The subjective well-being of widowed men and women


Material Information

The subjective well-being of widowed men and women the influence of social support, social interaction and religion
Physical Description:
x, 218 leaves : ; 28 cm.
Beck, Rubye Wilkerson, 1957-
Publication Date:


Subjects / Keywords:
Widowers -- Florida -- Gainesville   ( lcsh )
Widows -- Florida -- Gainesville   ( lcsh )
Bereavement -- Psychological aspects   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1986.
Includes bibliographical references (leaves 194-205).
Statement of Responsibility:
by Rubye Wilkerson Beck.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000924333
notis - AEN4950
oclc - 15911238
System ID:

This item is only available as the following downloads:

Full Text








I would like to thank Dr. Gordon Streib, my chairman, for all his

helpful comments and suggestions both at the beginning of this pro-

ject in 1981 and in the past year when I was writing the disser-

tation. His willingness to write down detailed comments on earlier

drafts has made the task of writing the dissertation "long-distance"

much more possible.

I would like to thank Dr. Felix Berardo for stimulating my in-

terest in widowhood through his published work and through what I

learned from him in his classes on death, dying and bereavement.

I would like to thank the other members of my committee,

Dr. Gerald Leslie, Dr. Charles Wood, and Dr. Samuel Hill, for their

contributions to this work.

Most of all, I would like to thank my husband, Scott, himself a

sociologist, who assisted me with my computer analysis, offered help-

ful comments, made numerous trips to the library while I was pregnant

to obtain materials I needed, shared fully in parenting responsibil-

ities, and was extremely supportive and encouraging throughout this

project. Without his help, the difficult task of completing a dis-

sertation out-of-state, while simultaneously holding several part-

time jobs and having and caring for two daughters, would have been



page number


ABSTRACT .. . ... ix


Importance of This Research . .. 1

Sociological Concerns in Studying Widowhood 2

Social Support and Social Interaction . 4

Culture, Norms and the Role of Religion 5

Other Issues . . 6

Summary . . 7


General Literature on Widowhood and Bereavement. 8

Prevalence of Widowhood . 8

Some Effects of Widowhood . 8

Stages of Grief . . 9

Physical Sequelae of Bereavement . .. 12

Widowhood as a Social Loss . ... 13

Economic Losses in Widowhood . 14

Practical Problems: Sex Differences ... 15

Loss of Caring Relationship . ... 16

Culture and the Experience of Bereavement .. 17

The Issue of "Death Denial" . .. 17


Works on Social Support and Social Interaction 20

Social Support and Social Interaction:
Initial Equation of the Two Concepts ...... 21

The Nature of Social Support . ... 22

Relationship Between Social Support and
Coping with Stressful Events . 24

Social Support and Subjective Well-Being .. 25

The Stress-Buffering Hypothesis ... 26

Importance of Subjective Perception of Social
Support and Social Interaction ... 28

The Issues of Reciprocity and Dependency .. 28

Possible Negative Consequences of Social Support .31

Works on Social Interaction . ... 32

Socioeconomic Status Differences in Social
Interaction . . 3

Differences in Social Interaction According to
Marital Status . . 34

Sex Differences in Social Support and Social
Interaction . . 35

Works on Religion and Widowhood ... .. ..... 38

Functions of Religion From a Sociological
Perspective . . ... 38

Religion and Aspects of Subjective Well-Being 38

Studies Specifically Concerning Religion
and Bereavement . ... 39

Social Aspects of Religion . ... 40

Religion as a Source of Meaning. ... 41

Other Issues . .... .. .... 44

Anticipatory Grief . ... 44

Bereavement Overload . ... 46

Summary of Literature Relevant to the Current Study 47

Literature on Social Support and Social Interaction 47

Definitions of Social Support . 48

Social Support and Coping . .. 48

The Stress-Buffering Hypothesis .. .49

The Subjective Dimension of Social Support
and Social Interaction . .. .49

The Issue of Dependency . ... 50

Socioeconomic Status Differences in Social
Interaction . . 50

Differences in Social Interaction According
to Marital Status . ... 51

Sex Differences in Social Support and Social
Interaction . . 51

Religion and Widowhood . .. 52

Other Issues . . 53


Introduction . .. 54

Sample . . 54

Sample Selection Process . .. 56

Data . . . 59

The Interview Schedule . .. 61

Methods of Data Analysis . .. 67

Critique of Methods in this Study and Others .. 68

Summary . . 72


Social and Demographic Characteristics .. 75

Social Supports . . 79

Attitudes and Subjective Well-Being .. 83

Social Interaction . ... 87

Religion . . ... 89

Life Events . . 94

Health . . . 96

Other Variables . . 98

Evaluation and Feedback. .. . .99

Summary. . .. .... 101


Social Support and Anxiety . 104

Social Supports and Depression. .. .108

Social Support and Life Satisfaction .. .110

Social Support and Happiness . .. .111

Perceived Lack of Social Support . .. 114

Telling One's Problems to Someone . .. 116

Turning to Someone in a Crisis. .. .118

Receiving Help with the Problems
of Widowhood . .... 121

Social Support and the Stress-Buffering Hypothesis 122

Summary .............. .... 125


Social and Demographic Correlates of Social
Interaction . . ... 130

Social Interaction and Subjective Well-Being 132

Number of Available Kin and Subjective Well-Being 132

Living Arrangement and Subjective Well-Being 133

Satisfaction with Amount of Contact with
Kin and Well-Being . .. 134

Social Activities with Others and Well-Being 137

Organizational Involvement and Well-Being 138

Church Attendance and Well-Being .. 139

Association Between Social Activity
With Others and Social Support .. 140

Kin Vs. Non-Kin: Effects of Social Interaction on
Well-Being . . ... 141

Friendship and Subjective Well-Being ... 143

Friendship and Social Activity . .. 144

Summary . . 145

CHAPTER 7: RELIGION . . .. 147

The Meaning of Suffering and Death: Relationship
to Well-Being . ... 148

Belief in an Afterlife and Well-Being ... 154

Reunion in the Afterlife and Well-Being .. .155

Believing That Good Can Come from Suffering .. 157

Strength of Religious Identification and Well-Being 158

Personal Importance of Religion and Well-Being 159

Doctrinal Orthodoxy and Subjective Well-Being 161

Denominational Orthodoxy and Well-Being .. .162

Religion as a Comfort . ... 162

Summary . . ... .164


Anticipatory Grief . .. 167

Bereavement Overload . ... 171

Sex Differences . . 173

Summary of Other Issues . ... 175


Summary of Findings: Social Support ... 177

Summary of Findings: Social Interaction .. .180

Summary of Findings: Religion . ... 182

Summary of Findings: Other Issues ... 184

Anticipatory Grief . ... 184

Bereavement Overload . .... 184

Sex Differences . . 185

Conclusions .... . 185

Critique and Suggestions for Future Research 190


RESPONDENTS . .... 206





Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Rubye Wilkerson Beck

December, 1986

Chairman: Dr. Gordon F. Streib
Major Department: Sociology

This study examines the influence of several factors on the sub-

jective well-being of a sample of 44 widows and 16 widowers inter-

viewed by the researcher in the Gainesville area during the summer of

1981. Major factors examined include social support, social inter-

action, and several aspects of religion. Secondary issues considered

are anticipatory grief, bereavement overload, and sex differences in

the experience of widowhood. Measures of the main dependent vari-

able, subjective well-being, include single-item indicators of life

satisfaction and happiness, and depression and anxiety subscales from

the Leighton Health Opinion Survey. Both quantitative and quali-

tative analyses are used.

It was found that social support had both positive and negative

effects on subjective well-being. A typology was developed which

suggests that the effects of social support depend partly on the atti-

tudes of the recipient toward social support. Fears of dependency

may reduce some of the positive effects of social support. Several

types of social interaction, including social activities with others,

were found to be positively associated with the dependent variables.

Religion, which included, for example, such issues as strength of

identification with one's religious group, personal importance of re-

ligion, and types of beliefs held, was found to have mixed effects on

subjective well-being. Some types of beliefs, for example, were

found to be helpful while others were not. It was concluded that the

influence of social support, social interaction, and religion on sub-

jective well-being can be positive, neutral, or negative, depending

on a variety of factors.


Importance of This Research

Students of aging have often focused their attention on the

losses which may accompany growing old. One of the most common

losses experienced in later life is the loss of one's spouse through

death. This research is concerned with the well-being of widowed men

and women during the second year of bereavement. The major focus is

on the influence of social support, social interaction and religion

on the subjective well-being of the widowed. This research is

primarily social-psychological in focus. There is a body of more

strictly sociological research focusing on the "family in crisis"

which has been reviewed most recently in an article by McCubbin et

al. (1980). This research is important for several reasons. Beyond

the age of 75, three out of every four women are widows and one out

of every four men is a widower. If only because of the sheer number

of people affected, widowhood is a phenomenon which demands the

attention of students of aging and family sociology. Indeed, Berardo

(1968) goes so far as to say that widowhood must be viewed as a

social problem and given the serious attention of sociologists. Yet

the numbers are only part of the story. Being bereaved of a spouse

is one of the most stressful events which can occur in an indi-

vidual's life; it is, for the most part, emotionally devastating.

Bereaved persons often experience anger, extreme sadness, anxiety,

depression and a host of other painful emotions. The bereaved may

also experience physical symptoms such as an increased risk of infec-

tion, cardiovascular disease, digestive problems, fatigue, sleep dis-

turbances and many others.

Sociological Concerns in Studying Widowhood

The social effects of widowhood are also quite dramatic. In a

couple-oriented society such as the United States, the loss of one's

spouse is often the loss of one's primary kin and primary source of

social interaction and social support. According to Leslie (1982),

marriage is the "structural keystone" of the kinship system in the

United States. Loyalties to husbands and wives, and to children, are

paramount. Relationships with other kin are much less important.

Older couples, whose children have, for the most part, left home,

have as their primary kinship unit their marriage. Consequently,

when a husband or wife dies, the surviving spouse is bereft of the

preeminent portion of his or her kin group. This may be more the

case for men than for women. According to Kessler and McLeod (1984),

women are more extensively involved in kinship networks outside the

nuclear family than are men. Therefore, when bereft of his wife, a

widower may be bereft of a greater portion, if not all, of his

significant kinship network than would be a widow bereft of her

husband. To a greater or lesser extent, widowhood disrupts the

patterns of social integration of the individual into kinship groups

and society. In addition to the actual loss of the spouse, there is

also the loss of the social role of husband or wife, the change in

status from married to widowed, and for most, a change in the way

one's time is spent. Sociologists have, ever since Durkheim,

emphasized the importance of social integration to the subjective

well-being of individuals and to the stability of society as a

whole. Therefore, anything which disrupts patterns of social

integration is of sociological concern. Widowhood presents a

situation in which the disruption of patterns of social integration

threatens to diminish the subjective well-being of individuals.

Indeed, Berardo (1968) claims that mental illness is connected to

social isolation and that the widowed are often "marooned" in an

environment which generally requires paired relationships for social

participation to occur. Yet there are alternative avenues of social

integration such as family, friends, and involvement in social

groups. One of the central foci of this research is to determine the

extent to which these alternate avenues of social integration affect

the subjective well-being of the widowed men and women in this


Social Support and Social Interaction

These avenues of social integration have been divided in this

study into the two general categories of social support and social

interaction. Social support, loosely stated, involves those things

which other people do for the widowed which may improve their lives

in some way. This includes both expressive and instrumental types of

support. The implicit assumption in examining social support in this

study is that the receipt of expressive and instrumental supports by

the widowed may be regarded as in some ways compensating for the

supports formerly provided by the spouse. A further assumption is

that those widowed persons whose emotional and practical needs are

met through these avenues of social support are better off in terms

of subjective well-being than those who do not receive any or as much

social support and whose needs may therefore not be met.

The other alternate avenue of social integration to be addressed

in this study is social interaction. An implicit assumption similar

to that made concerning social support is made concerning social

interaction. It is assumed that social interaction with family,

friends and others may operate in a compensatory way for the widowed,

to a greater or lesser extent replacing the social interaction

formerly experienced with the spouse. A further assumption is that

those who are more socially integrated, as measured by the social

interaction variables used in this study, are better off than those

who are relatively more socially isolated.


Culture, Norms and the Role of Religion

Another central focus of this research derives both from the

symbolic interaction perspective and the sociological focus on

culture and norms. Symbolic interactionists tell us that the way in

which people define a situation affects their response to it, its

meaning to them and its effects on them. The bereaved individual's

definition of the situation, in this case the situation of loss and

widowhood, is heavily influenced by the culture of which the indi-

vidual is a part. Culture provides a framework of interpretation, or

meaning, for life's events and norms for how one should think, feel

and behave when confronted with particular events and in particular

situations. Many writers have discussed both the secularization of

Western cultures and the lack of one clear view of the meaning of

death. Some have suggested that culturally, death is denied. Others

have suggested that it is dealt with not by denial, but by treating

it as solely a rational-physical phenomenon, both of which approaches

are seen by some as unhelpful to the bereaved. Berardo (1968) claims

that bereaved spouses face "normative ambiguity" concerning how to

behave when widowed. This "normative ambiguity" may elicit a great

deal of uncertainty and anxiety. Religion, it seems, is the

institution in society which deals most directly with death in all

its cognitive, emotional and spiritual aspects. Religion may provide

both a framework of meaning in which individuals may come to

understand their loss and consequent norms concerning how they should

feel, think and behave when bereaved. Religious organizations may

also serve as sources of social support and social interaction for

the bereaved. Based on these assumptions, religion variables are

used extensively in this study as a way to explore how bereaved

individuals view death, God, the afterlife and so on, and whether

their beliefs have any bearing on their subjective well-being.

