among
the lowest household incomes, regardless of educational
attainment and least represented among the highest household

incomes, in spite of educational attainment. Among White

90

Table 6.

Descriptive analysis of CES-D Scores within Pregnancy Status
and High School Attainment by,Race ICES-D scores; (casesll

 

 

>=High School <High School
WHITE BLACK WHITE BLACK
PREGNANT 12.04 17.52 16.48 20.55
(1079) (589) (194) (199)
NOT PREGNANT 11.93 16.38 16.78 19.44
(1283) (981) (310) (400)

 

cases, 6% of the cases had household incomes from $20,000 to
above $60,000 as opposed to only 1% among African American
cases. Racial differences among the lower household incomes
were slightly more modest, with those below $10,000 seen in
only 2% of the White cases and 1% of the Black cases.
Educated African American women comprise 31% of the African
American sample, while educated White women comprise 47% of
the White sample. Given the representativeness of African
American women among the educated sub-group, one might
expect a greater wealth distribution among this group of
women.

Descriptive relationships of income to psychological
distress also suggest race specific trends. Black women had
CES-D scores indicative of distress at much higher levels of
income than the White women in the present study. For
example, CES-D scores at or above the cut point of 16 were

91

found in all income categories up to $25,000 for Black
women, while White women demonstrated symptomatic levels of
CES-D only below income levels of $10,000.

Additional CES—D mean comparisons were completed for
the measures of marital status and age. The highest CES—D
scores were found among single women, regardless of race.
While not as high as among single women, African American
women who were married also demonstrate symptomatic CES-D
levels, regardless of pregnancy status (Table 7).
Symptomatic levels of CES—D among White women were found in
only the youngest age groups (19 and 20 years old). Black
women, on the other hand, showed no such prescribed trend,
with symptomatic levels found across all age spans,
especially within categories from 19-28 years.

Factors Predicting Psychological Distress in Women
.Measures of psychological distress in women who are pregnant
will significantly differ from psychological distress in
women who are not pregnant (Hypothesis 1).

Progressive Adjustment through Hierarchical Multiple
Regression

Psychological distress was regressed on all
sociodemographic variables in three hierarchical steps as
described in Tables 8, 9, and 10. At this level of
analysis, pregnancy status is not a significant predictor of

psychological distress, suggesting that pregnancy status in

92

Table 7.

Descriptive Analysis of CES-D Scores within Marital and
Pregnancy,Status by Race ICES-D score (casesll

 

 

Married Not Married
WHITE BLACK WHITE BLACK
PREGNANT 11.83 15.44 18.23 20.22
(1097) (319) (176) (469)
NOT PREGNANT 12.05 14.91 16.94 18.44
(1324) (458) (269) (923)

 

Table 8.

Model 1.1-Multiple regression of Psychological Distress on
Race and Pregnancy Status

 

Race 4.86*
Pregnancy Status .328
R2 .037
R .192
N 5034

 

*p<.OOl, unstandardized coefficients presented

Table 9.

Model 1.2-Multiple regression of Psychological Distress on
Race. Pregnancy Status. and Socioeconomic Status (SES)
measures (maternal education and household incomel

 

Race 2.536*
Pregnancy Status -474
Education —1.168*
Household Income --421*
R2 .092
R .303
N 5034

 

76<.001, unstandardized coefficients presented

93

Table 10.

Model 1.3-Multiple regression of Psyphological Distress on
Race. Pregnancy StatusL.Socioeconomic Status (SES) and
Marital Status

 

Race 1.75*
Pregnancy Status .515
Education —.917*
Household Income —.332*
Marital Status -1.932*
Age —.161*
R2 .099
R .315
N 5034*

 

p<.001, unstandardized coefficients presented

this sample of women, had no direct effect on their mental
health status. The measures of psychological distress in
women who are pregnant do not significantly differ from
psychological distress in women who are not pregnant, thus
the first hypothesis of this study is not supported. This
finding is further validated by ANOVA results and the F-test
which suggest no notable difference in the CES-D means
between women who are or who are not pregnant. While
psychological distress does not vary by pregnancy status,
these analyses demonstrate that psychological distress is
predicted by race, socioeconomic status, and marital status
in this sample.

The results for Model 1.1 (the first step of the
progressive multiple regression) are presented in Table 8
and demonstrate that race alone accounts for only 4% of the

94

variance in CES-D scores in this model. The first step in
testing hypothesis one of the study confirms previously
established relationships between race and depression
(Brown, 2000; Brown, 1995). African American women in the
present study tend to have greater psychological distress
and depressive symptoms than those of other races. In other
words, the mental health of Black women is worse than that
of women of other races. The t-value in this model
(t=13.902) reiterates that the race variable serves as a
good predictor of psychological distress. Although the
empirical indicators of the analysis evidence the strength
of this model, a large percentage of the variance in
psychological distress remains unexplained.

As demonstrated in Model 1.2 (Table 9),education level
and household incomes are both inversely related to
psychological distress; thus women with lower household
incomes and less education are more likely to experience
greater psychological distress and diminished psychological
well-being. Race remains a significant predictor of
psychological distress in this model, with Black women more
likely to experience poor psychological well—being than
White women in this study. The inclusion of the
socioeconomic variables, household income and maternal
education, in the model, results in a decline in the

regression coefficient for race and contributes to the

95

explanation of an additional 6% of the variance in CES—D
score. The lack of significance of the regression
coefficient for pregnancy is retained.

The relationship between race and psychological
distress does not disappear when adjusting for socioeconomic
status, however, the regression coefficient for race is
decreased, suggesting a mediating effect among these
measures (Model 1.2). Adjusting for education and income,
reduces the association between race and distress by 49%
(100 x [4.86-2.53]/4.86 = 47.94). This multiple regression,
which serves as the sociological equivalent of experimental
matching (Mirowsky, 1999), demonstrates that had education
and income been matched, the differences in psychological
distress between African American women and others would
remain. In other words, if Black women attained the same
educational and income levels as women of other races, the
differences in distress between these two groups would be
47% smaller. While this point is notable and impressive,
one must not overlook the additional 70% of the difference
between these groups which is not explained by education and
income.

The final step of this progressive regression model
involves the adjustment for marital status. The
relationship between race and psychological distress is

maintained. Marital status is the strongest predictor of

96

psychological distress in this model; stronger than both
education and household income. The inclusion of marital
status does not appreciably change the coefficients for the
other independent variables and they remain significant
predictors of psychological distress. The addition of
marital status in this step (Table 10), results in no
appreciable change. Only an additional 1% of variance in
the model is explained by the inclusion of marital status.
Furthermore, at the final step of this hierarchical model,
confidence in the predictive values of CES-D has increased
from .19 to .31 as indicated by the coefficient of multiple
correlation (R). The strength of the model appears solid,
with all t values greater than +2 or less than —2, with the
exception of the values for current pregnancy status.

Finally, analysis of variance (ANOVA) suggests that the
null hypothesis that there is no difference in means for the
dependent variable is supported. Thus, ANOVA results
suggest there is no statistical difference in the mean CES~D
scores between women who are and who are not pregnant,
thereby further reiterating the findings as discussed. In
summary, the presence of race in this model accounts for
the majority of the variance in CES—D. A comparison of each
step of the regression is summarized in Table 11. Although
the pregnancy measure fails to gain statistical

significance, the marked shift in coefficients and the

97

Table 11.

PROGRESSIVE ADJUSTMENT: Demonstration of measure of
Psychological Distress among women categorized by race and
pregnancy statusl adjusting for socioeconomic status.
marital status and age

 

 

REGRESSOR MODEL 1.1 MODEL 1.2 MODEL 1.3
Pregnancy Status .328 .433 .531
(.931) (1.264) (1.554)
Race 4.860* 2.474* 1.704*
(13.902) (6.699) (4.286)
Education -.697* -.643*
(-6.736) (-6.200)
Household Income —.439* —.375*
(—12.356) (-9.961)
Marital Status —2.254*
(-5.150)
Age —.l61*
(-4.301)
Constant 12.665 27.920 28.181
(45.720) (21.297) (21.535)
R .192 .302 .310
R2 .037 .091 .096

 

*p<.001; unstandardized coefficients reported with t values
in parentheses

generally small percent of variance accounted for by the
model suggest further analysis is warranted, including
assessment of interaction effects.

Despite suggestions in the literature (Hobfoll, 1995;
Gotlib, 1989; Hedegaard, 1993), the relationship between
pregnancy and psychological distress is not statistically
significant in the present study. As such, current
pregnancy status is not predictive of psychological

distress. The absence of significance in the relationship

98

between current pregnancy status and psychological distress
is unchanged at all levels of the model, thus ruling out the
presence of suppression effects as an explanation for the
lack of relationship between the two fundamental variables
in this study.
Interaction Effect and Subset Analysis by Race

It is the counter-intuitiveness of these initial
findings that prompted further testing for interaction
effect between race and pregnancy status as well as a subset
analysis by race. Regression predictions are confirmed by
testing for interaction effects between race and pregnancy
and the results are found in Table 12. Secondary analysis
in this investigation reveals that inclusion of the Non-
additive term for race and pregnancy does not improve the
fit of the full model and furthermore, is not statistically
significant. This finding holds for inclusion of the
interaction variable for the total model as well as for
analysis of the isolated regression relationships between
race and pregnancy alone. Thus, the relationships between
race and psychological distress as well as pregnancy and
psychological distress represent main effects and are not
the result of any intervariable dependencies. Additional
evidence for race—based effects are obtained from the

interaction analyses.

99

Table 12.

Non—additive Effects of Race and Pregnancy,on Psypholggical
w

 

Variable Name Regression Model of CES—D on
Race, Pregnancy &
Interaction Term

 

 

Race 4.389*
Pregnancy Status -.l64
(Race)(Pregnancy) 1.185
R2 .038
R .194
*p < .001

The primary interest of race in the present study,
prompted further testing of this question through sub-set
analysis. Among the White sub-sample, the pattern of
significance and predictors remains consistent in the
analysis of the sub—sample of White women in the study.
Interestingly, although erroneous, regression findings
within the sub—sample of African American women shift. Most
notably, current pregnancy status becomes predictive of
psychological distress (p < .05). However, upon testing for
interaction effects between pregnancy and race (Table 12),
the absence of significant interaction effects between these
variables proves that this sub-sample finding for Black
women is spurious; and pregnancy is not truly predictive of

PSychological distress.

100

Additional Findings from Interaction Effect Analysis

Tests for interaction effects among other independent
variables in association with the two fundamental measures
of pregnancy and race are now described, beginning with a
discussion of interaction effects involving pregnancy. In
addition to the interaction effect between pregnancy and
race, the following effects are also conceptually warranted:
1) interaction effect between pregnancy and marital status;
and 2) interaction effect between pregnancy and age.
Pregnancy interaction analyses were run on the full model
and confirmed by racially defined subsets analysis where
indicated.

