Quality of Life Research 12 (Suppl. 1): 25–31, 2003.
25
Ó 2003 Kluwer Academic Publishers. Printed in the Netherlands.
Social support and quality of life
Vicki S. Helgeson
Department of Psychology, Carnegie Mellon University, Pittsburgh, PA, USA
(E-mail: vh2e+@andrew.cmu.edu)
Accepted in revised form 9 February 2002
Abstract
Social support is a broad term, which includes the supportive ways that different people behave in the social
environment. Structural measures of the environment deal with the mere existence of social relationships.
Functional measures refer to the resources that people within an individual’s social network provide.
Structural support shows a linear relation to quality of life; the functional aspects of support demonstrate
the stress-buffering hypothesis. One of the main focuses of this article is to help researchers determine what
aspects of social relationships or what types of support need to be measured and to consider the mecha-
nisms by which support might influence quality of life. Also addressed is how to translate the correlational
research on social support and quality of life into the field of support interventions, taking into account
individual and situational differences.
It is important to start this article by defining social daughter, etc. There also are measures of social
support. Social support is not, however, an easily integration, which are usually composite indices of
definable term. People use the term social support, some of the structural measures of support. Social
or the phrase social support, to refer to a wide integration often includes marital status, mem-
variety of phenomenon that characterize the social bership within an organization, and frequency of
environment, or the people who surround indi- social contact.
viduals in their network. Social support is a broad Functional measures are typically what people
term, which includes the supportive ways that dif- think of when considering social support. Func-
ferent people behave in the social environment. tional measures refer to the resources that people
When referring to the social environment, peo- within an individual’s social network provide.
ple generally distinguish between structural and Functional measures are qualitative measures.
functional measures of support. Structural mea- There are a number of taxonomies of support
sures of the environment deal with the mere exis- functions. Most taxonomies consist of these three
tence of social relationships. Structural measures basic functions: emotional support, instrumental
describe the existence, the interconnections, and support and informational support. Emotional
the relations among network members. I refer to support refers to having people available to listen,
these measures as quantitative measures of sup- to care, to sympathize, to provide reassurance, and
port because they usually reflect an amount of to make one feel valued, loved and cared for.
people or the amount of contact an individual has Instrumental support, sometimes referred to as
in his or her social environment. Some examples of tangible assistance, involves people providing
structural measures are marital status, how many concrete assistance, such as help with household
friends a person has, frequency of interaction with chores, lending money, or running errands. In-
friends, and the number of personal roles an in- formational support involves the provision of in-
dividual has, such as student, worker, parent, formation or guidance.
26
Which of these aspects of the social environment physicians and nurses. Informational support, by
has the strongest implications for quality of life? contrast, is not perceived as helpful from all
The main hypothesis advanced by Cohen and sources. People want informational support from
Wills in 1985 [1] is the ‘main effects’ vs. the ‘stress- the experts, i.e. from nurses and physicians. People
buffering’ hypothesis. The main effects hypothesis are not always fond of informational support
states that the more social support an individual provided by family and friends. This may be
has, the better the quality of life, regardless of the viewed as unwanted advice. It is not clear if the
person’s level of stress. The relation between source of support influences whether instrumental
quality of life and social support is linear. The support is perceived as helpful.
stress-buffering hypothesis, however, states that Researchers also have suggested the matching
the relation of social support to quality of life hypothesis, that the kind of support needed de-
depends upon an individual’s level of stress. If pends on the nature of the stressor [2]. I tested this
there is no stress or little stress, social support is idea several years ago in a study of cardiac patients
unrelated to quality of life. Under conditions of [3]. I thought ahead of time about what kind of
high stress, however, social support serves as a support I thought patients and spouses would
buffer against the adverse effects of that stressor. most need after a first coronary event. I reasoned
The person who faces high stress with support that the patient, upon discharge from the hospital,
resources is almost as well off as the person who is would probably need informational support. The
not experiencing the stressor. Cohen and Wills patient would need information on how to make
found that the distinction between the main effects lifestyle changes and how to prevent a future heart
and stress-buffering hypothesis maps on to the attack. What type of support would spouses need?