Involvement in religious organizations is also viewed in terms of

social support and social interaction.

Other Issues

There are three other issues dealt with to a lesser extent in

this paper. The first is anticipatory grief. It has been hypo-

thesized by some that knowing in advance of the impending death of a

loved one allows one to grieve in advance of the death and thus

improves the adjustment after the actual death. By this reckoning,

sudden deaths, where there is no opportunity for anticipatory grief,

are harder to cope with. Others have argued that long illnesses

emotionally and physically drain the survivors and leave them less

able to cope when the actual death occurs. This issue will be

briefly explored in this paper.

Another issue involves the phenomenon termed "bereavement

overload." This involves the idea that multiple losses, especially

deaths, "overload" a person emotionally so that he or she becomes

less and less able to cope with each succeeding loss. Since this

study includes a life events scale which ascertains such losses, it

will be possible to determine if "bereavement overload" is affecting

the subjective well-being of the members of this sample.

The third issue is whether widowhood is more difficult for men or

women. Since this sample includes men, albeit a small number, this

comparison can be made to a limited extent. Such issues as economic

status, social support received, social interaction and comparative

well-being will be examined.


To recapitulate, the three major issues in this research are the

extent to which social support, social interaction and religion

influence the subjective well-being of the widowed in this study.

Secondary issues to be examined include anticipatory grief,

bereavement overload and sex differences in the experience of


Overall, what follows is a study of factors affecting the

subjective well-being of widowed men and women, based on in-person

interviews of 60 widowed persons in northern Florida. Due to the

nature of the sample, it is not expected that broad generalizations

about widowed persons will be made, but that issues will be explored

and tentative conclusions drawn.


General Literature on Widowhood and Bereavement

Prevalence of Widowhood

As of 1980, there were over ten million widows and two million

widowers in the United States (Lopata, 1980). This number is in-

creasing, but the ratio of widows to widowers has held steady at

about five to one. The proportion of the total female population of

the United States who are widowed at a given point in time is about

14%. The proportion for males is about 3%. The median age of widows

is 68; of widowers, 71 (Lopata, 1979). By the age of 75 and beyond,

75% of women are widowed, contrasted to about 25% of men in the same

age group. Clearly then, widowhood is an experience affecting a very

large proportion of the population, especially the female


Some Effects of Widowhood

The death of a husband or wife, besides being a common exper-

ience, is a devastating emotional experience, sometimes ranked

first on life events scales, but usually ranked as the second most

stressful event, with only the loss of a child ranked higher (Holmes

and Rahe, 1967). Emotional reactions to the loss of a spouse may

include anger, guilt, sadness, anxiety, depression and preoccupation

with thoughts of the lost spouse. The newly widowed person may

experience these emotions as bewildering, overwhelming, or

frightening. The fear of "going crazy" is a common one for bereaved

persons. In addition, the bereaved are more likely to experience

mental illness than are the non-bereaved (Berardo, 1968). Some

people describe grief, defined as the emotional response to the death

of a loved one, as an illness, but most agree that grief is a

"normal" and "self-limiting" experience, even though some of its

manifestations may resemble mental illness (Parkes, 1972).

Stages of Grief

An examination of the typical experience of grief for the widowed

follows. An understanding of the experience of grief may make it

clearer why the subjective well-being of widowed persons can be

viewed as tenuous and worthy of study.

Although there is a great deal of individual variation in the

nature of grieving, some general patterns have been found. Glick et

al. (1974) discuss the following as stages of grief. The first stage

usually involves a feeling of shock or numbness, or the inability to

feel anything at all, often accompanied by the inability to believe

the death has really taken place. Glick and his colleagues recount

one incident in which a woman was informed that her husband had been

killed in an automobile accident. She told herself it was an error;

it must be someone else. Even when she went to the morgue and saw

her husband's lifeless body lying on a table she refused to believe

that he was dead. She touched him, thought his skin was still warm,

and began screaming for the doctors to help him. While this is an

extreme example, it is illustrative of the power of denial and

disbelief. Generally, this first stage of grief lasts from a few

days to a few weeks, although in a few cases it may continue


The second stage is the most acute and intense phase of grieving,

sometimes referred to as Yearning and Protest. This stage involves a

psychological "search" for the lost spouse. Although intellectually

the widowed person knows the spouse is irretrievable, the emotional

connection is strong enough that it takes a while to emotionally "end

the relationship." During this period, which is generally thought to

last six months to a year, visual and auditory hallucinations of the

deceased are common, often contributing to the bereaved person's

feeling he or she is "going crazy". Protest and denial are operative

at this stage as well. The bereaved may feel angry at God, doctors,

even the deceased. There may be an obsessive review of events

leading up to the death with a tendency to worry over "if only's": if

only we'd caught the cancer sooner, if only he'd worn his seatbelt,

if only she had quit smoking, etc. Feelings of guilt are quite

common. Some people feel guilty over not having grieved enough, some

for having grieved too much, some for real or imagined shortcomings

in their treatment of the spouse prior to the death, or perhaps for

having argued with their spouse the last time they spoke.

The third stage, which some writers have merged with the second,

is sometimes termed Apathy, Aimlessness and Disorganization. A

person in this stage may be depressed and inclined to see no purpose

in life. This phase is extremely variable in length. The bereaved

may feel no enthusiasm for any activity, and the simplest of tasks

may seem overwhelming. A sense of helplessness and despair is


The fourth stage is termed Resolution, which involves the

resumption of a "normal" social life, the ability to have positive

feelings and to enjoy life to some degree. In a sense, it involves

"getting over" the feelings of grief. Glick et al. say this usually

occurs from one to two years after the death, although some people

never reach this phase; their grieving seems to be a permanent


One reason for the focus on the subjective well-being of the

widowed is, at least indirectly, to examine factors which may be

associated with "resolution" of grief. This involves assuming that

those who seem to exhibit greater degrees of subjective well-being

have resolved their grief more than those who exhibit lesser degrees

of subjective well-being. This tentative assumption is necessary in

the absence of longitudinal data with which some assessment of actual

resolution of grief might conceivably be made.

Physical Sequelae of Bereavement

In addition to the emotional reactions to bereavement, there is

evidence that there are negative physical consequences as well.

Therefore, the concept of subjective well-being in this research

includes measures of physical health. Physical responses to bereave-

ment may include shortness of breath, sighing, weakness, digestive

problems, fatigue, sleep disturbances, and tightness in the chest

(Troll, Miller, and Atchley, 1980). There is an increased suscepti-

bility to such problems as infections, cardiovascular disease, can-

cer, rheumatoid arthritis, and ulcerative colitis, among others

(Klerman and Izen, 1977). Additionally, there is a statistically

significant increase in mortality for widowed men under the age of

75. This higher mortality rate for widowers may last as long as six

years after bereavement, especially if they do not remarry. This

increased mortality results from an increased risk of death from

accidents, cardiovascular disease, and some infectious diseases. The

same does not seem to be true for widowed women (Osterweis, et al.,

1984). There is also an increase in suicide among the widowed, par-

ticularly among widowed men, which would seem to indicate that

bereavement may have devastating effects on mental health as well as

physical (Berardo, 1968; Bock and Webber, 1972).

Widowhood as a Social Loss

Many social changes and losses accompany the physical and

emotional reactions to conjugal bereavement. The first and perhaps

most obvious is the change in status from married to widowed. This

involves being seen differently by others. Glick et al. (1974) found

that many women were reluctant to use the term "widow" to describe

themselves, believing it would stigmatize them. Some still identi-

fied themselves as wives. Some widowed persons find that in a couple-

oriented society such as the United States, their new status causes

them to be left out socially by their still-married peers or, when

included, to feel like a "fifth wheel" (Lopata, 1979; Decker, 1973;

Caine, 1974). In view of the stated importance of social integration,

this would seem to be problematic. According to Leslie (1982) "what-

ever maladjustment is associated with widowhood is not a function of

widowhood and advancing age alone but is also a function of the indi-

vidual's position in a total social group" (p. 640). Widowhood

involves losing the social role of husband or wife. For those people

for whom the role of spouse has been very important, the loss of this

role may be nearly as devastating as the loss of the spouse. For

example, the woman who has devoted herself to keeping house for her

husband, being his companion, and supporting his career may be bereft

of her major role, in fact, her job. This loss may be less important

for a widower who still is employed. Less of his "life space" was

taken up with the spouse role so he has less empty "life space" to

fill when widowed (Lopata, 1979). The bereaved has lost not only his

or her role as spouse, but the roles the deceased spouse filled as

well, such as lover, co-parent, best friend, housekeeper, bread-

winner, etc. Particularly in the case of widowers, the bereaved may

have lost their main or only close "confidante" relationship. The

focus on social support in this research may be seen as a way to

determine if some of the roles filled and services provided pre-

viously by the spouse are being "replaced" with roles and services

filled and provided by others. If so, this may contribute to the

subjective well-being of the surviving spouses. According to Berardo

(1968), part of the task of the widowed is the reorganization and

reintegration of social roles to fit their new widowed status. To

the extent that this task can be accomplished, resolution of grief,

or subjective well-being, may be enhanced.

Economic Losses in Widowhood

Economic losses may also result. Lopata (1977) found that al-

most half of the widows in her study had incomes below the poverty

line. A decline in financial well-being is particularly common for

women who have been primarily homemakers. The new widow may find

herself having to obtain employment, perhaps for the first time, to

make ends meet, especially if her husband's death occurred before he

became eligible for pension benefits, if he had no pension or one

with no provision for widow's benefits, or if his life insurance

policy was inadequate. Given the combined effects of age and sex

discrimination in employment, and her possible lack of job experience

or marketable skills, this may be difficult. Even if she is old

enough to receive Social Security benefits based on her husband's

contributions, she may not have adequate income. Perhaps with the

increased participation of women in the labor force, future cohorts

of widows will be better off financially. Of course, if the widow's

husband left her with a paid-for home, a savings account, and a

generous pension, financial problems will be less severe. Even for

those widows who have sufficient money with which to live well, lack

of knowledge concerning financial planning or even complete lack of

experience in handling finances may be problematic. In some more

traditional marriages it has been customary for the husband to take

care of the couple's finances. Some widows have never, before

becoming widowed, written a check. A widow from such a marriage may

well be ill-equipped to manage her own financial affairs. Men are

less likely to suffer financially from their wives' deaths. The loss

of a wife usually does not entail the loss of the main source of

income or of pension benefits.

Practical Problems: Sex Differences

Practical problems may ensue with widowhood. Widowers may not be

well-prepared to cook for themselves, or care for their home and

clothing. They may be particularly ill-equipped to care for their

small children (Berardo, 1970). Widows may have no experience with

automobile maintenance, yard care or even changing fuses or driving a

car. The traditional division of labor among married couples

virtually guarantees that, upon widowhood, both men and women will

find themselves faced with the necessity of learning new skills in

order to cope with the tasks of daily living, or of finding someone

else to perform these tasks.

Loss of Caring Relationship

As devastating and difficult as these losses, changes and

problems are, it is probably the loss of the caring relationship that

is most difficult for the bereaved (Kalish, 1981). Absent is someone

to talk with, to sleep with, to show affection to and receive

affection from, to know just how you like your eggs cooked, to do

things with, to understand you and listen to you. Even if the

marital relationship was not seen as ideal, even if it was "conflict-

habituated," (Cuber and Harroff, 1965) its loss is painful. There is

evidence that even unhappy marriages may become idealized in the

memory of the survivors (Lopata, 1979), perhaps making adjusting to

the loss more difficult.

It has been said that death ends a life, but not a relationship.

The tie between husbands and wives is often so strong that it takes

considerable time and much painful grieving after the death for the

tie to be severed and the bereaved to "adjust" to their loss or "get

over" grieving. Of major concern in this research is the extent to

which other relationships, other social ties, may affect the adjust-

ment or subjective well-being of the bereaved.

Culture and the Experience of Bereavement

A major focus of sociological thinking has been the influence of

culture, or as Durkheim termed it,"collective representations," on

individuals. It is culture which provides individuals with norms

concerning how they should think, feel and act and with interpre-

tations for the meanings of events in their lives. It seems a log-

ical extension of this sociological idea to examine the ways in which

American culture deals with death and bereavement, as this presumably

affects the experience of bereavement in the lives of individuals.

The Issue of "Death Denial"

A frequent theme in literature concerning death and bereavement

is the "denial" of death in American society. Several writers

(Berardo, 1968; Becker, 1973; Feifel, 1977; Scheff, 1980; Jackson,

1979; Gorer, 1977) have claimed that there exists a culturally

ingrained "denial" of death which results in a paucity of norms

concerning how to grieve "properly" and how to behave toward bereaved

persons. As a result, there are strong feelings of anxiety and

discomfort in the presence of the bereaved, with a tendency to try

and avoid them as much as possible. The bereaved, being a part of

the same culture, try to conceal their grief as much as possible so

as not to make everyone else uncomfortable. It is also claimed that

this cultural "death denial" elicits a situation in which family,

friends, and clergy give support and solace to the bereaved for only

the minimal possible time, a few weeks, leaving the bereaved with few

supports just as the initial shock of loss and numbness begin to wear

off. The bereaved are thus often left to fend for themselves during

the most intense and painful phases of the grieving process. Both

Glick et al. (1974) and Gorer (1977) found the latter to be true in

their studies of recently bereaved persons. This "death denial" and

consequent lack of norms, it is claimed, have contributed to a

relative absence of formal ritual or structured supports for the


A somewhat different viewpoint is offered by Dumont and Foss

(1972) who assert that American culture simultaneously fosters

"denial" and "acceptance" of death. Their arguments for "denial" are

similar to those already discussed. They argue, in contrast, that

Americans have in some ways a very accepting attitude toward death,

for two main reasons. First, as advances in medicine and public

health measures have reduced the proportion of deaths at younger

ages, death is increasingly the province of the old. Death has come

to be viewed as the termination of a normal life cycle, as appro-

priate and acceptable, an inevitable occurrence of old age. The

second factor is the tendency toward rationality and reason in

American culture. Americans grow up learning to look at most things,

including death, in a relatively detached, rational manner. This

acceptance is manifested in the attitude that one should "be

realistic," "face up to" one's loss, and "get on with life."