Analysis of interaction between pregnancy and marital
status confirms the existence of interaction effects between
these two measures. Inclusion of the Non-additive term for
pregnancy status and marital status in the model resulted in
a significant interaction effect (p < .05) (Table 13). In
addition, the inclusion of the pregnancy/marital status
interaction term in the model results in the reversal of the
previously insignificant relationship between current
pregnancy status and psychological distress. Within the
model which tests for pregnancy/marital status interaction,
current pregnancy status becomes a significant predictor of
psychological distress (p < .05). In order to identify the

direction of the interaction effect, analyses of the model

101

Table 13.

Non-additive Effects of Pregnancy and Marital Status on
Psychological Distress

 

Variable Name Model with (Race) Model with
Pregnancy
(Pregnancy) Marital Status

Interaction Effect Interaction Effect

 

Race l.25** 1.761*
Pregnancy Status -.004 l.567**
Education —.914* -.917*
Income -.332* -.330*
Marital Status —1.954* -1.319**
Age —.161"r —.163*
(Pregn)(Marital Status) --—— -1.610*
(Race)(Pregn) 1.232 —-——
R2 .038 .100
R .194 .317

 

**p< .05; *p < .001

were repeated selecting cases by marital status (married or
unmarried/single). Among married cases, current pregnancy
status was not predictive of psychological distress.
However, among single cases, current pregnant status was
significant predictor of psychological distress (p < .05).
Thus, for unmarried women who are pregnant, their pregnancy

is predictive of psychological distress.

102

Further expansion of this questions was completed by
repeating the analysis controlling for race. The pattern of
significant prediction of psychological distress held with
one exception (Table 14).

Applying the previous findings that race and pregnancy
status do not interact, and that marital status and
pregnancy do interact, one intuitively expects the pregnancy
of all unmarried women to predict distress and the pregnancy
of all married women not to predict distress. The one
exception to this paradigm is likely a statistical residual.
Among single women of other races, pregnancy status was not
predictive of psychological distress. However, within this
subsample, which had the smallest sample size in this subset
analysis, the regression did not have a significant F
statistic. Analytical problems with the model in this
subset are also evidenced by the t statistics (ranging from
-1.184 to —.875) and the small size of the F statistic
(F=1.752). Thus, findings in the case of unmarried women of
other races, is more likely a result of small sample size in
comparison to the other sub-samples (n=440).

Finally, the pregnancy/marital status interaction
effects in this model are only significant when applied in
the full model using the total sample. Subset analyses by
race which included the pregnancy/marital status interaction

status fail to produce any significant evidence of

103

Table 14.

Regression Coefficients for Current Pregnancy Status
regressed on CES—D by,Race and Marital Status

 

 

Black Other Races Total
Sample
Unmarried —1.748* (1391) -1.096 (440) —1.567*
(1836)
Married -.319 (776) .152 (2420) .0028
(3197)

 

*p<.05; (n—sample size)

interaction effects, and are likely further evidence of the
spurious effects revealed in the previous analysis.

Analysis of pregnancy and age interactions are valid
questions, however findings were insignificant for the total
model. Thus, pregnancy and age have additive contributions
to the relationships found and represent main effect
relationships in this model. In addition, the inclusion of
the interaction term for pregnancy and age results in no
change in the relationship between current pregnancy status
and psychological distress, unlike the effect described for
pregnancy and marital status interaction.

The second foundational area of concern is that of
race, thus in addition to testing interaction effects
related to pregnancy, the evaluation of interaction effects
involving race are assessed. Notably, none of the race
interaction effects appear to be statistically valid, i.e.,

104

all Non-additive terms involving race lack significance;
including race/education, race/income, and race/marital
status. Thus, the race based relationships found in prior
analysis represent direct effects and the relationships
between race are not mediated by other variables in the
model.

Psychological Distress in Pregpant Women

Maternal psychological distress will vary by race,
socioeconomic, and marital status (Hypothesis 2).

Maternal psychological distress for the purposes of
this research question refers specifically to the
psychological state of pregnant women, thus analyses are run
only on the subset of women who were pregnant at the time of
the study. Closer investigation of psychological distress
among this sub—sample of women in the present study suggests
that race, socioeconomic status and marital status are all
predictive of mental health in pregnant women, providing
evidence in support of the second hypothesis of this study.
Regression results are presented in Tables 15, 16 and 17,
and provide evidence in support of the second hypothesis of
this study. Race is a significant predictor of
psychological distress at all levels of the regression, with
African American women at greater risk of psychological
distress than women of other races. In addition, the social

correlates education, income and marital status, and the

105

Table 15.

Model 2.1- Multiple regression of Psychological Distress on
Race among Pregnant Women

 

Race 5,57*
R2 .049
R .222
N 2060

 

*p<.001, unstandardized coefficients presented

Table 16.

Model 2.2-Multiple regression of Psychological Distress on
Race and Socioeconomic Status (SES) measures lmaternal

education and household income) among Pregnant Women

 

Race 3.062*
Education -1.356*
Household Income —.438*
R2 .116
R .341
N 2060

 

*p<.001, unstandardized coefficients presented

Table 17.

Model 2.3-Multiple regression of Psychological Distress on
Race. Socioeconomic Status (SES). Marital Status and Age

among Pregnant Women

Race 2.12*
Education -1.062*
Household Income -.325*
Marital Status —2.648*
Age —.168**
R2 .126
R .356
n 2060

 

*P<.001, **p<.05; unstandardized coefficients presented

106

dependent variable, psychological distress are all inversely
related and statistically significant. Thus, poorer, less
educated women who are unmarried are more likely to
experience greater psychological distress.

Comparative results from each progressive step of the
regression analysis are presented in Table 18. The
isolation of pregnant women for these analyses does not
appreciably vary the findings from those obtained in the
total population. The significance and direction of social
factors in relationship to psychological distress are the
same in Model 1.3 for the total sample (Table 11) and in
Model 2.3 for pregnant women (Table 18). Thus, social
correlates of psychological distress in women who are
pregnant result in no differing predispositions, than women
in the general population.

In an effort to again delineate the experiences of

women of color distinctively in this study, additional sub—

sample assessment of the experiences of Black women who were

pregnant were completed. As with race based analyses for

the first hypothesis of this study, the spurious nature of

subset findings was confirmed through the testing of

interaction terms. Initial differences in marital status

and age predictiveness for Black women who are pregnant in

relation to pregnant women of other races was not validated.

107

Table 18.

PROGRESSIVE ADJUSTMENT: Psychological Distress among
Pregnant women categprized by race and agjusting for age.
socioeconomic status and marital status

 

 

REGRESSOR MODEL 2.1 MODEL 2.2 MODEL 2.3
Race 5.574* 3.062* 2.122*
(10.330) (5.432) (3.532)
Education -1.356* —l.O62*
(—5.489) (—4.190)
Household Income -.438* -.325*
(-8.l92) (-5.608)
Marital Status —2.648*
(-3.880)

Age -.168**
(—2.859)
Constant 12.719* 23.572* 27.731*
(38.118) (24.563) (18.202)
R .222 .341 .356
R2 .049 .116 .126

 

*p<.001, **p<.05; unstandardized coefficients reported with

t values in parentheses

Finally, despite the absence of evidence for

interaction effects in the first hypothesis of the study,

this question is further extended by investigating

interaction effects among this sub~sample of pregnant women.

Interaction effects were completed for the following

interactions: race and education,

and marital status.

108

race and income,

and race

Consistent with the previously

described analyses, none of the interactions with race are
statistically significant. Thus among, pregnant women, the
effects of race on psychological distress are unmediated by
other variables considered in the model (as opposed to the
findings from hypothesis one, that the effects of pregnancy
status on psychological distress are mediated by marital
status).

Contributions of Pregnancy Outcomes to Maternal Distress
Current psychological distress will be significantly related
to previous pregnancy outcomes. (Hypothesis 3)

In order to test this hypothesis, the independent
variable which measured past pregnancy outcome, a
categorical variable with responses of 1) live birth, 2)
fetal death, and 3) infant death was recoded. Dummy
regression was used in this case. Each category was recoded
to account for the stated response (coded as 1) and all
other categories (coded as 0) resulting the creation of
three dummy variables. Dummy regression analysis requires
the inclusion of one fewer dummy variable in the equation
(g-l, where g = the number of categories in the nominal
variable which is dummy coded). In this case two dummy
variables were included in the regression equation.
Regression coefficients cannot be computed for an equation
that includes all the dummy variables because there will

then be perfect multicollinearity among the dummy variables

109

(LB dum + FD dum + ID dum = 1). The existence of perfect
multicollinearity precludes the estimation of the separate
effects of each independent (dummy) variable on the
dependent variable, psychological distress.

The omitted variable serves as the reference category
in dummy regression and is represented as the constant in
the analysis. The decision of which variable to exclude is
based upon theoretical support. In the case of prior
pregnancy outcome, the Live Birth group serves as the
reference group because that is the normative category and
is the standard to which public health and social well-being
perspectives aspire. Thus, Infant Deaths and Fetal Deaths
are looked at in comparison to Live Births which are the
preferred birth outcome and goal.

Results of the dummy regression which was completed in
two progressions, are presented in Table 19, with the Live
Birth dummy variable as the constant (excluded) variable.
Table 20 presents results of a second regression using only
the live birth variable and presented as confirmation of the
initial dummy analysis presented in Table 19. The
confirmatory analyses in Table 20 result in virtually
identical findings as obtained in the initial dummy
regression. The dummy regression results (Table 19) and
confirming regression using the live birth dummy variable

(Table 20) suggest that at least one of the categories of

110

Table 19.

Dummy regression with Progressive Adjustment of
Psychological Distress on Race. Pregnancnytatus and

WY

 

Pregnancy History:

Fetal Death Dummy 1.403** 2.52*
Pregnancy History:

Infant Death Dummy 1.990* 2.47*
Race -—- 5.18*
Current Pregnancy Status --- .361
R2 .005 .046
R .070 .214
N 5034 5034

 

*p<.001, p<.05; unstandardized coefficients presented; Live
Birth is excluded Dummy variable and serves as constant

Table 20.

Regression of Psychological Distress on Race. Pregnancy
Status and Pregnancy History

 

Race (Black = 1) 5.18*
Pregnancy History (Live Birth =1) —2.49*
Current Pregnancy Status (Pregnant=1) .363
R2 .046
R .214
N 5034

 

*p<.001; unstandardized coefficients presented

111

prior pregnancy outcomes has a different mean CES-D than the
others. The multiple correlation of determination (R5,
however indicates that only .5% of the variance in CES-D can
be accounted for by knowing the prior pregnancy outcome
alone. The t values validate the strength of the model
which confirms that the mothers in the infant death group
have significantly worse psychological distress. There is
no significant difference in the CES—D scores between the
fetal death group and the live birth group.