distinction I previously made between the struc- Would spouses also need informational support? I
tural and the functional aspects of support. The hypothesized that spouses would probably need
structural aspects of support fit the main effects instrumental support to assist them with the
hypothesis. Structural support shows a linear re- household activities that the patient is not allowed
lation to quality of life. When stress levels are low, to resume for a while. Spouses need people to give
simply being involved in a network, having people them concrete assistance with chores. Controlling
around, and feeling part of a group are related to for patient sex, the severity of the initial heart
an enhanced quality of life. The functional aspects problem, and level of distress in the hospital, I
of support demonstrate the stress-buffering hy- used regression analysis to predict which kind of
pothesis. When stress levels are high, resources are support (informational, emotional and instru-
needed from people within the social environment mental) predicted patient and spouse adjustment
to facilitate coping. 6 months later. For patients, the only type of
When considering the functional aspects of support related to reduced distress 6 months later
support, the next question typically asked is what was informational support. For spouses, the only
type of support function is best? Which kind of type of support related to reduce distress was in-
support best buffers stress? Many people believe strumental support.
that emotional support is the strongest stress The severity of the stressor also might indicate
buffer. There is some evidence that among the what type of support is going to be most strongly
three different types of support, emotional support related to quality of life. It has been argued that if
shows the strongest relation to quality of life. the stressor is controllable, informational support
However, this claim may be a bit overstated. To be may be the most helpful kind of support. People
fair, many investigators are so certain that emo- can provide you with information about what to
tional support is the most important kind of sup- do to actually alter the stressor. If the stressor is
port that they only measure emotional support. uncontrollable, emotional support may be most
The source of support may influence which kind of beneficial.
support is effective. Emotional support is helpful The timing of the stressor may influence the
no matter who the source is. Emotional support support needed. Jacobson [4] suggests there are
is helpful if it comes from family and friends or if three phases of a stressor: a crisis phase, a transi-
it comes from healthcare professionals, such as tion phase and a deficit phase. The crisis phase
27
occurs when one’s first made aware of the stressor. The difference between coders and participant
During this first phase, emotional support may be reports of a support transaction presents another
most needed. One needs to know that people will problem for research. In studies that measure both
be around and available to help if needed. It is perceived support and perceived receipt, perceived
during the transition phase that the individual support is more strongly related to quality of life
under stress actually copes with the stressor. At than received support. Why would this be? One
this time, people are needed to supply information possibility is that support perceptions are based on
on how to cope with the stressor. In the deficit personality. Certain people perceive support to be
phase, the individual is overwhelmed by the ex- available. People who perceive high levels of sup-
cessive demands of the stressor. Jacobson suggests port may be people who are more psychologically
that is when instrumental support is most needed. healthy. Another possibility is that support per-
There are some other distinctions you might ceptions are related to social skills. People who
want to consider in regard to support functions. perceive support have the social skills necessary to
One important distinction is between perceived elicit support. The literature suggests that per-
support and received support. Perceived support is ceived support is associated with some personality
typically measured by asking people to what extent characteristics, but is more than personality. Sup-
they believe people in their network are available port perceptions are likely to evolve from a history
to help them with household chores. For example, of having received support. However, this does not
‘To what extent is your spouse willing to listen to explain why perceived support is more strongly
your problems?’ ‘To what extent do you have related to quality of life than support receipt. One
friends who will give you information on how to possibility is that received support indicates a
take care of yourself?’ It is the perception of greater need for support. People who receive sup-
whether support is available that is being mea- port may be more distressed and have greater
sured. Received support, theoretically, should be needs for support. In this case, received support
whether someone actually received support from will appear to be related to more distress. I found
network members. Did someone listen to their support for this idea in the study of cardiac pa-
problems? Did someone run errands for them? Did tients previously mentioned [3]. Received support
someone provide information? The difficulty that was associated with greater distress, especially re-
arises is in how to measure received support. ceived informational support.
Typically, received support is measured by a per- Another aspect of the social environment that is
son’s perception of whether they received support. important to consider in quality of life research is
Such questions may ask, ‘In the last 3 months, did the negative side of social relationships. So far, I
someone listen to your problems?’ or ‘In the last have only addressed the positive aspects of the
3 months, did your physician give you information social environment. We know that social rela-
or advice about how to take care of your health?’ tionships also can be a source of conflict, stress,
What is really being measured is ‘perceived re- and tension.
ceipt’. The only true way to measure received In any study of the social environment, it is
support is to observe a support exchange. There important to know how frequently people argue
are some observational studies where people ac- with one another and how often conflicts occur.
tually observe support transactions and code what However, relationships can be a source of stress in
types of support exchanges occur. Those studies another way. Relationships are two-way streets.
are interesting. However, they usually are per- People in one’s social environment are not only
formed in a laboratory, which raises questions support resources but can be support burdens.
about generalizability to the natural environment. Certainly any study of caretakers should take into
We have performed these studies with college consideration burden or support provision in
students. We brought them into the lab, had them studies of quality of life.