Bereaved people are expected to refrain from public displays of

strong emotion after a "reasonable" time, usually considered to be a

few weeks. Often they are expected to return to work and to "normal"

life in the same short period of time (Jackson, 1979). In short,

according to this point of view, there are clear norms concerning how

to behave toward the bereaved and how to conduct oneself when

bereaved. The norms are that one should, after a short period of

allowing the bereaved to indulge their emotions, help them "keep

busy," "get their minds off it," etc. The bereaved are expected to

control their strong feelings, at least in public, after the first

few weeks and return to "normal" life. Both the "death denial" and

the "death acceptance" hypotheses lead to a similar conclusion: the

strong emotions of grief should be expressed only in a short span of

time and then the bereaved should get on with life. This may be

functional for society as a whole, minimizing the disruptions caused

by deaths. However, for the individual bereaved person, it is open

to debate whether or not this approach to death and bereavement is

helpful. One thing is clear; grief is expected to be and usually is

a private experience. Grief and other strong emotions are expected

to be expressed primarily within the family and among close friends.

It is expected, then, that adjustment to bereavement and subjective

well-being are strongly influenced by the attitudes and behaviors of

family and friends toward the bereaved. The latter is one of the

issues in this research.

Works on Social Support and Social Interaction

The sociological study of the importance of social support and

social interaction may be said to have begun with Emile Durkheim.

Durkheim argued that to understand individuals, one must begin with

society. To explain individual behavior, one had to use social

facts. Durkheim's study of suicide, especially egoistic suicide, is

a case in point. Egoistic suicide reflects a state of inadequate

solidarity in which the suicidal individual has too few socially

structured sources of support, that is, is too little involved in

society (Smelser, 1976). Durkheim's theme of psychic support

provided by social bonds is illustrated in the following quote.

There is, in short, in a cohesive and animated society a constant
interchange of ideas and feelings from all to each and each to all,
something like a mutual moral support, which instead of throwing the
individual on his own resources, leads him to share in the collective
energy and supports his own when exhausted. (Durkheim, 1897).

Durkheim believed that strengthening social bonds would increase

the subjective well-being of individuals. This theme of the impor-

tance of social bonds, or social integration, to the subjective well-

being of individuals is echoed in much of contemporary sociological

thought. Lowenthal and Haven (1968), in a much-cited article, pos-

ited the importance of intimacy for the subjective well-being and

morale of persons undergoing social losses. They argued that intimate

relationships serve as buffers in times of stress, such as widowhood

and retirement, and that maintenance of close ties with others is the

"center of existence up to the very end of life" (p.30). The authors

stated in particular that the presence of a confidante is positively

associated with measures of adjustment such as psychiatric status and

morale. Since Lowenthal and Haven' 1968 article, research and com-

mentary on social support and social interaction have dramatically


Social Support and Social Interaction: Initial Equation of the Two

The concepts of "social support network" and "social support

system" have been developed. According to a review article by Wellman

(1981), during the last decade or so, numerous studies have been con-

ducted exploring the relationships between social ties, or social net-

works, and subjective well-being, usually conceived of as some indi-

cator of physical and/or mental health. Many of these studies, fo-

cusing on the number of ties in a social network, the frequency of

contact with network members and the differential presence of kin and

friends in these networks, have shown that these ties often provide

the individual with emotional aid, information, and material re-

sources, all of which have been found to have a positive influence on

various measures of subjective well-being.

Much of this research, having found a positive relationship

between social ties, social interaction and subjective well-being,

has oversimplified the relationship between these variables by as-

suming that all social interaction is beneficial and supportive. A

great deal of the initial social support research thus focuses almost

exclusively on measures of frequency of interaction. Recent

research, however, has shown that the supportiveness of social ties

or interaction is open to question, and that in fact, some inter-

action may have negative effects (Wellman, 1981; Thoits, 1982). A

useful statement may be that social support always involves social

interaction but that social interaction does not always involve

social support.

The Nature of Social Support

Just what, then, is the nature of social support? Many def-

initions have been offered. Cobb (1976) defines social support as

information which prompts an individual to believe that he or she is

cared for and loved, esteemed and valued, and belongs to a network of

mutual obligation. This notion of mutual obligation comes up

frequently. Gottlieb (1981) delineates four types of support:

support which is emotionally sustaining, support which is problem-

solving oriented, support which involves indirect personal influence

and/or support which involves environmental action. House (1981)

posits social support as having four dimensions: emotional, involving

empathy, love and caring; appraisal, involving information relevant

to self-evaluation; information which can be used to aid coping; and

instrumental, which is tangible aid. Kahn and Antonucci (1980)

describe a trichotomy of social support: aid, which is instrumental

help; affect, which is emotional support; and affirmation, which is

acknowledgment or agreement with another's statement or act. Cohen

and McKay (1984) offer a more outcome oriented definition, seeing

social support as interpersonal processes presumed to result in

positive outcomes and to buffer stress. Walker, Macbride and Vachon

(1977) define a social support network as that set of personal

contacts which helps an individual to maintain his or her social

identity and receive emotional support, material aid, services,

information and new social contacts. Lopata (1979, p.4) defines a

social support as "an action or object which the society generally

defines as necessary or helpful in maintaining a style of life of a

category of its members." Lopata delineated and researched four

types of support: economic, service, social and emotional. Antonucci

(1985) offers the notion of a convoy of social support, a dynamic

concept of social support networks over the life course. Over the

life span, the individual develops a variety of relationships that

become the bases for the support convoy. Some relationships are

lifelong, some end through death or other exit events, and new ones

are added along the way. Support exchanges occur throughout the life

span and it is necessary to understand this history to understand

present support relationships. Antonucci explains that "for example,

an older widow may feel perfectly comfortable accepting support or

help from one member of her convoy but not another. With person A

she has a "support reserve," having nursed this friend through an

illness ten years ago, but with person B she has no such support

reserve and therefore feels uncomfortable accepting help" (p.97).

The support convoy notion seems particularly useful for this study,

in which respondents are at least middle-aged and family rela-

tionships are considered.

Relationship Between Social Support and Coping with
Stressful Events

Keeping in mind that the definition of social support varies

considerably from one study to the next, some research pertaining to

the relationship between social support and coping with stress,

especially the stress of bereavement, will now be reviewed. Several

well-respected researchers (Maddison and Raphael, 1975; Lopata, 1979;

Gorer, 1977; Maddison and Walker, 1967) have found that a widowed per-

son's perception of the supportiveness of his or her social network

during the first year of bereavement is important in determining the

subsequent outcome of the bereavement crisis. Maddison and Raphael

(1975) go so far as to claim that the lack of such a supportive

network during this period may contribute to psychological problems.

Glick et al. (1974), based on their study of widowed persons during

the first year of bereavement, assert that widowed individuals need

support during the first year in order to effectively work through

their grief. Berardo (1985), in discussing the potential for

survival after widowhood, claims that the maintenance of meaningful

human contacts may lead to increased longevity and subjective well-

being, especially among widowed men. A network of intimate social

ties is life-preserving.

Social Support and Subjective Well-Being

Turner (1981) examined the relationship between social support

and psychological well-being using secondary analysis of four studies

involving diverse populations. The use of a large and diverse sample

contributes to the value of Turner's work. He found a modest but

reliable positive association between social support and well-being

for all four populations. Turner claims that while social support

has significant main effects on well-being, it is most important in

stressful circumstances. Turner also suggests that the relationship

between social support and psychological well-being may well be recip-

rocal. That is, it is probable that one's psychological condition

will affect the availability of social support as well as social sup-

port affecting one's psychological condition. Turner also found

social class differences. For middle-class and upper-class

respondents, social support was important for psychological well-

being at all stress levels whereas for lower-class respondents,

social support affected psychological well-being only at high levels

of stress. He suggests that variations in socialization by social

class may account for this result. Turner speculates that middle and

upper-class persons may grow up in a situation where the elements of

social support such as being loved, esteemed and able to count on

others are routine but that this is less often the case for lower-

class persons. One other finding of interest is that social support

may be of greatest significance for those persons suffering from

depression. This latter finding has relevance for this research;

depression is one component of the measures of subjective well-being

used in this research.

The Stress-Buffering Hypothesis

The situation in which social support operates mostly as a

moderator between high stress, or negative life events, and sub-

jective well-being has been termed the stress-buffering hypothesis.

Turner's research suggests this formulation. Individuals with a

strong social support system will, according to this hypothesis, be

better able to cope with major life stresses and, conversely, those

with little or no support will experience more negative effects from

these major stresses (Thoits, 1982). This hypothesis is not without

its critics, however. Thoits suggests that while there is much sup-

port for this hypothesis, a great deal of it is based upon research

which fails to examine the main effects of social support upon stress

and confounds the direct effects of life events on social support

with the buffering effect of events on support, mostly because cer-

tain life events such as widowhood, divorce or moving directly reduce

available sources of social support. Many studies, according to

Thoits, also suffer from inadequate operationalization of social sup-

port. Thoits suggests it is important to assess the perceived amount

and adequacy of aid received. The subjective dimension of social

support may be as important as the objective dimension. The subjec-

tive dimension of social support is explored to a limited degree in

this research. The research on the stress-buffering hypothesis is

equivocal. In attempting to refute the stress-buffering hypothesis,

Norris and Murrell (1984) found that resources, which included social

support, did not buffer the effects of negative life events on de-

pression. Resources did, however, reduce the likelihood of exper-

iencing global stress and the chances of stress leading to hope-

lessness and despair. Levitt, Clark, Rotton and Finley (1983) found,

in their study of elderly residents of a residential area targeted

for massive redevelopment, that individuals having at least one close

support figure coped better with the stress associated with being

forced to relocate than those lacking a close personal relationship.

Mitchell and Moos (1984) examined the relationship between social

support and stress among patients being treated for depression. They

found that reported increases in stress levels actually led to de-

creases in reported supports received. The authors suggest that per-

sons experiencing long-term psychological distress may alienate

sources of support and that the persistence of such long-term

stresses may erode the individual's will to maintain social ties and

the capacity of the social network to help. However, they did find

that having more close friends reduced subjectively experienced

strain. Some of the reported decrease in supports could be sub-

jective. Sarason et al. (1983) found, in a study of college stu-

dents, that negative events were associated with decreased satis-

faction with perceived support but that they were not related to

actual amount of support received. Barrera (1981) found in a study

of pregnant adolescents that negative life events actually led to an

increase in receipt of helping behaviors but a decrease in

satisfaction with that help. So it seems that there is a difference

between actual supports received and perception of their adequacy.

Harel and Deimling (1984) echo this idea. They found that although

social interaction was not associated with measures of mental health,

the overall rating of social resources and perceived adequacy of

social resources were fairly strongly associated with most measures

of mental health. The stress-buffering literature, still fairly new,

is rife with disagreements. One consistent idea, however, is that

subjective interpretation of supports received is important in

determining outcome.

Importance of Subjective Perception of Social Support and Social

Jeanne Gibbs (1985), in a relatively small-scale study of older

widows, found that the subjective well-being of her subjects depended

more on the perceived quality of relationships with family members

than the quantity of interaction with them. Ward, Sherman and LaGory

(1984) echo this idea. In their research, subjective assessments of

the adequacy of support networks were more important in predicting

the subjective well-being of elderly respondents than were more ob-

jective measures of support and interaction.

The Issues of Reciprocity and Dependency

Perception of social support may involve more than just its

amount or adequacy. Perception of reciprocity, or of whether what

one receives is appropriate, is also important. Stoller (1985) found

that reciprocity in social support is important for morale. She

asserted that the inability to reciprocate for aid received, rather

than the need for assistance, has a greater negative effect on

morale. She claims this is more operative in relationships with

friends than with kin. Stoller speculates that with family,

reciprocity can be more general, including aid given in the past or

anticipated in the future whereas reciprocity with friends is felt to

be needed in the immediacy of the present. This is similar to

Antonucci's notion of support convoys. For the elderly persons in

Stoller's study, receiving more aid than they gave led to feelings of

dependency, especially when the aid was received from children.

Dependency on children runs counter to expectations for the parent-

child relationship in many segments of American culture. Avoiding

dependency is a theme which arose spontaneously fairly often in the

course of the interviews for the current study, indicating the

salience of this issue to the respondents. Opinions as to what con-

stitutes dependency may vary from one generation to the next.

Brody, Johnsen and Fulcomer (1984) assessed what a sample of 403

women representing three generations believed adult children should

do for their elderly parents. The older and middle-aged women pre-

ferred children as confidantes and counselors but said they would pre-

fer not to live with their children, receive financial aid from them

or receive most instrumental services from them. The youngest genera-

tion of women expressed a broader range of filial obligation than the

older generations, including subscribing to the belief that adult

children should offer their elderly parents a home, financial assis-

tance, and instrumental aid. It seems that there is a generational

difference in views of what types of social support it is desirable

to provide to one's elderly parents. One reason for this difference

is suggested in the work of Streib et al. (1984) on shared living for

the elderly. This generational difference of opinion may be partly

due to the influence of the mass media. Although it has been well-

established in the scientific literature that most older persons

prefer relationships based on "intimacy at a distance" with their

children and grandchildren, living apart from them, supporting

themselves financially, but maintaining close emotional ties, the

media do not communicate this. The media often romanticize three-

generation households and strongly imply that perspns who would

hesitate to open their homes to their elderly parents or other

relatives are bad or selfish. The elderly themselves are often pre-

sented as lonely, sick and poverty-stricken, a picture which does not

represent the majority of elderly persons but a minority. If this

discrepancy exists in particular families, it may influence the

receiver's perception of the adequacy or desirability of support

received. This apparent confusion over the appropriateness of certain

types of social support may have deleterious consequences for

bereaved persons.