When the dummy regression is progressively adjusted to
include race and current pregnancy status, the fit of the
model improves, accounting for 4.6% of the variance in CES-D
scores (Table 19). The inclusion of the additional
variables facilitates the evaluation of the differences
between the infant death and the fetal death groups in
comparison to the live birth group, while holding race and
current pregnancy status constant. Thus, the results
suggest that when race is held constant, infant death and
fetal death remain significant predictors of psychological
distress. As with prior analysis, current pregnancy status
is not a significant predictor of psychological distress and
has no meaningful effect on the differences found between
groups based upon prior pregnancy outcome. In summary, the

psychological distress in these women is predicted by the

112

pregnancy outcome of their previous pregnancy and hypothesis
three is supported.

In order to facilitate testing of interaction effects
within this research question, an additional variable was
created. Because interaction analysis must include both
variables as well as their Non—additive interaction term in
the regression, a prior pregnancy outcome measure was
created. The prior pregnancy outcome measure was recoded
using the pregnancy outcome data (live birth, fetal death,
and infant death) which was dummy coded such that 1: live
birth and 0=all other pregnancy outcomes (fetal death or
infant death). Findings from secondary testing of
hypothesis 3 using this variable are consistent with those
described previously (Table 21). Non—additive interaction
terms were created for race and prior pregnancy outcome,
race and current pregnancy, and current pregnancy and prior
pregnancy outcome. Analysis was completed on the total
sample, as well as subset analysis by race and pregnancy
status. No evidence of interaction effect was identified in
any case.

In addition, inclusion of the prior pregnancy outcome
variable in the foundational model (similar to that tested
in regression analysis for hypothesis 1) reiterates the
previous findings, including the lack of significance of

current pregnancy status, and confirming the significance of

113

 

Table 21.

Confirmatogy Analysis for Role of Prior Pregpancy70utcome.
including Interaction Effects

 

 

Variable Name Full Model Full Model Full Model

+ (PriorPrg) + (PrioPrg)

(CurrPreg)
Race 2.007* 1.560* 2.007**
Education -.984* -.980** -.985**
Income -.333* -.332** -.332**
Marital Status -2.140* -2.169** -2.l47**
Age —.171* --.171*"r —.171**

Prior Pregnancy
Outcome

(Live Birth=1) —3.133* —3.121** —3.069**
Current Pregnancy

Status .575 .116 .628
(Race) (Pregnancy)

Status Interaction -—— 1.107 -—-—
(Current Pregnancy)x

(Prior Pregnancy) -—— --- —.156

R2 . 1 13 . 1 14 .113

R .336 .337 .336

 

*p < .001; **p< .05; (spurious findings—not valid)

prior pregnancy outcome as a viable predictor in
psychological distress in the women in the present study.
Prior pregnancy outcome is negatively associated with

psychological distress, however one must pay particular

114

attention to the coding of the variable in the
interpretation of these results. In this case, a greater
value of the dummy coded prior pregnancy outcomes (live
birth) correlates to lesser values (less distress) in the
psychological distress measure. Thus, as expected, women
who experienced a live birth in their previous pregnancy are
more likely to have less psychological distress in their
current pregnancy.

Finally, within this framework, the non—additive
interaction term for prior pregnancy outcome and current
pregnancy outcome is then included, an addition which is
justified by the overarching interest in the role of the
pregnancy status in general. Results suggest that any
relationships demonstrated between psychological distress
and current pregnancy status or psychological distress and
prior pregnancy outcomes are distinct, with no evidence of
interaction effects. Finally, spurious findings were again
obtained when this research question is tested by race. As
with hypothesis one, significant prediction of distress by
current pregnancy status for Black women is not confirmed by
interaction effects of race and pregnancy status. The
failure to obtain significant evidence of interaction
effects by race and pregnancy status refute any support for

race specific findings.

115

Supplemental Considerations: Questions of Employment and

Parental Status

The literature addresses two additional factors of note
in maternal psychological well-being; the contributions to
psychological status made by current parenthood and
employment. Parenthood has been investigated from a number
of perspectives, both quantitatively (Lederman, 1998) and
qualitatively (Tietjen, 1985). Most commonly, sociological
work views parenthood subjectively, as representative of one
of the many stratified roles in which women are planted by
society (Oakley, 1980) and from which a woman’s standpoint
evolves (Collins, 1991). Sociology has also contributed to
the more clinical work which addresses issues of parenting
and numbers of children Lederman, 1998). The dynamics of
status as associated with employment and work are also
viewed as potential stressors among pregnant women
(Bramwell, 1997). Questions regarding type of work,
frequency of work and duration of employment are raised in
the literature (Taylor, 1997; Dragonas, 1997; Bramwell,
1997).

The NMIHS contained two variables which were
appropriate measures of these issues. The following survey
questions were used: 1) Did you work during the 12 months
before delivery of the baby named on this questionnaire?

(response categories- yes/no); and 2) List the number of

116

your own children living with you now (continuous, numerical
responses allowed). Unfortunately, the shortcomings of
significant amounts of missing data in responses to measures
for these issues, results in the inability to include these
measures in the present study. In addition to excessive
amounts of missing data for responses to each of the
independent measures themselves, cross tabulations of the
measures for children and employment with the CES—D measure,
reveal that there is also missing data for the dependent
variable among those responding to the parenting and
employment measures.

Identification of applicable cases for this
exploratory analysis resulted in 1722 cases who had
responded to the key questions of pregnancy status and CES—D
measurement, as well as having responded sufficiently to the
independent variables of interest for this exploratory
analysis. Findings obtained from the exploratory analysis
of the roles of parenting and employment contribute very
little to the understanding of mental health experiences of
women and pregnant women in particular. Bivariate analyses
reveal the measures for employment were not significantly
correlated to any other variable in the study. In the case
of parental status, zero- order correlations were
significant for all other independent variables, with the

exception of the newly added employment variable. In

117

addition, parental status was not significantly correlated
with the dependent variable, psychological distress.

Demographic distributions remain surprisingly
consistent with those of the original sample. CES-D total
score (mean = 14.73), education, and age (mean = 25.9) are
comparable to the original sample. This subset which
includes employment and number of children, is comprised of
a slightly poorer group with a higher percentage of women
who are not pregnant or not married. Closer scrutiny of the
key variables, race and pregnancy status reveal that among
cases for this exploratory analysis, 44% (n=212) are
pregnant women of color and 47% (n=470) being pregnant and
White. However, despite the characteristic similarities,
analysis of variance results suggest that there were no
significant differences in psychological well—being among
this subset. Furthermore, psychological distress was quite
low in the sub—sample, with the vast majority of respondent
categories at the non—symptomatic level. The only exception
to this trend in CES—D scores was found among those
respondents who had 4 children (n=110). This group
presented a CES-D measure which is indicative of
psychological distress, however, the score lie precisely at
the threshold of symptomatology (16.55).

Nevertheless, regression analyses in this group are

presented in Table 22. Repeat analyses for the fundamental

118

Table 22.

ExploratogyiSub-sample Analysis

 

 

Original Model + Work + Parental Parental
Status Status &
Work
Race 1.45* .445* 1.538* 1.533*
Pregnancy .375 .383 .277 .285
Education .959* .959* -1.029* —1.030*
Income .321** .320** -.327** —.326**
Marital Status .957 .963 -.965 -.971
Age .167* .169* -.149* -.151*
Work .430 -.434
Parental Status --230 --231
R .285 .286 .286 .287
R2 .081 .082 .082 .082
N 1379 1379 1379 1379

*p < .05; **

model presented in hypothesis one of the present study

result in relatively consistent findings, with the exception

of a lack of predictive value for marital status within this

sub—set. Inclusion of the two variables which characterize

this exploration was completed through the independent

analyses of each variable and through the inclusion of both

variables simultaneously.

In neither case,

were employment

or parental status predictive of psychological distress in

women. The loss of cases which would result from further

subset analysis by race or pregnancy status precluded

additional consideration at this point.

120

CHAPTER 5:

DISCUSSION

The present study is designed to investigate the mental
health experiences of pregnant women, with specific focus on
the differences defined by the race of the women
participants. There are four important findings from this
investigation. The first is the descriptive finding that
Black women, regardless of pregnancy status, in this
nationally representative sample experience significantly
greater psychological distress than women of other races
(p<.001). African American women report a mean CES—D score
of 17.64, with 45.7% reporting scores above the symptomatic
cut point of 16. In contrast, women of other races report a
mean CES—D score of 12.81, with only 28.8% scoring above 16.
Secondly, the psychological state of women is predicted more
by the social contexts of their lives than by the immediate
experience of pregnancy. Thirdly, the present study
confirms the importance of the role of marriage in the
psychological well—being of women who are pregnant. Lastly,
it is the experience of prior pregnancies, rather than the
current pregnancy, which effectively predicts psychological
distress in women.

African American women in the present study are at

greater risk for psychological distress and depressive

121

symptoms. Filling a distinct void in the current
literature, these findings empirically demonstrate that
mental health among Black women is indeed worse than among
women of other races. Psychologically distressed women in
this study tend to be poorer and less educated,
characteristics predominated by women of color in this study
and nationally. Furthermore, when considering the inverse
of those characteristics, it is notable that factors which
tend to enhance the psychological well—being of women of
other races, fail to demonstrate comparable advantages for
African American women. For example, the psychological
distress of highly educated Black women is more comparable
to that of uneducated and undereducated White women as
opposed to their White peers who are educated. In addition,
Black women remain distressed at much higher income levels
than women of other races in the present study.

The second finding of the present study is perhaps the
most unexpected. The absence of any effect of current
pregnancy status on psychological distress is a result which
was unanticipated yet is provocative. The causal effect
initially hypothesized for pregnancy and distress was not
confirmed, instead, the specific roles of marital status,
education, income, race and age are clarified. Quite
simply, psychological distress does not vary appreciably

between women who are pregnant and those who are not,

122

regardless of the race of the pregnant women. Pregnant

women who are poorer, less educated and unmarried are more
likely to experience greater psychological distress. The
pregnancy experiences of African American and White women,
however, are for the most part, undifferentiated in the
present study. Among pregnant women in this study, race,
socioeconomic status and marital status remain predictive of
mental health.