engage in support exchanges, transcribed and Another source of stress from the social envi-
coded the conversations [5]. We typically find that ronment is unintended support failures. Unin-
the support coded by independent judges is not tended support failures are based on the idea that
the same as the support coded by participants. people in one’s social environment try to help,
28
have good intentions, and want to be helpful, but personal hypothesis is that it is difficult for people
say or do something that is not really helpful at all. under high levels of stress to admit that a network
These kinds of negative interactions have been member is less than supportive. When we con-
documented in the literature on people with cancer ducted a study of women with breast cancer, we
[6, 7]. Here are some examples. One is forced interviewed them while they were undergoing
cheerfulness. People with cancer say that people in chemotherapy. We measured social support from
their social environment tend to encourage them their oncologist. The oncologist is someone to
to be optimistic, to look on the bright side of whom they, in essence, have entrusted their medi-
things, and tell them not to worry about their cal treatment, if not their life. We asked them
problems. When someone tells you not to worry whether they felt their oncologist was a good
about your problems, is that reassurance – emo- person and whether they were happy with this
tional support? It can be, but it also can be per- relationship. How likely is it that these women are
ceived as minimizing the person’s problems, which going to say that this doctor to whom they have
sometimes negates the experience. People with entrusted their lives is not supportive?
cancer complain that people are trying to get them Another issue to take into consideration is in-
to be optimistic and cheerful at times when they do dividual difference variables. Does everybody
not want to have those feelings, when they want to benefit from support in the same way? I am going
be able to express other feelings. People in one’s to describe three individual difference variables.
social environment may change the subject when One is attachment style. This is an individual dif-
one wants to talk about problems because they ference variable that comes from the develop-
have the perception that it is bad for someone to mental literature. It characterizes infants’ relation-
talk about their problems, or that it is counter- ships with their caretakers. Individual differences
productive to dwell on negative feelings. People in infant–caretaker attachment have been trans-
also may avoid the person completely. Network ferred to the adult literature to characterize adult
members may be uncomfortable because they do romantic relationships. People with secure at-
not know what to do or say. They do not know tachment styles believe their partner is responsive
how to respond to the person who is facing the to their needs. There are two insecure attachment
stressor, so they ignore the person. styles in which the person believes a partner is not
People in one’s social environment also may responsive. One insecure attachment style is called
provide too much support. I found evidence of this avoidance, which is a type of detachment from
in the study of cardiac patients [3]. I found that relationships. An avoidant attachment style is
people sometimes felt that their family and friends characterized by a general mistrust of others, a
provided too much informational support. This fear of closeness, and detachment – possibly be-
probably reflected unwanted advice. cause these people were rejected in the past. An
Negative interactions are not the opposite of anxious attachment style is characterized by a
positive interactions. They are typically uncorre- preoccupation with relationships and a desire for
lated. In studies of quality of life, negative inter- extreme closeness. In the past, these people prob-
actions are often a stronger predictor of quality of ably learned that they cannot depend on others to
life than positive interactions. Why? One possi- be there all of the time.
bility is that negative interactions are not as fre- In a study conducted by Carpenter and Kirk-
quent as positive interactions. Negative interac- patrick [8], college students came to the lab, either
tions are salient and unexpected, so they have a without or with their romantic partner. They were
stronger impact on us. There also may be ceiling led to believe that they were going to experience a
effects on measures of positive interactions. When stressful event. Carpenter and Kirkpatrick com-
measuring the functional aspects of support, it is pared how people who had a secure vs. an insecure
not uncommon for people to report extremely high (either avoidant or anxious) attachment style re-
levels of perceived support. There may not be sponded to stress with and without the presence of
much of a distribution on some of the positive their romantic partner. Insecure respondents
aspects of support. Thus, there is more variability showed an increase in blood pressure and heart
when measuring negative interactions. My own rate when their partner was present compared to
29
absent. Thus, the insecure people were actually influence quality of life is by altering how we ap-
harmed in a sense by having their partner present. praise a stressful event. Informational support or
The presence of a support person certainly did not emotional support (reassurance) can lead us to
buffer them from the distress associated with the perceive that the stressor is not as bad as originally
stressful task. believed. Once stress levels are lowered, support
A second individual difference variable is cynical resources may help us more effectively cope with
hostility, which is a construct that has been in- the stressor. People may provide information about
vestigated by Lepore [9] in a laboratory study with concrete things that can be done to exert control
a similar paradigm. In Lepore’s study, people ac- over the stressor. If people provide assistance by
tually underwent a stressful task. Lepore divided running errands and cleaning the house, the indi-
people into those who were high in cynicism and vidual’s quality of life may be enhanced. Both
those who were low in cynicism. In one condition functional and structural aspects of support influ-
they gave a speech alone, and in another condition ence health behavior, which could enhance quality
they gave a speech in the presence of a supportive of life. The relationships we have provide models
confederate. The supportive confederate repre- for us in terms of risk behaviors, such as smoking
sented emotional support. The confederate was and drinking. In terms of the functional aspects of
taught to nod, agree, and make a few reassuring support, the people in your social environment
remarks. Lepore found that people who were low might encourage good health behavior and
in cynicism had lower reactivity in the support encourage you to go to the doctor in a timely
condition than in the no support condition. Low manner.