Possible Negative Consequences of Social Support

Glick et al. (1974) suggest that too much help, or the wrong

kinds of help, offered to the widowed may increase their feelings of

dependency and cause them to lose confidence in themselves. On the

other hand, too little help, or preferred advice to "keep busy" may

hamper adjustment. Many of their respondents reported a conflicting

tendency between a desire for the help and support of others together

with a fear of dependency. Physicians were often sought out for sup-

port because it was clear to the bereaved that concrete help in the

form of medication could be obtained. It was seen as appropriate to

turn to physicians and not indicative of unwarranted dependency. It

may be the case that defining a problem as physical or medical is

more socially acceptable than acknowledging that one has emotional or

psychological problems or needs. The need for support in bereave-

ment was also complicated by the fact that most of the widows in

Glick's study regarded grief as a private experience and that to

attempt to alleviate it by sharing it with others was to burden them

inappropriately. On the other hand, about a third of their sample

felt that "everyone was helpful at first but they don't have time any

more" (p.186), in other words, that their family and friends had done

too little for them. Others felt oppressed by too much help which

they viewed as interference. In fact, Glick et al. view one sign of

"getting over" grief as coming to value independence more and the ad-

vice and help of kin and friends less. They assert that the recov-

ering bereaved begin to develop new respect for themselves as they

recognize their ability to cope with their new situation. Bankoff

(1983) found in her study of widowed women that the relationship be-

tween support and subjective well-being varied according to the stage

of the grief process. In the early stages of grief, support from fam-

ily has the strongest positive effect on subjective well-being.

Later in the grief process it is support from friends that helps the

most. It is possible that early in the grief process the bereaved

need to be heavily dependent on those with whom they have strong ties

of affection and obligation. Later, this much dependency can be

oppressive and support from friends, or equals, where more reci-

procity is involved, may be most helpful. One might conclude from

these studies that the relationship between grief, support and sub-

jective well-being is a complex one, involving an interplay between

the need for support from others and the need for a feeling of inde-

pendence and competence. This theme of avoiding undesired depen-

dency versus the need for support is of great importance in the cur-

rent study. The degree to which social support is helpful to the

widowed is complicated by the issue of dependency.

Works on Social Interaction

This section, which focuses on the relationship between social

interaction, social networks and subjective well-being, overlaps

somewhat with the previous one on social support, since much research

has considered social support, social interaction and social networks

as interchangeable concepts.

Socioeconomic Status Differences in Social Interaction

Much research has concentrated on differences in social inter-

action according to socioeconomic status variables. According to

Antonucci (1985) people from higher socioeconomic status groups have

larger networks consisting of both family and friends, whereas people

from lower socioeconomic status groups have smaller networks consis-

ting primarily of family members. Babchuk (1979) found that higher

education and more prestigious occupations are associated with less

isolation and with larger networks. Kernodle and Kernodle (1979)

reported that low socioeconomic status elderly have fewer contacts

with both family and friends. Parks and Pilisuk (1981) found that

those with lower education and income have relatively more contact

with professional aid providers (eg. clergy, physicians, social wor-

kers) than people with higher education and income who have more con-

tact with family and friends. Heltsley and Powers (1975) found that

higher socioeconomic status people have a more positive perception of

their contact with kin and friends, although there was no objective

difference in amount of contact. Harvey and Bahr (1974), however, sug-

gest that the often reported lower levels of social interaction and

morale of widowed persons may in fact be attributable to their lower

incomes. Lopata (1979) found that more highly educated widows scored

higher on measures of social integration than less educated widows.

This body of research is important for the current study since

socio-economic status, especially education, for this sample as a

whole is relatively high.

Differences in Social Interaction According to Marital Status

A few studies have found differences in social interaction

according to marital status. Anderson (1984), in a comparison of

married and widowed women, found that widowed women were more likely

to rely on distant kin than were married women. The widows in her

study reported more primary ties with all kin, especially siblings

and extended kin, after the spouse died. This seems to lend support

to the notion that social interaction with those other than the

spouse is in some way taking the place of, or compensating for, inter-

action formerly shared with the spouse. One focus of this research

is to determine if such interaction with others makes any difference

in the subjective well-being of widowed persons. The widows in Ander-

son's study were more likely to turn to their children for emotional

support when worried, but to their siblings for instrumental help

when ill or in need of money. Morgan (1984), using data from the

Longitudinal Retirement History study, found that the average fre-

quency of interaction with kin increases for both married and widowed

persons over time, with the greatest increase among women who become

widowed. The total number of social network members decreases over

time, however, for both widows and widowers. The widows in Morgan's

study, however, had a higher visiting rate per relative than either

widowers or married men and women. Petrowsky (1976), in a study of

273 elderly Floridians, found that widowed persons were no more iso-

lated from kin and friends than were married people. Mutran and

Skinner (1981) assert that differences in amount of contact may be

based on need. Those in greatest need-are more likely to interact

with and receive help from others. This idea may help to explain why

the other studies cited found more interaction for widowed persons.

They would presumably be more in need of interaction than would mar-

ried persons. It seems plausible that a person feeling a "need" for

interaction might, along with that need, feel a fear of becoming too

dependent on others. Interaction based on dependency may be seen as

a problem. Arling (1976), in a study of 409 widows ages 65 to 85,

found that interaction with friends made more positive difference in

morale than interaction with kin. He proposed that relationships

with kin are based on role reversal and dependency and that rela-

tionships with friends are based on reciprocity and mutuality of in-

terests and values. It is questionable whether or not relationships

with kin are usually based on dependency but, as mentioned earlier in

the discussion of social support, the fear of too much dependency may

be a factor in such relationships.

Sex Differences in Social Support and Social Interaction

Several studies have noted differences between men and women in

social interaction and receipt of social support. Antonucci (1985)

states that women have larger and more multifaceted social networks

than men and that men turn primarily to their wives for support

whereas women turn to husbands, children, other family members and

friends as well. Antonucci argues that this and the better inter-

personal skills of women give them the advantage in maintaining and

enhancing their support Convoys. Men are more limited, especially

widowed men. Longino and Lipman (1981), echoing this, found that

unmarried men, including widowers, had the smallest support networks

of all marital statuses. Nye and Berardo (1973) state that widowers

tend to be more socially isolated than widows, being the least likely

category of persons, classified by marital status, to have a high

degree of kin interaction, to be satisfied with family relationships,

to have friends, and to give aid to or receive aid from their chil-

dren. Lopata (1979) has found that widows are generally involved in

extensive support networks and that children, especially daughters,

are central in them. Seelbach (1977) found that women expected much

more personal care from children than did men, and that usually their

expectations were met. Apparently widowers both expect and receive


One aspect of women being involved in larger networks than men

may, however, have negative consequences. According to Kessler and

McLeod (1984), it is not that women actually know more people than do

men, but that "women consider more people important than men do" (p.

628). This leaves women more vulnerable to distress when someone in

their network experiences a negative life event. So, while women may

have access to larger networks from which to obtain support when in

need, they may have to pay the "emotional cost of caring," when

members of their network are in need.

Widowers may feel more isolated than widows since there is a much

smaller number of fellow widowers with whom to interact. Widows in

many cases are able to become part of a "society of widows." On the

other hand, widowed men are about five times as likely to be able to

remarry than are widowed women (Cleveland and Gianturco, 1976).

Thus, in a sense, men are more likely to "replace" the missing source

of interaction and support. Chappell (1983) claims that it is not the

quantity of interaction which improves the quality of life, but hav-

ing a confidante relationship with at least one person. Another

reason for the higher remarriage rates for widowed men may be that

men are more likely than women to have had only one confidante rela-

tionship, with the spouse. So, in this way, men may "need" to re-

marry more than women. Bell (1963) claimed that marriage is more

important to women than men so that women are more devastated by be-

coming widowed. This is similar to Lopata's claim (1979) that the

wife role takes up more of a woman's "life space" than the husband

role does for men, thus making the loss of the role more devastating

to women. It seems reasonable at this point to surmise only that the

experience of bereavement, supports received and social interaction

are different for widowed men and widowed women. This study will at-

tempt to add to current knowledge concerning these differences.

Works on Religion and Widowhood

Functions of Religion from a Sociological Perspective

In delineating the functions of religion, O'Dea (1966) includes

the idea that religion can provide individuals and groups with a

sense of the meaningfulness of life which may help them find support

and consolation for the difficult experiences of life. Religion puts

the many sorrows and tragedies of life in an understandable context;

it "explains" them. In addition, religion provides a relationship to

the transcendental realm which can lead to a feeling of security and

worth. Having a reference point beyond the present life can give

people the belief that there is an order to life beyond what appears

to them in the present. In particular, belief in an afterlife may

ameliorate the sense of meaninglessness or unfairness of losing loved

ones through death or of facing one's own death.

Religion and Aspects of Subjective Well-Being

Some studies have focused on the relationship between religious-

ness and fear of death. It has been found that fear of death was

least strong among both the deeply religious and the deeply irre-

ligious. People who were uncertain in their religious beliefs showed

the greatest anxiety and fear (Kalish, 1981). Kalish claims that it

is those people with strong inner beliefs who probably find more

meaning in life, regardless of whether or not those beliefs are


There is some evidence that certain religious beliefs may be life-

preserving. Stack (1983), in a study of suicide rates in 25 nations,

found that a high level of commitment to a few life-preserving reli-

gious beliefs, values and practices lowered suicide levels. Examples

of life-preserving beliefs include a focus on an afterlife, belief in

a responsive God, and an eternal perspective. Examples of the latter

would be seeing earthly suffering as a mere short moment in the con-

text of eternity, or a belief in role models for suffering from reli-

gious literature which give suffering a positive meaning, such as the

book of Job in the Old Testament.

A few researchers have found an association between certain as-

pects of religion and life satisfaction. Hunsberger (1985) found a

moderate positive relationship between degree of religiosity and life

satisfaction in a study of 85 elderly persons. Gray and Moberg

(1977) claim that both religious behavior and belief are causal fac-

tors in the life satisfaction of older people. One focus of this re-

search is whether or not religiosity and particular religious be-

liefs contribute to life satisfaction.

Studies Specifically Concerning Religion and Bereavement

Certain religious beliefs and types of participation may facil-

itate resolution of grief. Bryer (1979), discussing her observa-

tional study of the Amish way of death, notes that the Amish view

death as the "final rite of passage into a new and better life"

(p.257), and are socialized throughout life into a calm acceptance of

death. They also have specific rituals and customs concerning death

and bereavement. Certain families are assigned to the newly bereaved

for at least one year to offer comfort and support. She saw no

evidence of protracted or abnormal grief among this group. Amish

religion provides both a "source of meaning" and a "sense of

belonging" (Wuthnow et al., 1980), both important aspects of the ways

in which religion may address the needs of grieving persons. Glick

et al. (1974) suggested that religious beliefs sustain morale and are

a major source of comfort to the newly bereaved. Harvey and Bahr

(1980) found that for the widows in their study, perception of self

as very religious and participating in church activities both had

positive effects on perceived quality of life and overall happiness.

Social Aspects of Religion

Research has shown that religious organizations serve as social

support systems for many of their members (Glock et al., 1967;

Hadaway and Roof, 1978). Additionally, religious participation may

bring the bereaved into contact with others who may reinforce his or

her beliefs and provide normative guidelines for "appropriate"

grieving behavior. Participation in religious organizations may also

contribute to the subjective well-being of widowed individuals by

increasing their degree of social participation and social


Religion as a Source of Meaning

Several authors have discussed the provision of meaning as one of

the important functions of religion for the bereaved. Religious

faith may enable the widowed person to explain and account for his or

her loss, to give it meaning (Simpson, 1979; Jackson, 1979). Frankly

(1963) has stated that the ability to find meaning in one's life is

the most important task in attaining subjective well-being. This

task is made more difficult for persons experiencing the pain of

grief, a situation which may render some previously acceptable

structures of meaning irrelevant or impossible. For Frankl "to live

is to suffer, to survive is to find meaning in the suffering. If

there is a purpose in life at all, there must be a purpose in suf-

fering and in dying" (p.xi). Frankly writes from the perspective of a

psychologist who himself was a prisoner in a Nazi concentration

camp. He observed that fellow prisoners who were able to find some

source of meaning in their tortured lives were more likely to survive

than those who "gave up," who found no meaning in their lives at that

point or had no hope for the future. The situation of concentration

camp prisoners, while different from that of widowed persons, shares

the characteristics of extreme loss and the potential for hopeless-

ness. This seems particularly relevant for the current study since

one issue is whether or not attributing some positive or neutral

meaning to suffering and death is associated with the subjective well-

being of the widowed.