The only evidence of variation in the pregnancy
experiences of women in the present study comes not from
racially defined categories but from the evidence of
interaction between the pregnancy and marital status
measures. Marital status is inversely related to and
predictive of psychological distress. Upon inclusion of the
interaction term for pregnancy and marital status in the
model, the third distinct finding of the present study
becomes apparent. The interaction term is statistically
significant and discloses the existence of a suppression
effect in pregnancy status. Upon testing for interaction
effects in this model, current pregnancy status becomes
predictive of psychological distress. Thus, marriage
clearly plays an important role in the mental health
experiences of the women in this study. Psychological
distress does vary in the present study at different levels

of marital status. Women who are not married are at greater

123

risk for psychological distress and among single women,
being pregnant is predictive of distress (p<.05).
Interaction effects are only significant when applied in the
full model using the total sample. These findings are all
the more intriguing given the absence of interaction between
race and pregnancy or any of the independent measures
tested.

The fourth finding of the present study provides for
future consideration in regards to pregnancy outcomes. One
might expect the proximity of a current pregnancy to
supplant the effect of prior pregnancy outcomes in the
mental health of women. This study confirms the opposite;
rather it suggests that when race is held constant, prior
infant death and fetal death become significant predictors
of psychological distress and current. The experience of
having a live birth in the previous pregnancy is predictive
of lower CES-D scores and better psychological well-being.
The effects of prior pregnancy outcomes are unmediated by
the current pregnancy as confirmed by interaction testing.
Subset analyses by race are also reiterative of previously
described results and reveal no interaction or
differentiation by race.

The empirical findings of the present study lead to the
following revision in the conceptual model presented

initially:

124

Eigure_2

Revised model of Mental Health, Pregnancy, and Race

f,pi RACE

 

KP CURRENT
PREGNANCY MARITAL PSYCHOLOGICAL

————+>
EXPERIENCE STATUS DISTRESS

PRIOR
PREGNANCY
EXPERIENCE

The revised model confirms and highlights the significant
role of three predictors of psychological distress in women in
this study, 1) race, 2) current pregnancy status, and 3) prior
pregnancy outcome. The expectation of a greater role for
current pregnancy as opposed to prior pregnancy was the
rationale for initially excluding this term from the proposed
model. Findings not only heighten the importance of the prior
pregnancy but negate the original supposition about the
relationship of current pregnancy on psychological distress.

As a consequence, both measures are incorporated into the
final model. The relationships between current pregnancy and
race are exogenous, but not causal in this model. The measure

of current pregnancy and prior pregnancy outcomes are not

125

significantly correlated in these data and thus appear
unrelated. The contributions of marital status are much more
meaningful than originally hypothesized, with marriage serving
as a mediator of distress among pregnant women.

Race as initially hypothesized remains an important
predictor of mental health in African American women. As
noted, the present study is driven by a concern for the
disparate experiences and health outcomes of African American
women who are pregnant. The race based findings as obtained
are quite unexpected. The evidence for distress which is
categorized by race is irrefutable; Black women present
greater rates of distress in both the descriptive data and as
obtained through all regression analysis completed in the
study. Despite this clear pattern, race contributed nothing
to the contextual analyses and interpretations of this
investigation for the various subsets considered. All
research questions were evaluated across racial categories
through subset analyses and interaction testing. In no case
did race significantly interact with any of the other
independent variables in the study.

The descriptive trends described previously in regards to
race and the independent variables of the study reiterate the
findings of Krieger et al. (1993) which document the varying
utility of education and income for women of different races.

Why are educated Black women as depressed, if not more

126

depressed, than White women with much less education? The
findings also speak to the implications of interpreting the
value or meaning of wealth and poverty in the lives of African
American women. Income and educational attainment appear to
have differing advantages by race. It is perhaps the “socio—”
component of socioeconomics which portends the greatest
influence in the lives of women of color, perhaps even
overriding the quantitative contributions implied from the
“economic” component of the term. Similarly issues of context
effect the quality of life. For example, are the benefits of
marriage the same for Black and White women? The absence of
marriage does not negate the role of significant others in
ones life, an observation which may have greater effect in
Black women.

Interdisciplinary and cross-disciplinary models have all
expected pregnancy to hold some level of emotional influence
(Lips, 1985). Whether the hormonal/biomedical paradigms of
medicine, the intrapsychic models of psychology or the family
transitions exemplars of sociology, most have expected
pregnancy to play some role in the psychological state of
women. In addition, some researchers have found an
association between parity and depression during pregnancy
(O'Hara, 1986). Thus, one would expect this present sample of
women who have experienced at least one pregnancy prior to the

current conception, to consequently have higher rates of

127

depression and psychological distress. Therefore, the
fundamental findings of no relationship between current
pregnancy status and psychological distress is at once
conceptually counterintuitive and empirically unexpected. The
present study however, confirms the suspicion that for the
majority of women, pregnancy is not problematic (Thorpe,
1992). These intriguing findings bring important
clarification to this field of study.

The lack of support for a relationship between pregnancy
status and maternal psychological distress seems to indicate
that within the range of normally experienced psychological
variations in pregnancy, there are no important detrimental
effects that can be tied exclusively to the pregnancy itself.
This finding parallels those found in Istvan’s (1986) work on
psychosocial factors in childbirth. In the lives of most
women, the occurrence of pregnancy in and of itself, poses no
harmful contribution to mental health. It is perhaps
society’s reification of pregnancy which has supported the
misinterpretation of this life occurrence. Lips cautions
researchers not to attribute characteristics of emotional
symptoms in the pregnancy to the pregnancy itself rather than
to ‘environmental or dispositional' factors (Lips, 1985:631).
This is a critically important warning in contextual
considerations of the influences of pregnancy on psychological

distress. The findings of the present study suggest that

128

attributes of a woman psychological status, may not
necessarily be attributed to her pregnancy. Instead, they
challenge us to investigate other conditions of her life which
may more directly predict psychological status.

The findings of the present study hold some promise for
shifting pregnancy from the medicalized orientation which
seems to persist, to a more naturalized approach viewing
pregnancy as one of many role attainments that characterize
the lives of women. The findings support the “normalization”
of the pregnancy experience itself (Knafl, 1986; Anderton,
1998) and remove the mystical contributions attributed to this
otherwise reified condition (Mirowsky, 1989).

The state of pregnancy may not necessarily impart some
atypical influence, but rather may exacerbate the effect of
other conditions in the lives of women, e.g. being unmarried
while pregnant. Mirowsky reminds mental health researchers
that “sometimes the qualities of a situation determine the
effect of being in it” (Mirowsky, 1999:117). The qualities of
being pregnant while initially predicted to be characterized
by race are more influenced by marital status in the present
study. A greater caveat presented by Hobfoll (1995) found
that marital status was only predictive of depression before
delivery, with single non~cohabiting women at greater risk.
Despite the small size of the Hobfoll study and its placement

in an inner city setting (Hobfoll, 1995), it raises an

129

interesting corollary. The findings of the present study
confirm those of Hobfoll (1995), reiterating that unmarried
pregnant women are at greatest risk for psychological
distress. Intriguingly, while pregnancy is only predictive of
distress in unmarried women, this distress may resolve after
the delivery, perhaps even in the absence of a spouse.

Clearly the lives of women are composed of multilayered and
multidimensional facets, each with its own methodological
challenge for mental health researchers.

Given the additional descriptive findings that unmarried
women in this study are more distressed, regardless of current
pregnancy state, one must ponder the causal ordering of events
at work. Sociologists often grapple with causal/selection
models, a seemingly relevant line of inquiry for this
situation (Mastekaasa, 1992). At question is whether women
are depressed because they are single or are depressed because
they are pregnant. Further still, whether depressed women are
more likely to remain or become single or whether depressed
women are more or less likely to become pregnant. The
blending of these two selection paradigms with the causal
tiers just mentioned further complicates the empirical
elucidation of the contexts and events at hand.

Finally, one is struck by the final arm of the conceptual
model which frames the influence of prior pregnancy outcomes

on psychological distress. The separation of the prior

130

pregnancy measurement is further exacerbated by its lack of
correlation with the measure of current pregnancy. This seems
to suggest that indeed it is not the physiological occurrence
of conception and gestation which characterizes the lives of
women during that time but rather, the contexts and
consequences which follow. The biology of pregnancy remains
the same, whether measured in the present or retrospectively,
thus the receipt of varying findings appears representative of
events which surround the pregnancy and not necessarily the
pregnancy itself. As pregnancy state has changed throughout
the analyses of this study, what has remained consistent is
the influence of social contexts on the psychological well—
being of the women in question.

The greatest implications of this study emerge at the
intersection of mental health, pregnancy and race and evolve
from the absence of race based or pregnancy predicted
distress. Explanations for these effects seem lacking in the
current literature. Many studies recognize the numerous
factors which threaten the mental health of poor people
(Kessler, 1994). But as noted in the literature review, a
relatively small cadre of studies have investigated the mental
health of the special sub-sample of interest here, African
American women who are pregnant. Undeniably, the pregnancy
outcomes of African American women are deplorable in relation

to White women. What is equally as certain, is that these two

131

trends are unrelated and offer no useful mechanism for
directly intervening upon the Black/White gap.

Why is perinatal depression as important a consideration
as the more renowned post-partum depression? Depression seems
to increase during the third trimester and decrease after
delivery for most women. Few women actually develop post-
partum depression (Steer, 1992). However depression during
pregnancy has been correlated to poor health behaviors
(Zuckerman, 1989), and marital conflict (Kumar, 1984), yet its
role in the overall well-being of the pregnant and pregnant
woman remains contradictory in the literature (Kitamura,
1996). The occurrence of distress in the lives of pregnant
women is important not only as a quality of life concern, but
is of tantamount importance given the findings of the present
study which suggest that distress in the lives of women is not
necessarily a consequence of their pregnancy. Thus, while
this may be an important point of intervention for matters of
well—being, what is clear from the present study is that the
current pregnancy status is unrelated to psychological
distress and further that any clues to pregnancy outcomes may
unfold more cogently from retrospective considerations of past
pregnancies.

One might ponder the apparent double jeopardy found in
the dynamics between Black women and mental health and

potentially those between pregnancy and mental health. While

132

African American women may be at twice the risk for poor
pregnancy outcomes as well as at greater risk of poor mental
health, the two disadvantages present more as parallel
problems rather than tangential. This investigation fails to
substantiate any synergy in the effects of being Black and
pregnant as they relate to psychological distress. This
absence of findings is perhaps less a reflection of the
differences in the lives of Black and White women and more a
representation of what is captured in measuring ‘race.’ While
beyond the scope of this present work, these findings push us
to further consider emerging issues among race researchers
which speculate upon discrimination and other qualifications
in the meaning of race as it relates to health outcomes.