cynical people benefited from the presence of a The last issue I want to briefly discuss is how to
supportive confederate, whereas cynical people did translate the correlational research on social sup-
not. This study suggests that people who are cyn- port and quality of life into the field of support
ical may not benefit from support. interventions. First, you need to ask what kind of
The third individual difference variable is gen- support you should provide in your intervention?
der. There are more mixed effects of support on The descriptive and correlational research might
women’s quality of life than men’s quality of life. lead you to believe that emotional support is most
Woman have more support available to them than important. However, if you examine the interven-
men which should be related to an enhanced tion research, it appears that there are stronger
quality of life, but women also are more likely to effects for informational support than emotional
provide support to others than men which could support (see Ref. [10] for a review). You also might
be burdensome. Both men and women turn to want to ask who should provide the support in the
women for support. Providing support can be a intervention? Should the support come from ex-
tax on our resources, which might result in a re- isting network members or strangers? That is,
duced quality of life. This may be one explanation should you intervene by altering people’s existing
as to why the support findings for women are more network, or should you provide them with a new
complicated than they are for men. network? What are the advantages and disadvan-
One of the main focuses of this article has been to tages of each?
help researchers determine what aspects of social The advantage of altering the existing network is
relationships or what types of support need to be that these are the most critical support persons.
measured. Researchers also need to consider the The disadvantage is that it is not going to be an
mechanisms by which support might influence easy thing to do because it is difficult to change
quality of life. The mechanism might differ for the people’s existing social relations. People in the
structural and functional measures. First, let’s network are often involved with the stressor, so
consider mechanisms that might explain why struc- they might also be in need of their own support.
tural support enhances quality of life. The mere What about creating a new network of support
existence of social relationships can put us in a bet- providers for people? This is the premise of the
ter mood, provide us with a sense of identity, and be peer support group. The advantage of the peer
a source of companionship to share activities. support group is that group members are experi-
One way that functional measures of support might encing the same stressor and may understand the
30
situation better than network members. There homogenous group of caregivers was studied – all
should be less of a chance that unintended negative spouses.
interactions would occur among peers. There are A review of the literature for new parents has
also many disadvantages of peer support groups. shown few benefits of peer discussion group in-
It is not possible to control the nature of the in- terventions [17]. Again, one exception is a study
teractions that occur in a support group, so there is that included a longer-term follow-up [17]. It was
no guarantee that the support provided will be not until 18 months that improvements in the in-
helpful. Another concern about support groups is tervention relative to the control group occurred.
that they might lead network members to feel ex- Finally, reviews of the literature on divorced peo-
cluded. ple [18] and victims, largely rape victims [19], have
Recently, I reviewed the support group inter- shown no evidence for the benefits of peer support
vention literature in a number of areas – cancer, groups on quality of life.
caregivers of Alzheimer’s patients, new parents, One is more likely to find positive effects of
divorce, and bereavement [11]. All of these litera- peer support group interventions if: (1) informa-
tures reach the same conclusion. There is little tional support from experts is included, (2) emo-
evidence for the effectiveness of peer discussion tional support groups are of longer duration, and
group interventions. In the cancer literature, (3) emotional support groups are really group
support interventions often take one of two forms – therapy guided by clinicians.
emotional support provided by peers or informa- These are some of the issues that need to be
tional support provided by experts. There is more considered when studying the social environment
evidence for the benefits of informational support or the effect of social support on quality of life. Is
than emotional support. There are a couple of there a stressor? Do you want to measure the
exceptions. A study of women with metastatic structural aspects of support, or do you want to
breast cancer found survival benefits after 1 year measure the functional aspects of support? Who
of meetings [12]. A study of men who were HIV+ are the important sources? Which functions are
found more benefits from a peer support group most important? Are you interested in people’s
than a skills training group [13]. One thing that perceptions of support or whether they actually
both of these studies have in common is that the received support? Do you want to measure the
prognosis of the patients attending the group was negative aspects of the social environment? Do
poor. It is possible that emotional support is more you want to take into consideration whether par-
beneficial under more severe threat conditions or ticipants in your study are also providing support
when the stressor is less controllable. When the to these same network members? What are the
stressor is more controllable – when there are processes by which you expect support to be re-
things that you can do to control the stressor – lated to quality of life? Will everybody benefit or
informational support in the form of education do you want to select the people you think will
may be more helpful. benefit? Finally, do you want to intervene with
Two reviews of the literature on caregivers, one people’s existing relationships or create new rela-
of which is caregivers of Alzheimer’s patients, have tionships?
shown little effect of peer support groups [14, 15].