Wuthnow et al. (1980), in formulating a conceptual framework for

viewing the relationship between religion and bereavement, discusses

the effects of specific types of religious beliefs on the subjective

well-being of bereaved persons. These include the ideas that be-

lieving that God can bring good from evil (in this case suffering and

death) will be associated with subjective well-being, and believing

that something good for the self (such as greater understanding of

God's love) can come from bereavement, will be positively associated

with subjective well-being and that fatalistic conceptions of death

will be negatively associated with subjective well-being. Wuthnow

and associates point out that the relationships between religion and

bereavement are likely to vary by religious tradition. Some re-

ligious traditions more than others strongly adhere to beliefs which

may contribute to the subjective well-being of the bereaved than

others. Religious groups may also vary in the kinds of social sup-

port provided to members. In other words, the relationship between

religion and subjective well-being varies according to religious

group as well as individual religious characteristics such as

strength of commitment to religious beliefs and degree of involvement

in the church organization. This article stimulated some of the

thinking involved in this research, such as assessing whether or not

there are associations between such variables as degree of religious

orthodoxy, religiosity, belief in an afterlife and subjective well-


According to Batson and Ventis (1982, p.247) "how one is re-

ligious does indeed have a dramatic effect on the relationship

between religion and mental health." In other words, the con-

sequences of religion for mental health, or subjective well-being,

will vary according to the way in which a person is religious.

Batson and Ventis, for example, discuss three types of religion,

which may each have quite different consequences for well-being.

Religion may be a "means" to something else such as gaining respect

in the community, making friends, or improving one's business

contacts. This type of religion, or religiosity, has also been

termed "extrinsic." It probably does not offer the adherent a set

framework of meaning which would provide comfort in bereavement. It

may, however, provide the bereaved with a ready-made network of

caring others who will provide social support and social

interaction. The second type is religion as an end in itself or

"intrinsic" religiosity. Intrinsically religious people view their

faith as the central, integrating value in their lives. They often

strongly adhere to an orthodox and absolute set of beliefs which

provide them with certainly concerning spiritual matters. It seems

that it may be these people for whom religion is most likely to offer

the assurance that, for example, their loved ones are in heaven and

that they will be able to be reunited someday. The intrinsically

religious may be most likely to have a firmness of faith which,

although perhaps challenged by the death of the spouse, will remain

essentially intact. According to Batson and Ventis, however, this

category of the religious may also include those who are fanatical

and inflexible, which may be viewed as mentally unhealthy, or as

detracting from well-being. The third type of religion is termed the

"quest." Those whose religiosity falls into this category place a

high value on thinking about and questioning religious values.

Complexity, doubt, tentativeness, flexibility, and tolerance of the

beliefs of others often characterize their religious views. The

"quest" type of faith may not offer the same type of certainty and

comfort to the bereaved as the intrinsic, but it seems possible that

those who fit into this category might exhibit a self-directedness

and self-reliance which would be a source of strength to them in

bereavement. On the other hand, uncertainty may be a source of pain

and worry for the bereaved.

In sum, there seem to be two aspects of religion which may

contribute to the subjective well-being of the widowed: a belief

system which gives meaning to their losses and the availability of an

organization containing like-minded others who may enhance the degree

of social integration of the bereaved. Different types of religiosity

may provide these in varying degrees or in some cases, perhaps not at


Other Issues

Anticipatory Grief

Another issue to be explored in this research is whether or not

"anticipatory grief" is associated with the subjective well-being of

the widowed. The term "anticipatory grief" is used to mean any grief

occurring prior to a loss, as distinguished from that which occurs at

or after the loss (Aldrich, 1974). It has been suggested that antici-

pated deaths are less traumatic for the survivors because there is an

opportunity to work through some of the grief in advance of the

death. Parkes (1972) found in his study of young widows that those

whose spouses died suddenly fared worse, in terms of eventual resolu-

tion of grief, than those who had known of the impending death in

advance. Hendin (1973), in a review chapter, claims that grief reac-

tions are more normal when death has been expected. The work of

mourning is done more quickly because a certain amount of this work

has been accomplished prior to the death. Glick et al. (1974) found

that expecting death in advance did not reduce the intensity of

grief, but that there was a better eventual adjustment. Sudden death

can be a shock that overwhelms the individual's ability to cope and

makes the death more difficult to accept (Ward, 1984).

Studies by Clayton et al. (1980) and by Fulton and Gottesman

(1980) did not find support for the helpfulness of anticipatory

grief. Gerber et al. (1975) found a negative relationship between

length of illness and outcome. The longer the illness, the more

difficult the grieving period was for the survivor. This could be

because an extended period of terminal illness is quite stressful for

the spouse, who is usually the main caretaker of the dying spouse.

The caretaking spouse's emotional resources may be severely depleted

by the time the spouse dies, leaving the survivor less able to cope

with the actual loss.

Thus, there is a controversy over whether or not "anticipatory

grief" helps or hinders adjustment to bereavement. This issue will

be addressed in the current research.

Bereavement Overload

Another important issue to be explored in this study is the

concept of "bereavement overload" (Kastenbaum, 1969). This concept

refers to a situation in which someone, usually an older person, suf-

fers several losses of significant others through death in a relative-

ly short time. Before the person is able to recover from one major

loss, he or she is confronted with another. This may lead to a sit-

uation in which subsequent deaths are much more difficult to work

through because previous losses have not been adequately mourned

(Kalish, 1981). As far as this author is aware, there is little empir-

ical research on bereavement overload. One study mentioned earlier,

by Kessler and McLeod (1984), may, however, suggest a possible sex

difference in propensity to experience "bereavement overload." They

contend that women care more about more people than do men. It would

seem, then, that women may be more affected by the deaths of a wider

group of people than men. Bereavement overload is a factor which will

be taken into account in trying to explain degree of adjustment to

widowhood in the current study. It seems logical to include this vari-

able when studying a group of middle-aged and older widowed persons

who are presumably more "at risk" for bereavement overload than the

population as a whole.

Summary of Literature Relevant to the Current Study

The prevalence of widowhood and the degree to which it is a devas-

tating emotional, physical, economic, and social experience clearly

mark widowhood as a subject worthy of attention. The current study

focuses on the subjective well-being of the widowed in the second

year of bereavement, focusing on the major sociological areas of so-

cial support, social interaction, and religion and the secondary

areas of sex differences in the experience of widowhood, anticipatory

grief, and bereavement overload.

Literature on Social Support and Social Interaction

Ever since Durkheim, it has been recognized that the social in-

tegration of the individual into society is important to the subjec-

tive well-being of the individual. The situation of widowhood pre-

sents a unique opportunity to study the social integration of those

whose major source of social integration, marriage, has been disrupt-

ed by the death of the spouse. In an oft-quoted article in 1968,

Lowenthal and Haven claimed that intimate social ties were of crucial

importance to the subjective well-being of persons undergoing social

losses. Since then, the literature on what are now called social sup-

port and social interaction has grown dramatically. Barry Wellman

(1981) presents a very useful summary of much of this literature.

His summary leads to the conclusion that social interaction often

provides social support which, in turn, has a positive effect on

subjective well-being. As Thoits (1982) and Wellman himself point

out, much research literature has assumed that social interaction and

social support are one and the same. Thoits and Wellman claim that

they are not and that some social interaction may actually be neg-

ative or harmful.

Definitions of social support

Numerous authors (Cobb, 1976; Gottlieb, 1981; House, 1981; Kahn

and Antonucci, 1980; Cohen and McKay, 1984; Walker, Macbride, and

Vachon, 1977; Lopata, 1979; and Antonucci, 1985) have defined social

support in useful but different ways. The definition which comes

closest to approximating that used in this study is the first two

parts of Kahn and Antonucci's (1980) tripartite definition of social

support as aid (or instrumental help) and affect (or emotional

support). Antonucci's (1985) concept of a convoy of social support,

or social support based on relationships built over a lifetime, is

especially useful in this study in helping to understand why a

widowed person may be comfortable accepting social support from one

person but not another, based on his or her past relationship and

past support exchanges.

Social support and coping

A number of researchers have found that social support is helpful

in coping with stress, especially the stress of bereavement. The

works of Maddison and Raphael (1975), Glick et al. (1974), Lopata

(1979), Gorer (1977) and Maddison and Walker (1967) have all pointed

out that social support is important in determining the outcome of

bereavement or the subjective well-being of the widowed person.

The stress-buffering hypothesis

Turner (1981) and others have said that the effects of social

support on subjective well-being may be strongest in stressful

situations, such as widowhood. This has been termed the "stress-

buffering hypothesis" and is one issue of interest in this study.

Others (Thoits, 1982; Norris and Murrell, 1984; and Mitchell and

Moos, 1984) have challenged the "stress-buffering hypothesis" on

various grounds. It is hoped that this study will add to the current

state of knowledge in this debate.

The subjective dimension of social support and social interaction

The subjective dimension of both social support and social inter-

action are important in this study. Several researchers (Thoits,

1982; Sarason, 1983; Barrera, 1981; Harel and Deimling, 1984; Gibbs,

1985; Ward,Sherman, and LaGory, 1984) have found that the indi-

vidual's perception of and feelings about social support and social

interaction have greater effects on subjective well-being than the

actual support received or interaction experienced. This is an impor-

tant idea for this study. Although objective measures of social

support and social interaction are used, indicators of the subjective

dimension are also used and considered to be important.

The issue of dependency

Another important issue which arose during the course of conduct-

ing the interviews for this study is dependency. Dependency, espec-

ially dependency on one's adult children, runs counter to expec-

tations in many sectors of American culture. Several researchers

have found that issues of dependency may complicate or even negate

the positive effects of receiving social support (Stoller, 1985;

Glick et al., 1974). Others have pointed out that people disagree as

to what constitutes appropriate and acceptable social support and at

what stage in the grief process particular types and amounts of sup-

port are desirable (Bankoff, 1983; Brody, Johnsen, and Fulcomer,

1984). People's feelings about the appropriateness of support

received or not received, and particularly issues of dependency, are

explored in this research.

Socioeconomic status differences in social interaction

Amount and types of social interaction have been found to vary by

socioeconomic status (Antonucci, 1985; Babchuk, 1979; Kernodle and

Kernodle, 1979; Heltsley and Powers, 1975; Harvey and Bahr, 1974;

Lopata, 1979). This is relevant to the current study since

socioeconomic status in this sample, especially education, is

probably higher than for a comparable national sample.

Differences in social interaction according to marital status

Some researchers have found differences in social interaction

according to marital status, most finding that widowed persons are no

more isolated than married persons and may in some cases be involved

in more social interaction than their married cohorts (Anderson,

1984; Morgan, 1984; Petrowsky, 1976). There is some indication that

widowed persons, especially widowed women, may expand their social

interaction to include a wider network of persons than that in which

they were previously involved when married. This seems to lend

support to the notion that social interaction after widowhood may be

to some degree compensating for the loss of social interaction with

the spouse.

Sex differences in social support and social interaction

Most of the literature on sex differences between widowed men and

widowed women implies that widowers are more socially isolated than

widows (Antonucci, 1985; Lopata, 1979; Longino and Lipman, 1981).

Men are, however, more likely than women to remarry (Cleveland and

Gianturco, 1976). It may be that men, because of their greater

social isolation, may "need" to remarry more than do women. Obvi-

ously, the greater availability of potential marriage partners for

men is the major reason for their higher rates of remarriage, but the

"need" factor may enter in as well. This study will compare the

social interaction and social support of the men and women in this


Religion and Widowhood

Religion is important in this study in two ways, first as an

institution which provides individuals with a framework of meaning

for death and bereavement and, second, as a source of social support

and social interaction.

There is some indication that possessing a religiously derived

sense of the meaningfulness of death and bereavement may contribute

to the subjective well-being of the widowed (Stack, 1983; Glick et

al., 1974). There is also evidence that simply perceiving oneself as

religious is associated with better morale and life satisfaction

(Hunsberger, 1985; Gray and Moberg, 1977; Harvey and Bahr, 1980;

Glick et al., 1974). Being a part of a religious group which provides

clear norms concerning how to grieve and which provides members with

comfort and support in their grief may contribute to subjective well-

being (Bryer, 1979). In an extremely helpful theoretical article,

Wuthnow et al. (1980) point out that religious traditions vary in the

extent to which they provide beliefs, norms and social support which

may enhance the subjective well-being of the bereaved. Certain

beliefs, for example the belief that God can bring good from evil,

may be especially helpful to newly widowed persons. A focus on the

contributions of particular types of religious beliefs and degrees of

religious involvement to the subjective well-being of the members of

this sample is another component of this research.

Other Issues

Two other issues to be explored in a more limited way in this

study are anticipatory grief and bereavement overload. Some studies

have found that a death which has been expected, or anticipated, al-

lows for some "grief work" to be completed in advance of the death

and thus results in better and quicker resolution of grief after the

death occurs (Parkes, 1972; Glick et al., 1974; Ward, 1984). Other

studies find, by contrast, that the opportunity for anticipatory

grief does not seem to be particularly helpful (Clayton et al., 1980;

Fulton and Gottesman, 1980; Gerber et al., 1975). The issue of the

contributions of anticipatory grief to the subjective well-being of

the widowed is explored briefly in this research.

Although virtually absent from empirical studies of grief and

bereavement, the concept of "bereavement overload" is explored in

some of the theoretically oriented literature (Kastenbaum, 1969;

Kalish, 1981) and will be examined in the context of this study.



This research is an attempt to glean information on a wide vari-

ety of characteristics and experiences which may be associated with

the subjective well-being of widowed individuals. One of the best

methods available with which one may obtain such information is the

structured interview. Thus, this was the method of data collection

chosen for this study. Sixteen widowers and forty-four widows were

interviewed by the researcher using a combination of forced-choice

and open-ended questions. The interview included questions on socio-

economic status characteristics, social interaction, social support,

health and subjective well-being, and religion.

This chapter includes a description of the sample and the sample

selection process, the process of obtaining the interviews, a descrip-

tion of the content of the interview schedule and methods of data



The sample consists of 16 widowed men and 44 widowed women re-

siding in the Gainesville area at the time of the interviews in the

summer of 1981. The sample was obtained from newspaper obituary

listings of surviving spouses. It was felt that this method of

sample selection would result in a more unbiased sample than going

through organizations such as senior citizen centers or churches.