The starkness of these findings in relation to many of
those found in the literature as well as the newness of these
contributions to the literature, require close scrutiny to any
weaknesses of this study. Many of the methodological concerns
are mediated by the size and scope of the source study and
many of the most meaningful concerns are more conceptual than
procedural. Lips (1985) found that emotional symptoms
increased over the time of the pregnancy (measured symptoms
three times over the pregnancy). The present study does not
account for stage of pregnancy but given the methodological
goal of interviewing women as soon as possible after the

critical event (live birth or fetal infant death) one might

133

assume they were relatively early in the pregnancy (certainly
not third trimester).

Despite the attention in the present study to contextual
interpretations of findings, the conceptualization of
socioeconomic status, as limited by the source study,
reinforces a monolithic interpretation of SES especially as
related to race. Krieger reiterates the need to recognize the
heterogeneity within groups of women categorized by race
(Krieger, 2000). The descriptive analysis of these data also
suggest variable worth of education and income across race and
reiterate the need for creativity in this regard as described
by Krieger (2000). Future research should strive to apply
more meaningful and interpretative measures of socioeconomic
status.

These data also prevent the controlling for other
variables which may effect psychological well—being such as
prior history of mental illness or any other preexisting
psychological conditions (Hobfoll, 1995). However the
nationally representative nature and large sample size
enhances the likelihood that rates of mental illness among
this sample will not exceed national norms or unduly bias the
findings. This study also did not control for substance abuse
or alcohol, factors which have held relevance in poor
pregnancy outcomes (Parker, 2000) as well as poor health

behaviors (Zuckerman, 1989).

134

Despite these conceptual challenges, the present study
affords great confidence in the validity of its findings and
provides a major contribution to this field of work. As noted
previously, the current literature is plagued by small and
specialized study populations, placing the applicability of
most other works to date in question, despite some intriguing
findings. Most authors have noted this liability in their
work (Lips, 1985; DaCosta, 2000). The methodological rigor of
the source survey and the data analysis applied in the present
study resolve most of those concerns. In addition, many of
the studies of pregnant women have been retrospective
(Collins, 2000). This study not only applies a nationally
represented sample of large sample size, but assesses mental
health while the women remain pregnant.

The present study provides elementary explanations for
very fundamental and legitimate questions about the
psychological health of pregnant women. It also suggests the
need for a shift in direction regarding these questions. The
investigation of psychological well—being during pregnancy is
indeed a valid research question. It does not appear,
however, to be a critical consideration in questions of
pregnancy outcome or racial differences in pregnancy
experiences. Future research may elect to either investigate
the general psychological health of women during pregnancy,

recognizing that for most women there is little difficulty in

135

this regard, or to concentrate on more clinical indices and
diagnoses, which again limits the impact to small population
with clinical occurrences. Indeed, Black women tend to have
worse mental health and greater psychological distress than
their White peers and these same women also are at greater
risk for poor pregnancy outcomes. However, this investigation
seems to suggest that these two juxtaposed events are simply
complicated by the other and not caused by interactions
between them.

Future work should strive to more fully describe the
social context of the lives of women which exist during their
pregnancies. The National Maternal Infant Health Survey is
rich in information on various social domains. Subsequent
levels of analysis with these data should incorporate and
expand social context considerations, such as household
density (crowdedness with primary residence (Kitamura, 1996),
employment type and job title (Bramwell, 1997) as well as a
longitudinal analysis of depressive symptomatology, using the
follow-up data in this survey. Despite its detachment from
pregnancy status, additional investigation of psychological
distress and depressive symptoms are warranted in light of the
current literature and in consideration of the significantly
greater morbidity among the African American women in the

present study.

136

Conclusion

The familiarity with this field of research acquired
through academic training and lived experience establishes in
this investigator a particular affinity for this subject as
well as a critical mass of knowledge and information. As
such, it is a privilege to contribute to a field which is
emergent and cutting edge. Certainly the specific focus of
this investigation is maternal health and well—being but its
connection to the well—being of infants and families is an
equally important driving force. This study appears to be the
first of its magnitude and scope to empirically test important
questions regarding the psychological well-being of pregnancy
women of color. One of the methodological challenges of
research is the occasional but pertinent need to document what
has not happened. It is perhaps in this area that the present
study contributes most meaningfully, albeit inadvertently.

The empirical documentation of the absence of relationships,
which would seem otherwise reasonable and expected, is a Vital
contribution to directing the next step in the work on
pregnancy and mental health.

The methodological strength of this study combined with
its unanticipated findings may actually raise more questions
than answers, a characteristic not uncommon in good studies
(Krieger, 2000). It is with some incredulity that we secede

the expectation for race—based differences in the pregnancy

137

experiences of women. However this findings entreats us to
avoid reductionist interpretations of the experiences of women
of color. Race is more than a descriptor but rather frames
the nature of the experiences of certain communities of women
in this society. Epidemiology and biomedical sciences are
just awakening to the discovery that race is more than innate
biology or genetic variation (Krieger, 2000), a point long
known and established in sociology. Given the recent and
emergent interest in the relationship of social contributions
to health and health outcomes, sociology is conspicuously
overdue in re-engaging its voice and imagination in this
discourse.

Understanding and interpreting the social worlds and
contexts of the lives of individuals is what we do as
sociologists. Race is a social construct which has clearly
delineated the pregnancy experiences and outcomes for women.
It is now the charge to creatively identify the true
differences which exist between women of varied races which
precipitate disparate outcomes. The findings of this
investigation confirm that pregnancy is not the context; women
experience no differences in that domain. On the other hand,
what is known from this study is that the experiences of women
vary along social domains and clearly along the prior
pregnancy outcome. The next aspiration becomes the

clarification of particular dynamics, given the emergent

138

 

knowledge that pregnancy is less likely the culprit than

suspected.

139

LIST OF REFERENCES

Aderibigbe, Y.A., O. Gureje, O. Omigbodun. 1993. “Postnatal
Emotional Disorders in Nigerian Women: A Study of
Antecedents and Associations.” British JOurnal of
Psychiatry 163:645—650.

Anderton, Joan, Helen Elfert, and Magdalene Lai. 1989.
“Ideology in the Clinical Context: Chronic Illness,
Ethnicity and the Discourse of Normalisation.” Sociology
of Health and Illness 11:253—278.

Aneshensel, Carol 8., Ralph R. Frerich, and Virginia A. Clark.
1981. “Family Roles and Sex Differences in Depression.”
Journal of Health and Social Behavior 22:379-393.

Aneshensel, Carol 8., Carolyn Rutter, and Peter Lachenbruch.
1991. “Social Structure, Stress, and Mental Health:
Competing Conceptual and Analytical Models.” American
Sociological Review 56: 166-178.

Aneshensel, Carol S. 1992. “Social Stress: Theory and
Research.” Annual Review of Sociology 18: 15—38.

Aneshensel, Carol. 1999. “The Sociology of Mental Health:
Surveying the Field.” In Handbook of the Sociology of
Mental Health, Carol Aneshensel and Jo Phelan, Eds.
Plenum Publishers, New York.3—18.

Arno, P.S., T.E. Awerbuch, A.T. Geronimus, James S. House, I.
Kawachi, B. Kennedy, R. Levins, Bruce Link, Jo C. Phelan,
D.B. Smith, R. Wallace and David R. Williams. 1998.
“Statement of Working Group on Social Determinants of
Health on Healthy PeOple 2010 Objects.” Memorandum to
Dr. David Satcher, Assistant Secretary for Health and
Surgeon General. December 15.

Avison, William R. 1995. “Roles and Resources: The Effects of
Family Structure and Employment on Women’s Psychosocial
Resources and Psychological Distress." Research in
Community Health 8: 233—256.

Barber, Jennifer, William Axinn, and Arland Thornton. 1999.
“Unwanted Childbearing, Health, and Mother~Child

Relationships,” JOurnal of Health and Social Behavior
40:231-257.

140

Beck, Cheryl. 1999. “Available Instruments for Research on
Prenatal Attachment and Adaptation to Pregnancy.”
Maternal Child Nursing 24:1, 25—32.

Benedict, Mary, Lisa Paine, Lori Paine, Diane Brandt, and
Rebecca Stallings. 1999. “The Association of Childhood
Sexual Abuse with Depressive Symptoms During Pregnancy,
and Selected Pregnancy Outcomes.” Child Abuse and Neglect
23:7, 659—670.

Bhopal, Raj and L. Donaldson. 1998. “White, European, Western,
Caucasian, or What? Inappropriate Labeling in Research on
Race, Ethnicity, and Health.” American JOurnal of Public
Health 88: 1303-1307.

Bramwell, Ros. 1997. “Studying the Impact of the Psychosocial
Work Environment on Pregnancy Outcome: A Review of
Methodological Issues.” Journal of Reproductive and
Infant Psychology 15:3/4, 257—269.

Brett, K.M., D.S. Strogatz, and D.A. Savits. 1997.
“Employment, Job Strain, and Pre-term Delivery Among
Women in North Carolina.” American JOurnal of Public
Health 87: 199—204.

Broman, Clifford L. 1996. “The Health Consequences of Racial
Discrimination: A Study of African Americans.” Ethnicity
and Disease 6:148—153.

Broman, Clifford L., V.L. Hamilton, W.S. Hoffman, and R.
Mavaddat. 1995. “Race, Gender, and the Response to
Stress: Autoworkers’ Vulnerability to Long-term
Unemployment. American Journal of Community Psychology 23
(6): 813-842.

Broman, Clifford. 1993. “Race Differences in Marital Well—
being.” Journal of Marriage and the Family 55: 724-732.

Brown, Anita C., Gene Brody, and Zolina Stoneman. 2000. “Rural
Black Women and Depression: A Contextual Analysis.”
JOurnal of Marriage and the Family 62: 187—198.

Brown, D. R., F. Ahmed. 1995. “Major Depression in a Community
Sample of African Americans.” American JOurnal of
Psychiatry 152: 373—378.

Bruce, M.L., K.M. Kim. 1992. “Differences in Effect of Divorce

on Major Depression in Men and Women.” American Journal
of Psychiatry 149 (7): 914—933.

141

Bryce, R.L., F.J. Stanley, and MW Engin. 1988. “The Role of
Social Support in the Prevention of Pre-term Birth. Birth
15:19—22.

Beaching, D.P., M.L. Glasser, and D.A. Frame. 1986.
“Progression of Depression in the Prenatal and Post-
partum Periods.” WOmen and Health 61-78.

Carmichael, S.L. 1998. “Changes in the Black—White Infant
Mortality Gap from 1983-1991 in the United States.”
Journal of Preventive Medicine 15: 220—227.

Christensen, Bruce J. 1992. “In Sickness and in Health: The
Medical Costs of Family Meltdown.” Policy Review, Spring:
70—72.

Cobb, Sidney 1990. “Social Support as a Moderator of Life
Stress.” Psychosomatic Medicine 38: 300-314.