One notable exception is a study by Mittelman
et al. [16] in which people were assigned to attend
References
a series of individual counselling, family counsel-
ling, and then group discussions. The multiple 1. Cohen S, Willis TA. Stress, social support, and the buffer-
components of this intervention make it difficult ing hypothesis. Psychol Bull 1985; 98: 310–357.
to discern the effect of the group discussion. The 2. Cutrona CE. Stress and social support – in search of opti-
intervention also was of longer duration and mal matching. J Soc Clin Psychol 1990; 9: 3–14.
included a long-term follow-up. In fact, the dif- 3. Helgeson VS. Two important distinctions in social support:
Kind of support and perceived vs. received. J Appl Soc
ferences between the intervention and control Psychol 1993; 23: 825–845.
group were not significant until 8 months after the 4. Jacobson DE. Types and timing of social support. J Health
study began. The group also is unique in that a Soc Behav 1986; 27: 250–264.
31
5. Mickelson KD, Helgeson VS, Weiner E. Gender effects on depressed, HIV infected persons. Am J Psychiatry 1993;
social support provision and receipt. Pers Relationships 150(11): 1679–1686.
1995; 2: 211–224. 14. Bourgeois MS, Schulz R. Intervention for caregivers of
6. Peters-Golden H. Breast cancer: Varied perceptions of so- patients with Alzheimer’s disease: A review and analysis of
cial support in the illness experience. Soc Sci Med 1982; 16: content, process, and outcomes. Int J Aging Hum Dev
483–491. 1996; 43(1): 35–92.
7. Dakof GA, Taylor SE. Victim’s perceptions of social sup- 15. Lavoie JP. Support groups for informal caregivers don’t
port: What is helpful from whom? J Pers Soc Psych 1990; work! Refocus the groups or the evaluations? Can J Aging
58: 80–89. 1995; 14(3): 580–595.
8. Carpenter EM, Kirkpatrick LA. Attachment style and 16. Mittelman SM, Ferris SH, Shulman E, Steinberg G, Am-
presence of a romantic partner as moderators of psycho- binder A, Mackell JA. A comprehensive support program:
physiological responses to a stressful laboratory situation. Effect on depression in spouse-caregivers of AD patients.
Pers Relationships 1996; 3: 351–367. Gerontologist 1995; 35(6); 792–802.
9. Lepore SJ. Cynicism, social support, and cardiovascular 17. Cowan CP, Cowan PA. A preventative intervention for
reactivity. Health Psychol 1995; 14(3): 210–216. couples becoming parents. In: Boukydis CFZ (ed.), Re-
10. Helgeson VS, Cohen S. Social support and adjustment search on Support for Parents and Infants in the Postnatal
to cancer: Reconciling descriptive, correlational, and Period, New York: Ablex, 1986; 225–251.
intervention research. Health Psychol 1996; 15: 135– 18. Hughes R Jr. Divorce and social support: A review. J Di-
148. vorce 1988; 11(3/4): 123–145.
11. Helgeson VS, Gottleib BH. Support groups. In: Cohen S, 19. Coates D, Winston T. Counteracting the deviance of de-
Underwood LG, Gottlieb BH (eds), Social Support Mea- pression: Peer support groups for victims. J Soc Issues 1983;
surement and Intervention: A Guide for Health and Social 39: 169–194.
Scientists, New York: Oxford University Press, 2000; 221–
245.
12. Spiegel D, Bloom JR, Kraemer HC, Gottheil E. Effect of Address for correspondence: Vicki S. Helgeson, Department of
psychosocial treatment on survival of patients with meta- Psychology, Carnegie Mellon University, Pittsburgh, PA 15213,
static breast cancer. Lancet 1989; 2: 888–891. USA
13. Kelley JA, Murphy DA, Bahr R, et al. Outcome of cogni- Phone: þ1-412-268-2624
tive-behavioral and support group brief therapies for E-mail: vh2e+@andrew.cmu.edu