Obituary listings dating from 24 months to 13 months prior to the

interviewing period were used. Persons who had been widowed for 13

to 24 months were selected for 3 reasons. First, since Glick et al.

(1974) found that the most intense phase of grieving lasts about a

year, it was thought desirable to interview people who had had enough

time to work through some of the worst of their grief. (There is, of

course, controversy on the issue of the duration of grief, some

saying one never completely "gets over it.") Second, it was believed

to be important to avoid interviewing people exactly one year after

the death of their spouse, due to the "anniversary reaction." The

"anniversary reaction" entails the resurgence of intense feelings of

grief on or near the anniversary of the death (Maddison and Raphael,

1975). Interviewing someone at that time might have unduly upset him

or her as well as presenting a negatively biased picture of his or

her situation. One interview happened to occur on the birthday of

the deceased spouse, which was unavoidable since that information was

not available to the interviewer in advance. The "anniversary

reaction" also may occur on birthdays, wedding anniversaries,

Christmas and any other day significant to the surviving spouse.

Third, it is theoretically important to this work to obtain a picture

of what life is like for widowed persons in the long-term, at least

beyond the first few months when, it has been claimed, supports are

most abundant and the full force of grieving is ameliorated by shock

and numbness.

Sample Selection Process (see table 3-1)

Spouses of those who were age 50 or older at the time of death

were selected. A list of 156 names was compiled. Addresses and

phone numbers were obtained from the 1980 and 1981 Gainesville and

Alachua County phone directories. No attempt was made to contact

those without listed phone numbers (31) as phone calls were to be

used to make interview appointments after introductory letters were

sent. Of the remaining 125, an attempt was made to contact 116 by

letter. Because of the failure of the researcher to obtain consent

for interviews from any black persons in the pilot study for this

research (conducted in the spring of 1981) it was decided to limit

the sample to non-blacks, primarily because of a concern that the

sample of blacks would prove to be too small to be used in data anal-

ysis. Nine persons who were believed to be black based on obituary

listings of membership in all-black churches were not contacted. Of

the 116 contacted by letter, 12 letters were returned; 10 of these

were addressees who had moved with no forwarding address and 2 were

deceased. Of the remaining 104, 12 were unable to be contacted be-

cause the listed phone number had been disconnected or a person an-

swering the phone claimed that the person no longer lived at the num-

ber and that they were unable to give further information. Attempts

to find new addresses and phone numbers for these persons failed.

Ninety-two persons were contacted by phone after being sent an

introductory letter explaining the nature of the research (see

APPENDIX A for copy of introductory letter). Twenty-nine persons

refused to be interviewed. Of these 29 refusals, there were four

cases in which a family member refused the interview for the widowed

person without letting him or her speak on the phone to the re-

searcher. Some of the reasons given over the phone by widowed persons

themselves for refusing to be interviewed included not having the

time, being too ill, feeling the interview would be too emotionally

upsetting, not wanting to "get involved," and none of the

interviewer's business. Several potential respondents phoned the

researcher upon receipt of the initial letter to convey the latter

message, some with hostility and cursing. The interviewer was

surprised and somewhat dismayed by this phenomenon. A few women who

had been previously interviewed on the subject of widowhood by a

doctoral student in clinical psychology at UF a few months earlier

did not wish to be interviewed again. A few people kept telling the

researcher to call back in a few days and did not actually refuse the

interview until the third or fourth call-back. Ultimately, 63

interviews were obtained. Of the 63, 3 persons turned out to be

black and were not included in the analysis. These three persons

belonged to integrated churches and so escaped the screening method

of the interviewer. The interviews were conducted, though, as the

appointments had been made and it would have been rude and awkward to

have cancelled.

Table 3-1 Sample Selection Process

Step 1: 156 names selected from obituary listings of spouses of
deceased men and women aged 50 or older at the time of death

Step 2: Addresses and phone numbers obtained for 125 of the 156

Step 3: 9 blacks dropped from sample, leaving 116 persons

Step 4: Initial contact letters mailed to 116 persons
12 letters returned to sender as undeliverable
(10 had moved with no forwarding address; 2 were

104 potential respondents

Step 5: The researcher attempted to make phone contact to arrange
for interviews with 104 persons.
12 of the 104 had disconnected phones or their whereabouts
were unknown to answerer of phone call

92 potential respondents were contacted by phone

Step 6: 29 persons refused to be interviewed
63 persons agreed to be interviewed
3 of the 63 interviewees were black and were dropped from
the analysis

60 usable interviews were obtained

The refusal rate was about 32%. It is not possible to know exact-

ly how the refusers differ from those interviewed, yet it may be the

case that those who refused were more likely to be ill, living with a

relative who took care of them, or more depressed or angry than those

who agreed to be interviewed. It seems likely that the final sample

is biased in favor of those who were somewhat healthier and more inde-

pendent or for whatever reasons more willing to talk with a stranger

about their lives as widowed persons.


The interviews (see APPENDIX B for copy of interview schedule)

were in-person and all but one were conducted by the researcher in

the homes of the respondents. The one exception was a widower who

preferred to be interviewed at his place of work. Consent forms were

signed prior to each interview as required by the University Human

Subjects Committee (see APPENDIX C for copy of consent form). Confi-

dentiality of the respondents' answers and identities was assured by

having no identifying information on the interviews and by allowing

no one except the researcher to have access to the list of names and

addresses of respondents or actual data.* The average length of the

interviews was one and one-half hours with a range of from 45 minutes

to over 4 hours. In the latter case the interviewer was invited to

stay for lunch and was shown family photo albums. The photos includ-

ed many of the respondent's cat who had died of old age a few days

prior to the interview. The respondent, a woman in her 80's, talked

about the cat as if it had been a person and seemed very upset over

its death. It had been her only companion since her husband's

death. In another case, the respondent was babysitting her infant

grandson while the interview took place. The baby's bottle nipple

was not working very well whereupon the woman became visibly upset.

*In one case the interviewer was told by the respondent to tear up
the notecard bearing the respondent's name and address before the
interview could begin. The card was later repaired with cellophane

It was only after the researcher went out and bought another bottle

nipple that the interview could continue. In several cases, respon-

dents seemed to regard the interview as an important social occasion,

wearing very nice clothes and providing refreshments. In one case

the respondent, a woman in her 70's, had set out her best china and

served tea and freshly baked cake. This type of thing seemed to

occur more often with respondents who had lower levels of social

interaction than the sample on average. Surprisingly, only a few

respondents cried or seemed upset during the interview. One man in

his 70's, however, cried for nearly 30 minutes at the end of the in-

terview. The interviewer stayed with him until he seemed to feel bet-


No attempt was made to limit the length of the interviews or the

talkativeness of the respondents. In several cases, additional

information was obtained after the "official" interview was over

which added to the overall picture of the respondent's situation.

Much of this information was written down by the researcher in her

car immediately following the interview while it could still be

remembered clearly. This information included comments made by the

respondents, observations about their homes such as the woman who had

14 photos of her husband in her living room, or observations about

their behavior such as the woman who petted her small dog so

furiously during the interview that the interviewer was concerned for

the dog's safety. The latter widow said the interview had not upset

her but her manner would seem to indicate otherwise. Several re-

spondents commented that the interviewer looked like or reminded them

in some way of a niece or granddaughter. It could be that identi-

fying the interviewer with a female relative in this way made the

respondents feel more comfortable in answering her questions. Others

commented that the interviewer was much too young to understand such

a thing as widowhood. Ideally, such interviews should probably be

conducted by another widowed person or one closer to the

respondents' age range. Many respondents became less reticent after

the interview was over when the interviewer's pen and paper were put

away. A tape recorder was not used since respondents in the pilot

study said, when asked, that a tape recorder would have made them

feel uneasy and less willing to discuss sensitive issues.

The Interview Schedule

The interview consisted of both open-ended and forced-choice

questions, including questions in the following categories:

(1) Social and Demographic Variables-- Some examples in this

category are age, education, employment status, family composition,

living arrangement (alone or with others), place of birth, length of

residence in Alachua County, and length of marriage to deceased


(2) Social Support Variables-- Some examples in this category

include types and sources of support received, perceived help and

lack of help with problems associated with widowhood, and questions

ascertaining whom the respondent tells his or her problems, whom he

or she turns to in a crisis, and whom he or she most enjoys being


(3) Attitudes and Subjective Well-Being-- This included global

questions on life satisfaction and happiness (two of the main depen-

dent variables in this study), satisfaction with amount of contact

with children and with other relatives, and an open-ended question

concerning the main problems experienced in the first few months of

widowhood. The latter is not an attitude per se but is attitudinal

in the sense that it asks the respondents for their subjective assess-

ments of what the main, most important, or most difficult problems


(4) Social Interaction-- This category is distinct from social

support. Some research has used social interaction as an indicator

of support but, while there is some overlap, they are not inter-

changeable concepts. Social interaction may be supportive, neutral

or even negative. For example, one woman reported that her son and

daughter-in-law came to see her several times a week but that their

purpose in so doing was to talk her into giving up her home so that

they might have it. She perceived this as exploitative and was very

distressed by it. Some examples in the social interaction category

are a social activity scale (which includes information on with whom

the respondent participates in the listed activities), a question on

organizational involvement, number of close friends and number of

widowed close friends, frequency of church attendance and extent of

involvement in other activities in church.

(5) Religion-- At least two aspects of religion are important in

this study. First, degree of religious participation may be an indi-

cator of social integration into a network of like-minded others who

are a potential and/or real source of support and comfort for the be-

reaved. This is borne out by the substantial proportion of respon-

dents who listed church members, and a few clergy, as sources of help

with problems, as close friends, and as those to whom they would turn

in a crisis. Secondly, religious faith and some types of beliefs

offer the bereaved a meaningful framework in which to understand and

come to terms with their loss. The religion questions include re-

ligious affiliation, strength of religious identification, atten-

dance, other church involvements (such as choir, Sunday school, board

of trustees), personal importance of religion, and several questions

concerning beliefs. Questions on beliefs include whether or not the

respondents believe in an afterlife and if so, if they believe they

will be reunited with family members after death. It is supposed

that such a belief would ameliorate the painfulness of grief by inter-

preting the death as less than a permanent separation from the

spouse. Even a belief in an afterlife that does not include reunion

may offer comfort by assuring the bereaved that the deceased spouse

is not completely annihilated but exists in some other form. Several

questions ascertain the orthodoxy of the beliefs of the respondents.

Another very important question is an open-ended query concerning the

meaning of suffering and death. This is used to determine whether

the respondent's beliefs have provided an explanation for death and

suffering and the nature of that explanation. This was, in some

cases, a difficult question for respondents to answer. Some

respondents were able to respond easily and articulately while others

were less able to put into words their beliefs and feelings on this

question. To some extent, educational and occupational differences

may explain the differences in ability to verbalize on this issue.

For example, some of the respondents who had been teachers seemed

very much at ease in explaining their beliefs, while others, such as

the 80-year-old widow who had always been a homemaker, were less able

to formulate an answer. Despite these limitations, the question is a

useful addition to the forced-choice questions which may have more

demand characteristics and may also fail to tap respondents' specific

beliefs as well as this question. With these constraints in mind, it

is supposed that those who were unable to answer the question at all

and those who gave negative answers (for example, "Suffering is pun-

ishment for killing Jesus") would score lower on measures of sub-

jective well-being than those who gave coherent neutral (for example,

"Death is a natural part of life.") or positive (for example, "It is

passing over to the other side, a better, happier life.") answers.

That is, the researcher believes that in spite of differences in

ability to answer the question based on education, verbal ability,

etc., differences in ability to answer and type of answer given do to

a large extent indicate differences in the degree to which the

respondents' beliefs provide for them an explanation for suffering

and death. It is further believed that the nature of that

explanation affects subjective well-being. For example, it seems

logical that a respondent who believes that death and suffering are

completely negative phenomena will experience a lower degree of

subjective well-being than one who believes that, perhaps, suffering

and death build character and bring families closer together. In

other words, it is thought that subjective well-being would be better

in those people whose religious beliefs encompassed the idea that

life is basically good and that death does not negate that essential

goodness. A subsequent question more explicitly asks if any good can

come from suffering. The above logic applies here as well. It is

supposed that those who say yes and can give examples ("gives you

more appreciation of the good health you have") will score higher on

measures of subjective well-being. Another question asks whether or

not the respondent's religion has been a comfort since the spouse

died and if so, in what way. This is to determine what, if anything,

about the respondents' religion has been helpful in their own


(6) Life Events-- This is a shortened version of the Holmes and

Rahe (1967) Life Events Scale, including ten of the most stressful

and most probable events for older persons. It was decided that the

Life Events Scale in its entirety was too long for inclusion in an

interview containing so many types of questions. It seemed a good

idea, then, to include only those few items which, in the research-

er's judgment, were most likely to have occurred in the lives of this

particular group of respondents. Another criterion was to include

only those items which Holmes and Rahe had rated as particularly

stressful, since the purpose of including life events at all was to

assess what major stresses the respondent had experienced in addition

to becoming widowed. For example, birth of a child was omitted while

death of a family member was left in. The portion of the scale used

in this analysis includes deaths of friends and family members, ill-

nesses, retirement, etc. Inclusion of other deaths is particularly

important in assessing possible "bereavement overload." Persons exper-

iencing multiple losses of significant others in a relatively brief

period, three years in this case, may be more at risk for "bereave-

ment overload."