Cohen, S. and S.L. Same. 1985. Social Support and Health.
Academic Press, Orlando, FL

Collins, James W., Richard David, Rebecca Simons, Ads Handler,
Stephen Wall and Lisa Dryer. 2000. “Low—Income African
American Mothers’ Perception of Exposure to Racial
Discrimination and Infant Birth Weight.” Epidemiology 11:
337-339.

Collins, Nancy L., Christine Dunkel—Schetter, Marci Lobel, and
Susan Scrimshaw. 1993. “Social Support in Pregnancy:
Psychosocial Correlates of Birth Outcomes and Post—partum
Depression.” Journal of Personality and Social Psychology
65: 1243—1250.

Collins, Patricia Hill. 1991. Black Faminist Thought.
Routledge, Chapman, and Hall, Inc. New York, NY.

Cooper, R.L., N. Elder, R. Ramsey, and F. Johnson. 1996. “The
Pre-term Prediction study: Maternal Stress is Associated
with Spontaneous Pre-term Birth at Less than Thirty-five
Week’s Gestation.” American Journal of Obstetrics and
Gynecology 175: 1286-1292.

Covington, Sharon, and Susan Theut. 1993. “Reactions to
Perinatal Loss: A Qualitative Analysis of the National
Maternal and Infant Health Survey,”American Journal of
Orthopsychiatry 63: 215-222.

142

Cowie, C.C., F.K. Port, R.A. Wolfe, P.J. Savage, P.P. Moll,
and V.M. Hawthorne. 1992. “Disparities in Incidence of
Diabetic End-Stage Renal Disease According to Race and
Type of Diabetes.” The New England JOurnal of.Medicine
321: 1074-1079.

Coyne, J.C. and G. Downey. 1991. “Social Factors and
Psychopathology: Stress, Social Support and Coping
Processes.” Annual Review of Psychology 42: 401-425.

Crnic, K.A., M.T. Greenberg, A.S. Ragozin, N.M. Robinson, and
R.B. Basham. 1983. “Effects of Stress and Social Support
on Mothers and Premature and Full—term Infants.” Child
Development 54: 209-217.

DaCosta, Deborah, Julie LaRouche, Maria Drista, and William
Brender. 2000. “Psychosocial Correlates of Pre-partum and
Post-partum Depressed Mood.” Journal of Affective
Disorders 59:31-40.

David, Richard J. and John Collins. 1991. “Bad Outcomes in
Black Babies: Race or Racism?” Ethnicity and Disease 1:
236-244.

Deal, Lisa W., Victoria L. Holt. 1998. “Young Maternal Age and
Depressive Symptoms: Results from the 1988 National
Maternal and Infant Health Survey.” American Journal of
Public Health 88: 266-270.

Dragonas, Thalia, Konstantine Petrogiannis, and Helen Adam.
1997. “Working Women, their Emotional Well—being and
Pregnancy in Greece.” JOurnal of Reproductive and Infant
Psychology 15: 239-256.

Dunkel-Schetter, Christine. 1998. “Maternal Stress and Pre—
term Delivery.” Prenatal and Neonatology Medicine 3: 39-
42.

Dunston, Patricia J. 1990. “Stress, Coping, and Social
Supports: Their Effects on Black Women.” In Handbook of
.Mental Health and Mental Disorders Among Black Americans.
Edited by D. Ruiz. Westport, CT: Greenwood Press.

Elliott, Timothy R., Richard Shewchuk, Cheryl Richeson, Heidi
Pickelman, and Kristin Franklin. 1996. “Problem-solving
Appraisal and the Prediction of Depression During
Pregnancy and in the Post-partum Period,” JOurnal of
Counseling and Development 74: 645—651.

143

Fiscella, Kevin. 1996. “Ethnicity, Poverty, Health, and the
American Dream.” Family Systems and Health 14: 361-363.

Flack, J.M., H. Amaro, W. Jenkins, S. Kunitz, J. Levy, M.
Mixon, and E. Yu. 1995. “Epidemiology of Minority
Health.” Health Psychology 14: 592-600.

Fullilove, Melinda T. 1998. “Comment: Abandoning “Race” as a
Variable in Public Health Research: An Idea Whose Time
has Come.” American Journal of Public Health 88: 1297-
1298.

Goldberg, K.C., A.J. Hartz, S.J. Jacobsen, H. Krakauer, and
A.A. Rimm. 1992. “Racial and Community Factors
Influencing Coronary Artery Bypass Graft Surgery Rates
for all 1986 Medicare Patients.” JOurnal of the American
Medical Association 267: 1473—1477.

Goldenburg, Robert L. and Dwight J. Rouse. 1998. “Prevention
of Premature Birth.” Medical Progress 339: 313-320.

Goldman, N. and Y. Hu. 1993. “Excess Mortality Among the
Unmarried: A Case Study of Japan.” Social Science and
.Medicine 36(4): 533—546.

Gotlib, Ian H., Valerie E. Whiffen, John H. Mount, Kenneth
Milne, and Nikkie I. Cordy. 1989. “Prevalence Rates and
Demographic Characteristics Associated with Depression in
Pregnancy and Post—partum.” JOurnal of Consulting and
Clinical Psychology 57: 269-274.

Hedegaard, Morten, Tine B. Henriksen, Svend Sabroe, and Niels
J. Secher. 1993. “Psychological Distress in Pregnancy and
Pre-term Delivery.” British Journal of Medicine 307:
234-39.

Hemstrom, Orjan. 1996. “Is Marriage Dissolution Linked to
Differences in Mortality Risks for Men and Women?”
Journal of Marriage and the Family 58: 366-378.

Henly, Julia R. 1997. “The Complexity of Support: The Impact
of Family Structure and Provisional Support on African
American and White Adolescent Mothers' Well-being.”
American Journal Community Psychology 25: 629—655.

144

 

Hobfoll, Stevan E., Christine Ritter, Justin Lavin, Michael
Hulsizer, and Rebecca Cameron. 1995. “Depression
Prevalence and Incidence Among Inner—City Pregnant and
Post-partum Women,” JOurnal of Consulting and Clinical
Psychology 63: 445-453.

Hobbs, Daniel F. and Sue P. Cole. 1976. “Transition to
Parenthood: A Decade Replication.” Journal of Marriage
and the Family, November: 723-731.

Hoffman, S. and M.C. Hatch. 1996. “Stress, Social Support and
Pregnancy Outcome: A Reassessment Based on Recent

Research.” Paediatric and Perinatal Epidemiology 10: 380
405.

Hogue, Carol J. and M.A. Hargraves. 1993. “Class, Race, and
Infant Mortality in the United States.” American JOurnal
of Public Health 83: 9-12.

Horwitz, Allan. 1999. “The Sociological Study of Mental
Illness: A Critique and Synthesis of Four Perspectives.”
In Handbook of the Sociology of Mental Health, Carol
Aneshensel and Jo Phelan, Eds. Plenum Publishers, New
York: 57—78.

House, James, Bruce Link, and David Williams. 1998.
“Memorandum to Surgeon General David Satcher: Statement
of Working Group on Social Determinants of Health on
Healthy People 2010 Objectives-Draft for Public Comment,
by the Working Group on Social Determinants of Health of
the Robert Woods Johnson Foundation Investigator Awards,”
December 15, 1998.

Hu, Y. and Nancy Goldman. 1990. “Mortality Differentials by

Marital Status: An International Comparison.” Demography
27: 233—249.

Hummer, Robert A., Carl Schmertmann, Isaac Eberstein, and
Susan Kelly. 1995. “Retrospective Reports of Pregnancy
Wantedness and Birth Outcomes in the United States,”
Social Science Quarterly 76: 402-418.

Hummer, Robert A. 1993. “Racial Differentials in Infant
Mortality in the US: An Examination of Social and Health
Determinants,” Social Fbrces 72: 529—554.

Istvan, J. 1986. “Stress, Anxiety, and Birth Outcomes: A

Critical Review of the Evidence.” Psychological Bulletin.
100: 331-348.

145

Jacobs, Janet. 1994. “Gender, Race and Class, and the Trend
toward Early Motherhood,” Journal of Contemporary
Ethnography 22: 442-462.

Jackson, James, Tony Brown, David Williams, M. Torres, S.L.
Sellars, K. Brown. 1996. “Racism and the Physical and
Mental Health Status of African Americans: A Thirteen
Year National Panel Study.” Ethnicity and Disease 6:132-
147.

Kawachi, I, B.P. Kennedy, K. Lochner, and D. Prithrow-Stith.
1997. “Social Capital, Income Inequality and Mortality.”
American JOurnal of Public Health 87: 1491—1498.

Keppel, G. 1991. Design and Analysis: A Researcher’s Handbook.
Third Edition. Prentice Hall, Upper Saddle River, NJ.

Kennerley, H. and Gath, D. 1989. “Maternity Blues:
Associations with Obstetric, Psychological, and
Psychiatric factors.” British Journal of Psychiatry 155:
367-373.

Kerlinger, Fred N. and Elazar J. Pedhazur. 1973. MUltiple
Regression in Behavioral Research. Holt, Rinhart and
Winston, Inc. New York, NY.

Kessler, Ronald, K. McGonagle, S. Zhao, C.B. Nelson, M.
Hughes, S. Eshelman, H. Wittchen, Kenneth Kendler. 1994.
“Lifetime and 12-month prevalence of DSM-III—R
Psychiatric Disorders in the United States.” Archives of
General Psychiatry 51: 8-19.

Kitamura, T., M. Sugawara, K. Sugarwaru, M.A. Toda, and S.
Shink. 1996. “Psychosocial Study of Depression in Early
Pregnancy.” British Journal of Psychiatry 168: 732—738.

Kitson, Gay C. and Leslie A. Morgan. 1990. “The Multiple
Consequences of Divorce: A Decade Review,” JOurnal of
Marriage and the Family 52: 913—924.

Knafl, Kathleen and J. Deatrick. 1986. “How Families Manage
Chronic Conditions: An Analysis of the Concepts of
Normalization.” Research in Nursing and Health 5:4—14.

KOgan, Michael, Greg Alexander, Milton Kotelchuck, David
Nagey, and Brian Jack. 1994. “Comparing Mothers’ Reports
on the Content of Prenatal Care Received with Recommended
National Guidelines for Care.” Public Health Reports 109:
637-646.

146

 

Kost, Kathryn, Jacqueline Forrest. 1995. “Intention Status of
U.S. Births in 1988: Differences by Mother’s
Socioeconomic and Demographic Characteristics.” Family
Planning Perspectives 27: 11-17.

Krieger, Nancy. 2000. “Epidemiology, Racism, and Health: The
Case of Low Birth Weight." Epidemiology 11: 237—239

Krieger, Nancy, David Williams, and Sally Zierler. 1999.
“‘Whiting Out' White Privilege Will Not Advance the Study
of How Racism Harms Health.” American Journal of Public
Health 89: 782.

Krieger, Nancy 1990. “Racial and Gender Discrimination: Risk
Factors for High Blood Pressure.” Social Science and
Medicine 30: 1273~1281.

Krieger, Nancy, Diane Rowley, A.A. Herman, Billie Avery, and
M.T. Phillips. 1993. “Racism, Sexism, and Social Class:
Implications for Studies of Health, Disease, and Well-
being.” American Journal of Preventive Medicine
Supplement 9: 82—122.

Krieger, Nancy, and S. Sidney. 1996. “Racial Discrimination
and Blood Pressure: The CARDIA Study of Young Black and
White Adults." American Journal of Public Health 86:
1370-1378.

Kumar, R. and Kay M. Robinson. 1984. “A Prospective Study of
Emotional Disorders in Childbearing Women.” British
Journal of Psychiatry 144: 35-47.

Landrine, Hope, Elizabeth A. Klonoff, J. Gibbs, V. Manning,
and M. Lund. 1995. “Physical and Psychiatric Correlates
of Gender Discrimination.” Psychology of Women Quarterly
19: 473—492.

LaVeist, Thomas A. 1996. “Why We Should Continue to Study
Race...But do a Better Job: An Essay on Race, Racism and
Health.” Ethnicity and Disease 6: 21—29.

Lederman, Regina and Diana Miller. 1998. “Adaptation to
Pregnancy in Three Different Ethnic Groups: Latin-
American, African American, and Anglo-American.” Canadian
Journal of Nursing Research 30:3; 37—51.

147

 

 

Levin, Jeffrey S. and Richard S. DeFrank. 1988. “Maternal
Stress and Pregnancy Outcomes: A Review of the
Psychosocial Literature.” Journal of Psychosomatic
Obstetrics and Gynecology 9: 3-16.

Lindberg, Laura, Freya Sonenstein, Leighton Ku, and Gladys
Martinez. 1997. “Age Differences Between Minors Who Gave
Birth and their Adult Partners.” Family Planning
Perspectives 29: 61—66.

Lips, Hilary M. 1985. “A Longitudinal Study of the Reporting
of Emotional and Somatic Symptoms during and after
Pregnancy.” Social Science and Medicine 21: 631-640.

Lobel, Marci. 1992. “Prenatal Maternal Stress and Prematurity:
A Prospective Study of Socioeconomically Disadvantaged
Women.” Health Psychology 11: 32—40.

Lobel, Marci. 1994. “Conceptualizations, Measurements and
Effects of Prenatal Maternal Stress on Birth Outcomes.”
Journal of Behavioral Medicine 17: 225—272.

Lomas, Johnathan. 1998. “Social Capital and Health:
Implications for Public Health and Epidemiology.” Social
Science and Medicine 47: 1181-1188.

Mamelle, N., A. Measson, F. Munoz, M. Audras de la Bastie, P.
Gerin, M.T. Hanauer, F. Collet and J. Guyotat. 1989.
“Development and Use of Self-administered Questionnaire
for Assessment of Psychologic Attitudes Toward Pregnancy
and their Relation to a Subsequent Premature Birth.”
American Journal of Epidemiology 130: 989-998.

Martin, Emily. 1997. “Medical Metaphors of Women's Bodies.” In
Medina, N. and S. Stanbury, Writing the Body. Columbia
University Press, New York.

Mastekaasa, A. 1992. “Marriage and Psychological Well—being:
Some Evidence on Selection Into Marriage.” Journal of
Marriage and the Family 54: 901—911.

Mirowsky, John. 1999. “Analyzing Associations between Mental
Health and Social Circumstances.” In Carol Aneshensel and
Jo Phelan (Eds), Handbook of the Sociology of Mental
Health. Plenum Publishers, New York, NY.

Mirowsky, John and Catherine Ross. 1989. “Psychiatric

Diagnosis as Reified Measurement.” Journal of Health and
Social Behavior 30: 11—25.

148

Morbidity Mortality Weekly Report. 1995. 44:261-264

Murrell, Nancy L., R. Smith, G. Gill, and G. Oxley. 1996a.
“Racism and Health Care Access: A Dialogue with

Childbearing Women”. Health Care for WOmen International
17: 149—159.

Murrell, Nancy L. 1996b. “Stress, Self—esteem, and Racism:
Relationships with Low Birth Weight and Pre-term Delivery
in African American Women.” Journal of National Black
Nurses Association 8: 45—53.

Myrdal, Gunnar, Richard Sterner, and Arnold Rose. 1944. An
American Dilemma: The Negro Problem and Medern Democracy.
Harper and Brothers; New York, NY.

Myers, S.T. and Grasmick, H.G. 1990. “The Social Rights and
Responsibilities of Pregnant Women: An Application of
Parson’s Sick Role Model. “ Journal of Applied Behavioral
Science 26: 157—172.

National Advisory Mental Health Council Behavioral Science
Workgroup. 1999. Translating Behavioral Science into
Action. Website
http://www.nimh.nih.gov/council/bswsummary.cfm

Najman, J.M., M.J. Anderson, W. Boor, M.J. O’Callaghan, and
G.M. Williams. 2000. “Postnatal Depression — Myth and
Reality: Maternal Depression before and after the Birth
of a Child.” Social Psychiatry and Psychiatric
Epidemiology 35: 19-27.

Neumann, Joy P. 1986. “Gender, Life Strain, and Depression.”
Journal of Health and Social Behavior 27:161-178.

Norbeck, Jane S. and N.J. Anderson. 1989a. “Life Stress,
Social Support and Anxiety in Mid and Late Pregnancy

among Low Income Women.” Research in Nursing and Health
12: 281—287.

Norbeck, Jane S. and N.J. Anderson. 1989b. “Psychosocial
Predictors of Pregnancy Outcomes in Low—Income Black,

Hispanic, and White Women.” Nursing Research 36: 204-209.

Oakley, Ann. 1980. Women Confined: Towards a Sociology of
Childbirth. New York, NY: Schocken Books, Inc.

149

Oakley, Ann, L. Rajan, and A. Grant. 1990. “Social Support and
Pregnancy Outcomes.” British JOurnal of Obstetrics and
Gynecology 97: 155-162.

Oakley, Linda D. 1986. “Marital Status, Gender Role Attitude,
and Black Women’s Depression.” Journal of the National
Black NUrses Association 1(1): 41-51.

O’Campo, P., X. Xue, M. Wang, and M.O. Caughy. 1997.
“Neighborhood risk factors for low birth weight in
Baltimore: A Multi-level Analysis.” American Journal of
Public Health 87: 1113-1118.

O’Hara, M.W., D.J. Neunaber, and E.M. Zekoski. 1984.
“Prospective Study of Post—partum Depression: Prevalence,
Course, and Predictive Factors.” Journal of Abnormal
Psychology 93: 158-171.

O’Hara, M. 1986. “Social Support, Life Events, and Depression
during Pregnancy and the Puerperium.” Archives of
General Psychiatry 43: 569—573.

Orr, Suzanne T. and Sherman James. 1984. “Maternal Depression
in an Urban Pediatric Practice: Implications for Health
Care Delivery." American Journal of Public Health 74:
363-365.

Orr, Suzanne T., Sherman James, B.J. Burns, and B. Thompson.
1989. “Chronic Stressors and Maternal Depression:
Implications for Prevention.” American Journal of Public
Health 79: 1295—1296.

Orr, Suzanne T. and C.A. Miller. 1995. “Maternal Depressive
Symptoms and the Risk of Poor Pregnancy Outcome: Review
of the Literature and Preliminary Findings.”
Epidemiologic Reviews 17: 165—171.

Paarlberg, Marieke K., A.D. Vingerhoets, J. Passchier, G.A.
Kekker, and H.P. Van Geijn. 1995. “Psychosocial Factors
and Pregnancy Outcomes: A Review with Emphasis on
Methodological Issues.” Journal of Psychosomatic Research
39: 563-595.

Pappas, George. 1994. “Elucidating the Relationships between
Race, Socioeconomic Status and Health.” American JOurnal

of Public Health 8: 892-893.

Parker, J.D. 2000. “Birth weight Trends Among Interracial
Black and White Infants.” Epidemiology 11:242-248.

150

Pearlin, Leonard I., E.G. Menaghan, M.A. Lieberman, and J.T.
Mullan. 1981. “The Stress Process.” Journal of Health and
Social Behavior 22: 337-356.

Pearlin, Leonard I. 1989. “The Sociological Study of Stress.”
Journal of Health and Social Behavior 30: 241-256.

Pearlin, Leonard. 1999. “The Stress Process Revisited:
Reflections and Concepts and their Interrelationships.”
In Aneshensel, C and Phelan, J. (Eds.) Handbook of the
Sociology of.Mental Health. New York: Plenum Publishers.

Pfost, K.S., M.J. Stevens, and C.U. Lum. 1990. “The
Relationship of Demographic Variables, Antepartum
Depression, and Stress to Post-partum Depression."
Journal of Clinical Psychiatry 46:5, 388-592.

Powell, S. and D. Drotar. 1992. “Post—partum Depressed Mood-
The Impact of Daily Hassles.” Journal of Psychosomatic
Obstetrics and Gynecology 13: 255—266.

Pritchard, C.W., P.Y. prhm. 1994. “Pre-term Birth, Low Birth,
and the Stressfulness of the Household Role for Pregnant
Women.” Social Science and Medicine 38: 89-96.

Radloff, L.S. 1977. “The CES-D Scale: A Self-reported
Depression Scale for Research in the General Population.”
Applied Psychological Measurement 1: 385-401.

Ramsey, Christian N., Troy D. Abell, and Lisa C. Baker. 1986.
“The Relationship between Family Functioning, Life
Events, Family Structure, and the Outcome of Pregnancy.”
Journal of Family Practice 6: 521—527.

Raskin, Valerie, Judith Richman, and Cheryl Gaines. 1990.
“Patterns of Depressive Symptoms in Expectant and New
Parents.” Clinical Research Reports 147: 658-660.

Reiger, D.A., J.K. Myers, M. Kramer, Lee Robins, D.G. Blazer,
R.L. Hough, W.W. Eaton, and 8.2. Locke. 1984. “The
National Institutes of Mental Health Epidemiologic
Catchment Area Program: Historical Context, Major
Obstacles and Study Population Characteristics.” Archives
of General Psychiatry 41: 934-941.

Roberts, E.M. 1997. “Neighborhood Social Environments and the

Distribution of Low Birth Weight in Chicago.” American
Journal of Public Health 87:597-603.

151

Robins, Lee N., J. Wing, H-U Wittchen. 1988. “The Composite
International Diagnostic Interview: An Epidemiologic
Instrument Suitable for Use in Conjunction with Different
Diagnostic Systems and in Different Cultures.” Archives
of General Psychiatry 45: 1069-1077.