(7) Health and Subjective Well-Being-- This category includes a

global question ascertaining self-rated health and the Leighton

Health Opinion Survey, a 20-item scale including Depression, Anxiety

and Physical health problems subscales (Macmillan, 1957; Langner,

1962; and Leighton et al., 1963). The Depression and Anxiety

subscales are used as dependent variables indicating degree of

subjective well-being or adjustment to widowhood.

(8) Other-- Other questions ascertain the spouse's cause of

death, place of death and length of illness. The latter is important

in determining whether relatively sudden deaths are associated with

lower scores on measures of subjective well-being than deaths fol-

lowing longer illnesses. This variable will allow for some analysis

of whether or not "anticipatory grief" is operative or helpful for

this sample.

(9) Evaluation-- The final portion of the interview asks if the

respondents were upset or offended by the interview and gives them a

chance to make any additional comments about their lives as widowed


Methods of Data Analysis

The sample of sixty widowed persons in this study constitutes a

non-probability sample of widowed persons in the Alachua County

area. Because this is a non-probability sample, the use of infer-

ential statistics such as regression is not appropriate. Another

reason that regression is not used has to do with the dependent var-

iables used in this study. The global questions on life satisfaction

and happiness have ordinal response categories. It is therefore more

appropriate to use categorical measures when these two dependent vari-

ables are used. The depression and anxiety variables are interval.

However, the skewedness of responses would make the use of regression

misleading. However, certain measures of association such as Pearson

correlation, Kendall's Tau-C and Chi-square, and cross-tabulation

tables will be useful in determining the direction and strength of

relationships between variables for this particular sample. The appro-

priate measures will be used in each case depending on whether the

variables under consideration are at the interval, ordinal or nominal

level. Although it is not appropriate to generalize the results of

statistical analysis with this sample, the use of various statistical

measures will enable this researcher to draw some conclusions about

factors affecting the subjective well-being of these respondents.

The implicit assumption throughout most of this study is that the

measures of subjective well-being are dependent variables. That is,

other variables are assumed to affect the dependent variables. There

are two reasons for making this assumption. The first is based on

intuitive logic. It seems reasonable and logical to assume this

pattern of directionality. The second is based on previous

research. The vast majority of research reviewed for this study ex-

amines the influence of various factors on well-being. Tests of

significance, since they involve assumptions about probability sam-

ples, will be used cautiously. In addition, "case study" examples

and quotes from respondents will be used to complement, support or

challenge the findings from statistical analysis.

Critique of Methods in This Study and Others

One limitation of this study is the relatively small sample size

of 44 widows and 16 widowers. It would have strengthened the study

to have had a larger sample, yet the high refusal rate and the fact

that the entire study, including all of the interviewing, was done by

only one person in a limited time frame prevented the sample from

being any larger. Other reputable studies of widowhood have involved

similarly small samples. For example, Glick et al. (1974) had a sam-

ple of 49 widows and 19 widowers. Although it might have been pos-

sible to have enlarged the sample by using mailed questionnaires

and/or phone interviews, one of the advantages of in-person inter-

views is the opportunity to obtain actual quotes from respondents,

make behavioral observations and in general obtain more and better

information than is possible with other data collection methods.

-Aty-oin hindsight, it seems it might have been better to go outside

Alachua County as well as inside for the sample. Several widows

had been interviewed by another University of Florida student for her

dissertation on the psychology of widowhood and did not wish to be

interviewed again. Still others seemed to be suspicious of college

students in general. These factors may have increased the refusal

rate. Also, the "college town" sample turned out to include a higher

proportion of college graduates than would probably be typical of

this age group in the United States as a whole.

In this study's favor, it can be said that there was a moderately

successful attempt to include widowed men as part of the sample. The

majority of studies of widowhood have focused exclusively on women,

including probably the most famous recent study by Lopata (1979) on

the social support systems of widows.

Another positive feature of this research is the inclusion of

specific questions about social support as well as those concerning

social interaction. Many studies have viewed social support solely

as emotional support while, at the same time, measuring support by

assessing frequency of social interaction (Bankoff, 1983). Thus, sev-

eral studies have equated social interaction with social support.

This study makes it possible to distinguish between the two con-

cepts. One criticism of this study might be that it does not measure

frequency of interaction with social network members, just types of

interaction and persons with whom interaction takes place. In the

pilot study, frequency of interaction with all possible family net-

work members, friends and neighbors was ascertained. It simply took

too long, was tedious, and wore out the respondent. In one case, a

widower in his 80's had 11 children, 40 or so grandchildren (he

wasn't sure of the exact number), and 100 or so friends. For him to

answer the questions on frequency of interaction took over an hour

and was very taxing. For this reason the frequency-of-interaction

questions were dropped in favor of a broader range of questions. It

must be remembered that the respondents in this study were,

presumably, under some stress as a result of their bereavement and

that many were elderly and/or ill with limited energy levels. For

these reasons, keeping the length of the interview manageable was an

important concern. As it was, the interview was probably too long

for several of the respondents, although all of the interviews were


The questions on religion make this study unique. There are few

studies of widowed persons which include questions on religious

participation and beliefs. Most studies which do include questions

on religion limit themselves to questions on denominational affili-

ation and frequency of attendance as indicators of religiosity. It

has been found in an earlier study by this researcher that salience,

or importance, of a person's religion to him or her combined with

measures of orthodoxy, is a better predictor of some types of atti-

tudes than denominational affiliation or orthodoxy alone (Beck,

1980). The current study includes a question on the personal impor-

tance of religion and questions on specific religious beliefs, which

are hypothesized to be predictive of subjective well-being in widow-

hood. One question which, in hindsight, it seems could have been

better stated is the open-ended query concerning the meaning of

suffering and death. The two ideas should have been separated into

two questions.

Another comment concerns the measures of the dependent

variables. The two global questions on life satisfaction and

happiness are widely used and widely accepted as measures of

subjective well-being or adjustment. The other measure of subjective

well-being, the Leighton Health Opinion Survey, is less well known.

The Leighton Health Opinion Survey was developed as a method to

distinguish between persons experiencing psychological distress, or

mental impairment, and those who were not, by ascertaining the

presence or absence of a number of psychological and physiological

symptoms commonly associated with mental illness (Leighton et al.,

1963; Langner, 1962). Even a recent assessment (de Marco, 1984)

rates the Leighton Survey and its slightly modified form, the 22-item

Langner Mental Health Scale, as a useful index of mental impairment

or psychological distress. The Health Opinion Survey, however, has

been criticized on the basis that its inclusion of physiological

symptoms may yield high scores for those with physical disease as

well as psychological impairment (Crandell and Dohrenwend, 1967;

Gaitz and Scott, 1972). This may be more of a problem in samples of

older persons more likely to have symptoms of true physical disease.

A similar criticism of this type of measure comes from Thoits

(1981). She claims that this measure includes many questions on

physiological and psychophysiological symptoms which may be

confounded with similar items in life events scales in studies which

include both. This is not a major problem in this study since the

scale used includes only one question concerning the health problems

of the respondent. The HOS does, however, have the advantage of

including subscales on depression and anxiety as well as purely

physical symptoms, which may reduce some of the confounding of

physical and psychological symptoms. In this researcher's opinion,

while there may be better scales available to measure psychological

subjective well-being, the Leighton Health Opinion Survey, with its

subscales, is appropriate and useful for this study in distinguishing

between those who are experiencing relatively more psychological

distress and those who are not. Also, it has the advantage of not

being obvious about assessing mental health. A more obvious approach

might have elicited more socially desirable responses and fewer

truthful responses. The researcher feels that this is a real concern

for this sample, many of whom seemed to be trying quite hard to

convey the impression that they were coping well.


This chapter includes a description of the process by which the

sample was obtained, a discussion of the process of obtaining inter-

views and the difficulties encountered therein such as refusals, peo-

ple who had moved with no forwarding address and so on. The process

and problems of the interview process, such as crying respondents,

the willingness of most respondents to be quite open, and observa-

tions made outside the "official" interview process were discussed.


The items in the interview schedule and the rationale for their inclu-

sion were discussed. The chapter ended with a discussion of the meth-

ods of data analysis employed in this study and a critique of their



This chapter gives the reader a detailed description of the

sample. Distributions of respondents on most variables are dis-

cussed. Along with this descriptive material are included all of the

hypotheses used in the research. The hypotheses are inserted immedi-

ately after the relevant variables are discussed. Wherever it is

applicable, references suggesting or supporting each hypothesis are


The sample includes 44 widows and 16 widowers. Their mean age at

the time of the interview was 68.8 years with a range of 41 to 88

years old. The mean education was 12.4 years, or just beyond high

school. The range was from completing the second grade to Ph.D.

level. Seventeen respondents had completed less than 12 years of

schooling. Twenty were high school graduates. Six had attended

college but not graduated. Eight had bachelor's degrees and nine had

graduate degrees. These respondents probably are, on the average,

more highly educated than their age-mates nationally.

Hypothesis 1: It is hypothesized that those with higher levels

of education will have higher levels of social participation (Lopata,

1979; Kernodle & Kernodle, 1979).

Social And Demographic Characteristics

Table 4-1 Social and Demographic Characteristics of Respondents
Interval Level Variables





Length of Residence
in Alachua County

Length of Marriage

Number of Sons

No. of Daughters

No. of Brothers

No. of Sisters

No. of Grandsons

No. of Grand-

4ean St

68.8 yrs.

12.4 yrs.


27 yrs.

39.4 yrs.







andard Deviation














41-88 yrs.

2 yrs. to Ph.D.


63 yrs.

64 yrs.






0 to 24

Table 4-2 Social and Demographic Characteristics
Dichotomous Variables


Currently Employed? 17 43

Born in Florida? 17 43

Currently Live with
Someone Else? 22 38

Currently Remarried? 4 56

The mean annual income was $19,700 but the median was only

$13,500. The mean was skewed upward by one man's atypical income of

$230,000. The range of annual income was from $2,000 to $230,000.

Thirteen people had incomes of $5,000 or less, thirteen in the $5,001

to $10,000 range, fifteen in the $10,001 to $20,000 range, eleven be-

tween $20,001 and $30,000, five between $30,001 and $50,000 and one

had the aforementioned income of $230,000. Two people chose not to

answer this question. Both appeared to the interviewer to probably

be in the lower end of the income range, based on their homes.

Several others were initially reluctant but gave the interviewer the

information after the interview was officially over, sometimes as the

interviewer was walking out the door. One widow commented that with

all the things she'd revealed to the interviewer, there was really no

need to keep her income a secret.

Hypothesis 2: It is hypothesized that those with higher incomes

will have higher levels of social participation as measured by the

social activity scale, organizational involvement, and church atten-

dance (Antonucci, 1985).

Forty-three people were either retired or had never been employed

outside the home (homemakers). Eight had part-time jobs and nine

worked full-time. Information on occupation (or deceased spouse's

occupation) was not obtained for the 43 who were not employed at the

time of the interview. The question was designed primarily as an

indicator of social involvement rather than socio-economic status.

Hypothesis 3: It is hypothesized that those who are employed

part or full-time will have higher levels of social participation,

since they probably come into contact with other people with whom

they may interact at work.

Two respondents were born in Gainesville, 15 elsewhere in Flor-

ida, 15 elsewhere in the southeastern United States, 26 somewhere

else in the United States and two in other countries.

The mean length of residence in Alachua County was 27 years with

a range of from 2 to 63 years. In several cases, the respondent had

retired in Florida with his or her spouse shortly before the spouse's


Hypothesis 4: It is hypothesized that persons living in the area

for longer periods of time will have higher levels of social partici-

pation, as they would have had a longer period of time in which to

make friends and become involved in various social activities.

The mean length of marriage was 39.4 years with a range of from 1

to 64 years. Fourteen respondents had been married before being mar-

ried to the deceased spouse. Of these, five had been previously wid-

owed and nine had been divorced.

Hypothesis 5: It is hypothesized that length of marriage will be

positively associated with measures of subjective well-being since

death may seem more "appropriate" to the surviving spouse in mar-

riages of longer duration in which, presumably, the spouses are

older. Controlling for age is expected to reduce the effect of this

variable (Parkes, 1972).

Family composition was as follows. The mean number of living

sons was 1.3 with a range of from 0 to 6. The mean number of daugh-

ters was also 1.3 with a range of from 0 to 6. Only six respondents

had no living children. The mean number of living brothers was .88

with a range of 0 to 4. The mean number of living sisters was 1.11

with a range of 0 to 5. It was quite common for a respondent to have

had more than one sibling die in the recent past. The mean number of

grandsons was 3 with a range of 0 to 25. The mean number of grand-

daughters was 2.5 with a range of 0 to 24.

Hypothesis 6: It is hypothesized that the number of available

kin will have no influence on measures of subjective well-being.

This is hypothesized to be true since it is actual support and inter-

action, not just their potential which will affect subjective well-

being (Gibbs, 1985; Wellman, 1981).

Living arrangement was also assessed. Thirty eight respondents

lived alone. Twenty-two lived with another person, in most cases an

adult child who had come to live with the respondent and in a few

cases one who had not yet left home for the first time. Four res-

pondents (3 men and 1 woman) had remarried and consequently lived

with their new spouses. Only two persons in this sample had ever

gone to live in the home of a relative. Apparently some of the peo-

ple who refused to be interviewed, or whose children refused for

them, did live with relatives.

Hypothesis 7: It is hypothesized that living alone will be

positively associated with measures of subjective well-being. This

is because living with others is considered in many groups in Ameri-

can society to be less desirable than independent living (Brody, John-

sen, and Fulcomer, 1984). Also, the strain of having an adult child

move back home may possibly outweigh the benefits of having them

around. Those four who are remarried will be removed for this part

of the analysis.