Robins, Lee N. and D.A. Reiger. 1991. Psychiatric Disorders in
America: The Epidemiologic Catchment Area Study. The Free
Press, New York, NY.

Rowley, Diane L., Carol J. Hogue, Cheryl Blackmore, Cynthia
Ferre, Kendra Hatfield-Timajchy, Priscilla Branch, Hani
Atrash. 1993. “Preterm Delivery Among African American
Women: A Research Strategy.” In Racial Differences in
Preterm Delivery: Developing a New Research Paradigm,
Diane Rowley and Heather Tosteson, Eds., American Journal
of Preventive Medicine 9: 1-6.

 

Rowley, Diane L. 1995. “Framing the Debate: Can Prenatal Care
Help to Reduce the Black-White Disparity in Infant
Mortality?” JOurnal of American Medical WOmen’s
Association 50:187-193.

Rubin, Reva. 1984. Maternal Identify and the Maternal
Experience. Springer Publishing, Inc. New York, NY

Sagrestano, Linda, Pamela Feldman, Christine K. Rini, Grace
Woo, and Christine Dunkel-Schetter. 1999. “Ethnicity and
Social Support During Pregnancy." American Journal of
Community Psychology 27(6): 869-898.

Sanderson, M., P.J. Placek, and K.G. Keppel. 1991. “The 1988
National Maternal Infant Health Survey: Design, Content,
and Data Availability." Birth 18: 26-32.

Schmeelk, Karen, Douglas Granger, Elizabeth Susman, and George
Chrousos. 1999. “Maternal Depression and Risk for Post-
partum Complications: Role of Prenatal Corticotrophin-
Releasing Hormone and Interleukin-1 Receptor Antagonist.”
Behavioral.Medicine 25:88-94.

Schoendorf, KC, Carol J. Hogue, J.C. Kleinman, Diane Rowley.
1992. “Mortality among Infants of Black as Compared to
White College—educated Parents.” New England Journal of
Medicine 326: 1522-1526.

152

Séguin, Louise, Louise Ptovin, Michele St-Denis, and Jacinthe
Loiselle.1999. “Socio-environmental Factors and Post-
natal Depressive Symptomatology: A Longitudinal Study,”
Wbmen and Health 29: 57-71.

Sheehan, T. Joseph. 1998. “Stress and Low Birth Weight: A
Structural Modeling Approach Using Real Life Stressors.”
Social Science and.Medicine 47:10, 1503-1512.

Simon, R.W. and K. Marcussen. 1999. “Marital Transitions,
Marital Beliefs, and Mental Health.” JOurnal of Health
and Social Behavior 40: 111-125.

Smith, G.E., M. Gerrard, and F.X. Gibbons. 1997. “Self-esteem
and the Relation Between Risk Behavior and Perceptions of
Vulnerability to Unplanned Pregnancy in College Women.”
Health Psychology 16: 137—146.

Somervell, Philip D., Philip J. Leaf, Myrna Weissman, Dan G.
Blazer, and Martha L. Bruce. 1989. “The Prevalence of
Major Depression in Black and White Adults in Five United
States Communities.” American JOurnal of Epidemiology
130:725—735.

Stancil, Tonya R., Irva Hertz—Picciotto, Margaret Schramm, and
Margaret Watt-Morse. 2000. “Stress and Pregnancy Among
African American Women.” Paediatric and Perinatal
Epidemiology 14: 127-135.

Steer, Robert A., Theresa O. Scholl, Mary L. Hediger, Richard
L. Fisher. 1992. “Self-reported Depression and negative

Pregnancy outcomes.” Journal of Clinical Epidemiology 45:
1093—1099.

Steer, Robert A. and Theresa Scholl. 1990. “Revised Beck
Depression Inventory Scores of Inner-city Adolescents:
Pre and Post-partum.” Psychological Quarterly 66: 315-
320.

Steinberg, Susanne and Fancois Bellavance. 1999.
“Characteristics and Treatment of Women with Antenatal
and Post—partum Depression.” International Journal of
Psychiatry in Medicine 29:209-233.

Stern, G. and L. Kruckman. 1983. “Multi-disciplinary
Perspectives on Post—partum Depression: An
Anthropological Critique.” Social Science and Medicine
17: 1027-1041.

153

Sullivan, Patrick F., Ronald Kessler, Kenneth Kendler. 1998.
“Latent Class Analysis of Lifetime Depressive Symptoms in
the National Comorbidity Survey.” American JOurnal of
Psychiatry 155:1398—1406.

Sword, Wendy. 1999. “A Socio-ecological Approach to
Understanding Barriers to Prenatal Care for Women of Low
Income.” Journal of Advanced Nursing 29: 1170-1177.

Syme, S. Leonard. 1998. “Social and Economic Disparities in
Health: Thoughts about Intervention.” The Milbank
Quarterly 76: 493-505.

Takeuchi David, R.C. Chung. 1998. “Lifetime and Twelve-Month
Prevalence Rates of Major Depressive Episodes and
Dysthmia Among Chinese Americans in Los Angeles.”
American Journal of Psychiatry 155: 1407-1414.

Takeuchi, David, R.K. Adair. 1992. “The Exposure and
Vulnerability of Ethnic Minorities to Life Events.
Research in Community and.Mental Health 7: 111-124.

Taylor, Hazel and Deborah Baker. 1997. “Employment, Parity,
and Single Parenthood: Their Impact of Health in
Pregnancy.” Journal of Reproductive and Infant Psychology
15 (3/4): 221—237.

Thompson, V.L.S. 1996. “Perceived Experiences of Racism as
Stressful Life Events.” Community Mental Health Journal
32: 223-233.

Thorpe, Karen J., T. Dragonas, and J. Golding. 1992. “The
Effects of Psychosocial Factors on the Emotional Well-
being of Women during Pregnancy: A Cross Cultural Study
of Britain and Greece.” Journal of Reproductive and
Infant Psychology 10: 191-204.

Tiedje, Linda Beth, Camille B. Wortman, G. Downey, C. Emmons,
M. Biernat, and E. Lang. 1990. “Women with Multiple
Roles: Role-compatibility Perceptions, Satisfaction, and
Mental Health.” Journal of Marriage and the Family 52:
63-72.

Tiejten, A.M. and C.E. Bradley. 1985. “Social Support and
Maternal Psychosocial Adjustment during the Transition to
Parenthood.” Canadian Journal of Behavioral Science
17:109-121.

154

Turner, R. Jay, Grindstaff, Carl and Phillips, Norman. 1990.
“Social Support and Outcomes in Teenage Pregnancy."
JOurnal of Health and Social Behavior 31: 43-57.

Turner, R. Jay, Sorenson, Ann, and Turner, J. Blake. 2000.
“Social Contingencies in Mental Health: A Seven Year
Follow—up Study of Teenage Mothers.” JOurnal of Marriage
and the Family 62: 777-791.

Ulbrich, Patricia M., George J. Warheit, and R.S. Zimmerman.
1989. “Race, Socioeconomic Status, and Psychological
Distress: an Examination of Differential Vulnerability.”
Journal of Health and Social Behavior 30: 131-146.

Umberson, Debra, M.C. Chen, James S. House, K. Hopkins, and E.
Slaten. 1996. “The Effect of Social Relationships on
Psychological Well-being: Are Men and Women Really so
Different?” American Sociological Review 61: 837-857.

Umberson, Debra. 1992. “Gender, Marital Status, and the Social
Control of Health Behavior.” Social Science and.Medicine
34(8): 306-319.

Upchurch, Dawn and Carol Aneshensel. 1999. “Neighborhood and
Family Contexts of Adolescent Sexual Activity.” Journal
of Marriage and the Family 61: 920-933.

US Health and Human Services. 1998. “Vital and Health
Statistics. 1998 National Maternal Health and Infant
Survey: Methods and Response Characteristics.” DHHS
Publication No. 98-1399

Villarosa, Linda. 1994. Body and Soul: The Black Women’s Guide
to Physical Health and Emotional Well—being. New York:
Harper Collins Publishers.

Visness, Cynthia and Kathy Kennedy. 1997. “Maternal Employment
and Breast-feeding: Findings from the 1988 National
Maternal and Infant Health Survey.” American J0urnal of
Public Health 87: 945-950.

Warren, Barbara J. 1997. “Depression, Stressful Life Events,
Social Support, and Self-Esteem in Middle Class African
American Women.” Archives of Psychiatric Nursing 9:107-
116.

Weissman, Myrna M. and Gerald Klerman. 1977. “Sex Difference

and the Epidemiology of Depression.” Archives of General
Psychiatry 34: 98-111.

155

Weissman, Myrna M. and Mark Olfson. 1995. “Depression in

Women: Implications for Health Care Research.” Science
269: 799—801.

Wheaton, Blair. 1999. “Social Stress.” In Aneshensel, C and
Phelan, J. (Eds.) Handbook of the Sociology of Mental
Health. New York: Plenum Publishers.

Wijma, K. and B. Wijma. 1992. “Changes in Anxiety During
Pregnancy and After Delivery.” In Wijma, K. and von
Schoultz, B. (Eds.) Reproductive Life. Carnforth, U.K.:
Parthenon Publishing Group.

Williams, David R. 1992a. “Black-White Differences in Blood
Pressure: the Role of Social Factors." Ethnicity and
Disease 2: 126-141.

Williams, David R., David Takeuchi, and Russell Adair. 1992b.
“Socioeconomic Status and Psychiatric Disorder among
Blacks and Whites,” Social Forces 74: 179-294.

Williams, David R. 1993. “Race in the Health of America:
Problems, Issues, and Directions: Summary of Plenary
Presentations.”.MOrbidity and.Mbrtality Weekly Report 42
(No.RR—10), 8—10.

Williams, David R. 1996. “Racism and Health: A research
Agenda.” Ethnicity and Disease 6: 1—6.

Yu, Stella, Kenneth Keppel, Gopal Singh, and Woodie Kessel.
1996. “Preconceptional and Prenatal Multivitamin-Mineral
Supplemental Use in the 1988 National Maternal and Infant
Health Survey.” American Journal of Public Health 86:
240-242.

Zachariah, Rachel. 1996. “Predictors of Psychological Well-
being of Women during Pregnancy: Replication and

Extension.” JOurnal of Social Behavior and Personality
11:127-140.

Zuckerman, Barry, Hortensia Amaro, Howard Bauchner, and Howard
Cabral. 1989. “Depressive Symptoms during Pregnancy:
Relationship to Poor Health Behaviors.” American Jaurnal
of Obstetrics and Gynecology 160:1107—1111.

156

 

S .

 

 

11111 11 111111111111 11111

1293 021777