Social Supports

The respondents were asked whether or not they received each of

nine different types of aid, and if so, from whom this aid was

received (see table 4-3). Overall, the mean number of types of

support received was 4 with a range of from 1 to 8. The types of

support were divided into two categories, expressive support and

concrete (or instrumental) support. Expressive support includes

having someone care for you when you are ill, help you make important

decisions, and help you emotionally when you are sad or depressed.

Concrete support includes help with transportation, household

repairs, shopping, housekeeping, yard work and financial aid. Legal

aid was originally one of the categories of concrete support but was

dropped early in the interview process as being too confusing,

overlapping too much with decision-making and being received

exclusively from lawyers. Expressive support had a range of from 0

to 3 types received with a mean of 1.7. Six people reported

receiving no expressive support; four of these were men. Concrete

support had a range of from 0 to 5 types received, with a mean of

2.6. Three people reported receiving no concrete support.

Table 4-3 Number of Social Supports Received

Mean Standard Deviation Range

Social Support Overall 4.3 1.8 1 to 8

Expressive Support 1.7 .95 0 to 3

Concrete Support 2.6 1.3 0 to 5

The mean amount of social support received from kin was 3.58

types compared with 2.25 types from non-kin sources. The mean amount

of expressive support received from kin was 1.60 compared with a mean

of .75 types from non-kin. The mean amount of concrete support re-

ceived from kin was .68 types compared with .62 for non-kin. Gener-

ally speaking, then, these respondents received more social support

from kin than from non-kin, especially expressive support.

Hypothesis 8: It is hypothesized that support received will be

positively associated with measures of subjective well-being (Mad-

dison and Raphael, 1975; Turner, 1981). This hypothesis is based pri-

marily on the notion from the "stress-buffering" hypothesis that

social supports ameliorate the stress experienced from negative life

events. Therefore, social supports should improve subjective well-

being during bereavement (Turner, 1981; Thoits, 1982).

Hypothesis 8a: It is further hypothesized that expressive

support will have stronger effects than concrete support.

One question asked whether the respondents perceived any lack of

help with the problems of being widowed. Only 10 persons reported

such a lack. When asked what type or types of help were lacking, the

most common response was basically that no one would listen sympathet-

ically and nonjudgmentally to the expressed feelings of the widowed

person or that there was no one with whom the respondent could "real-

ly talk."

Hypothesis 9: It is hypothesized that a reported lack of help

will be negatively associated with subjective well-being (Sarason et

al., 1983; Barrera, 1981; Ward, Sherman, and LaGory, 1984).

Another question asked to whom the respondent told his or her

problems. Thirteen told no one, 5 told God and 4 told themselves

only. Therefore, fully 22 people told no other living person their

problems. It may be that Gorer (1967) was correct in his assertion

that death is a subject which can rarely be mentioned openly or dis-

cussed directly in polite conversation. Gorer has further stated

that happiness is seen as an obligation by many people so that talk-

ing about death and its attendant unhappiness may be discouraged.

Wass (1979), in reviewing research on the willingness of older per-

sons to talk about death, points out that whether or not older per-

sons talk about death depends on whether they find themselves in a

restrictive or permissive environment. One may feel restricted by

being with, for example, younger members of one's family but rela-

tively free to talk about death and the problems of widowhood in a

more permissive environment, such as with other widowed persons.

There may also be ethnic, religious, and other social group differ-

ences in the degree to which it is permissible to discuss death and

related topics. Additionally, many Americans, including many widowed

persons, value independence quite highly and may try to remain stoic

rather than ask for a shoulder to cry on, no matter how sorely needed

that shoulder may be. Others may feel that problems are appropriately

shared amongst family and friends and be less reticent to share their

needs and concerns. This sample would appear to lean toward valuing

stoicism and independence. Of those who did tell their problems to

someone, 13 told daughters, 3 told sons, 3 told sisters, 6 told other

family members and 13 told friends. (These figures are based on the

first person listed for each respondent. Later analyses take into

account the second person listed as well as the first).

Hypothesis 10: It is hypothesized that those who told their

problems to no one will have lower scores on measures of subjective

well-being than those who told someone.

Hypothesis 10a: It is also hypothesized that those who told non-

kin their problems will have higher scores on measures of subjective

well-being than those who told kin. This is based on the research

literature claiming a more positive effect on morale of interaction

with friends than with kin (Arling, 1976; Bankoff, 1983).

A related question is to whom the respondent would turn in a cri-

sis. Three respondents said they would turn to no one, six to God

and five to themselves only. Thirteen would turn to a daughter, six

to sons, four to siblings, nine to other family members, eight to

friends and only two to a minister. One man said he would turn to

his tutelaryy guardian," whom he emphatically denied was the same as

God, but declined to explain further. In the latter two questions,

daughters were relied upon more than any other category of persons.

This lends support to the common finding in gerontological research

that adult daughters figure prominently in their parents' support

networks. Friends were also frequently relied upon.

Hypothesis 11: It is hypothesized that those who said they would

turn to no other person in a crisis will have lower scores on mea-

sures of subjective well-being than those who would turn to someone


Hypothesis 11a: It is further hypothesized that those who would

turn to kin will score lower on measures of subjective well-being

than those who would turn to non-kin. The argument for this is the

same as that mentioned for the previous question concerning to whom

problems are told.

The respondents were also asked who they most enjoyed being

with. Seventeen chose friends, twelve chose daughters, one chose a

son, one chose an in-law, four chose fellow church members, four

chose family in general, eight cited children without specifying sex,

three cited grandchildren, 2 cited people in general, 3 chose the new

spouse, and 5 preferred to be alone.

Attitudes and Subjective Well-Being

This section includes a discussion of the questions on life

satisfaction, happiness, satisfaction with amount of contact with

children and with other relatives and a discussion of the problems of

widowhood cited by the respondents themselves.

In answering the global question on life satisfaction, 12 people

reported themselves completely satisfied, 32 said they were pretty

satisfied and 16 were unsatisfied. In the first two categories,

several people added the qualifier "under the circumstances" or "as

satisfied as I'll ever be without (spouse's name)." Thus it seems to

the researcher that some of these people might really have been less

than completely satisfied or pretty satisfied but resigned to being

widowed and determined to make the best of it. On the other hand,

they might really have been less satisfied than they were willing to

admit. In a few cases, other comments the respondents made about

their lives made their claims to be completely or pretty satisfied

somewhat less than convincing. The fact that they were interviewed as

widowed persons may have biased their answers in a positive direc-

tion. If they had been interviewed as, for example, Gainesville res-

idents, perhaps they would have answered in a broader context.

In answering the global question on happiness, 13 respondents rat-

ed themselves as very happy, 30 said they were pretty happy and 17

claimed to be unhappy (not very happy and very unhappy were collapsed

into one category since there were only 2 respondents in the very

unhappy category).

There is a high degree of overlap in life satisfaction and

happiness, with those being completely satisfied being likely to also

be very happy, and so on. The Chi-square for this comparison is sig-

nificant at the .01 level (see table 4-4).

Table 4-4 Cross-Tabulation of Life Satisfaction by Happiness


Very Happy Pretty Happy Unhappy
Life Satisfaction

Completely Satisfied 7 5 0

Pretty Satisfied 5 22 5

Unsatisfied 1 3 12

Chi-square = 32.8; df=4; p = .01

In assessing satisfaction with amount of contact with children,

thirty-six respondents were found to be very satisfied, 14 were satis-

fied, 2 were unsatisfied and 1 was very unsatisfied. For purposes of

analysis, the latter three categories will be treated as one, since

there are so few respondents in the latter two categories. In a cou-

ple of cases, respondents reported that their children came to see

them too much. Still others' adult children lived with them. In some

cases, this was seen as problematic. For example, one woman whose 40-

year-old divorced daughter lived with her was terribly concerned over

what she felt was her daughter's promiscuous and unacceptable life-


Hypothesis 12: It is hypothesized that those who were very satis-

fied with the amount of contact with their children will have higher

subjective well-being scores than those who answered differently

(Gibbs, 1985).

In the assessment of satisfaction with amount of contact with

other relatives, 27 claimed to be very satisfied, 23 were satisfied

and only 5 were unsatisfied. Of those making additional comments on

this question, most said that geographical distance and the expense

of travel prevented their seeing family members as much as they would

have liked.

Hypothesis 13: It is hypothesized that those who were very satis-

fied with the amount of contact with other relatives will have higher

subjective well-being scores than those who answered differently

(Gibbs, 1985).

Another question asked the respondents what were, in their opin-

ion, the main problems experienced in the first year of widowhood.

Loneliness was by far the most often cited, listed first by 28 re-

spondents. Other problems included numbness and shock, depression,

emotional problems, financial problems, decision-making, getting used

to living alone, physical illness, fear of going crazy, feeling ex-

ploited by family members, receiving obscene phone calls, guilt, and

one attempted suicide. Only one person claimed to have had no prob-

lems. When asked who helped with these problems, children, especial-

ly daughters, and friends were mentioned first, with ministers, pro-

fessionals, and neighbors being mentioned by a few. Fifteen people

said they had received help from no one. Some were satisfied with re-

ceiving no help, saying the problems were just something they had to

cope with alone; others felt abandoned and resentful because of the

lack of help.

Hypothesis 14: It is hypothesized that those who reported re-

ceiving help from no one will score lower on measures of subjective

well-being than those who reported receiving help with the problems

of widowhood (Maddison and Raphael, 1975).

Social Interaction

On the scale measuring activities shared with others, the range

of number of activities shared was from 2 to 11 with a mean of 7.6

and a mode of 9. The two social activities engaged in by the great-

est number of respondents were talking on the phone (engaged in by

59) and correspondence (52). While family members, especially daugh-

ters, were most frequently the ones with whom shopping trips and hol-

idays were shared, friends were most often cited for every other

activity on the list. Apparently, social interaction on a daily

basis takes place mostly with friends for this group of respondents.

This is probably a contributor to subjective well-being since,

according to Leslie (1982,p.639), "extensive association with friends

is an important mechanism of adjustment to old age and widowhood."

Hypothesis 15: It is hypothesized that higher levels of shared

activity will be positively associated with measures of subjective

well-being (Anderson, 1984; Morgan, 1984).

Hypothesis 15a: It is also hypothesized that higher levels of

activity will be positively associated with measures of social


Hypothesis 15b: It is expected, in addition, that activities

shared with non-kin will have different effects on measures of sub-

jective well-being than those shared with kin (Arling, 1976).

Organizational involvement was assessed as an aspect of social

interaction. Thirty-five respondents belonged to one or more or-

ganizations and twenty-five belonged to none.

Hypothesis 16: It is hypothesized that those who belong to one or

more organizations will have higher scores on measures of subjective

well-being than those who do not. This hypothesis is based on the

notion that organizational involvement provides the opportunity for

social interaction.

The number of close friends was also assessed as a measure of

social interaction. The mean number of close friends was 14.8 with a

mode of 4 and a range of from 0 to 100. Five people said they had no

friends, 21 had from 1 to 5, 11 had from 6 to 10, fourteen had from

11 to 20 and fifteen had more than 20. Of the latter group, 3

claimed to have 100 close friends. In all 3 cases, the respondents

were actively involved in a church and counted fellow members as

close friends. One could reasonably argue that the definition of

"close friend" varies greatly from respondent to respondent. The

major consideration in this study, however, is the respondent's

perception of having no, a few, or many close friends, not numerical


Hypothesis 17: It is hypothesized that those having no or a few

close friends will differ on measures of subjective well-being from

those having more close friends (Lowenthal and Haven, 1968).

Hypothesis 17a: It is also hypothesized that those with more

close friends will have higher activity levels, since they would

presumably participate in those social activities with their


The number of close friends who were also widowed was also as-

sessed in an attempt to ascertain whether the respondents' close

friends were "in the same boat" and possibly more sympathetic than

non-widowed friends. Sixteen people had no widowed friends, thirty

had from 1 to 5, and nine had 6 or more. (Five had no close friends

at all.)

Hypothesis 18: It is hypothesized that those with no widowed

friends will score lower on measures of subjective well-being than

those with 1 or more widowed friends (Glick et al., 1974).

Church attendance is also viewed as a type of social interaction,

or at least potential interaction. Thirty-six respondents attended

weekly or more often, six attended one to three times a month, three

a few times a year, four once a year or less and eleven never

attended. About a third of the sample (22) participated in

activities in their churches in addition to the worship services.

Hypothesis 19: It is hypothesized that church attendance will be

positively associated with measures of subjective well-being.


Attendance, considered both a part of the religion section and

the social interaction section, has been discussed in the preceding

social interaction section. The religious affiliations of the sample

members (see table 4-5) include one Congregationalist, four Epis-

copalians, one Christian Church member, eleven Methodists, ten Presby-

terians, six Catholics, one American Baptist, fourteen Southern Bap-

tists, eight sectarians (such as Jehovah's Witnesses, Apostolic, Holi-

ness, etc.), and three who claimed to have no religious affiliation.

One man refused to answer this question, as well as all the other

questions on religion. The distribution of religious affiliations is

typical of religious affiliations in the South, which is dominated

mainly by Southern Baptists and Methodists (Hill et al., 1972).

Table 4-5 Religious Affiliations of Respondents

Congregationalist 1
Episcopalian 4
Christian Church 1
United Methodist 11
Presbyterian 10
Catholic 6
American Baptist 1
Southern Baptist 14
Sectarian groups 8
No affiliation 3


Respondents' Strength of identification with their church was

assessed. Seventeen respondents said they were very strong in their

identification with their church, 24 were somewhat strong and 16 were

not very strong. This variable is used to see how strongly the re-

spondents identify themselves as Methodist, Baptist, etc.

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd