Journal of Counseling Psychology Copyright 1989 by the American Psychological Association, Inc.
1989, Vol. 36, No. 2, 170-175 0022-0167/89/$00.75
Social Support and the Effectiveness of Group Therapy
Division of Psychological and Quantitative Foundations
University of Iowa
Collected self-report data from 37 pretest clients and 26 of the same clients at posttest who
participated in six counseling center therapy groups. Questionnaires assessed six functionally
different types of social support provided from two sources, therapy group members vs. persons
outside the group, together with pre- and posttest levels of three distress symptoms, which were
depression, self-esteem, and psychological symptoms of stress. Significant improvement in
symptoms was noted during the eight-week interventions, and this improvement was related to
the availability of social support, depending on the type and source of support. In general,
support from sources outside the therapy group appeared to have the most impact. Levels of
certain types of support differed in groups depending on whether or not the group was composed
of members with a common presenting concern.
Social support is an important coping resource for persons
experiencing stressful life changes (see Cohen & Wills, 1985;
Leavy, 1983, for reviews). Mutually exchanged social support
is also an important aspect of theme-oriented therapy groups
that bring together persons who have experienced the same
trauma, life transition, or other presenting problem, for ex-
ample, victims of incest (Hall, Kassees, & Hoffman, 1986),
military veterans (Gressard, 1986), or persons coping with
bereavement (R. Schwab, 1986). In addition, social support
has been identified as contributing to the effectiveness of
general process therapy groups composed of members with a
wide variety of presenting concerns (Yalom, 1985). However,
little is known about how social support specifically benefits
members of either type of group.
Studies of social support in nontherapeutic contexts have
typically found a modest relation between perceived support
and lower levels of stress symptoms, but recent writers have
criticized many of these studies for assessing social support
with one omnibus measure and measuring stress with a single
index of accumulated life changes (cf. Thoits, 1985). Advo-
cates of the specificity hypothesis contend that social support
is multidimensional and functions in a stressor-specific man-
ner. Different types of support provide different coping re-
sources, and because stressors vary in adaptational demands,
a given type of support will be effective only when the coping
resources it provides are matched to the demands of the
stressor (Cohen & Wills, 1985; Wilcox & Vernberg, 1985).
Unidimensional indicators of support fail to capture this
I gratefully acknowledge Jackie Engle, Greg Keilin, Judy Naginey,
Nancy Downing, Dan Dworkin, Alan Farber, Beth Firestein, Susan
MacQuiddy, Ann Naplin, Randy Swaim, and Patty Vigil, who served
as group leaders or provided other data for this project, and Charles
Claiborn and Carolyn Cutrona, who reviewed an earlier version of
Correspondence concerning this article should be addressed to
Brent Mallinckrodt, who is now at the Division of Counseling and
Educational Psychology, College of Education, University of Oregon,
Eugene, Oregon 97403-1215.
Evidence supporting the specificity hypothesis has accu-
mulated from studies that used multidimensional measures
of social support, For example, Cutrona and Russell (1987)
developed a six-factor measure of social support based on a
model of social provisions, which are forms of emotional or
tangible assistance obtained from relationships with others.
This model was originally conceived by Weiss (1974) as a
theory of loneliness. Subsequent studies have found that not
only does the perceived lack of a particular type of support
(social provision) lead to differences in subjective feelings of
loneliness as Weiss predicted (Russell, Cutrona, Rose, &
Yurko, 1984) but types of support also appear to differ in
usefulness as coping resources depending on the particular
stressor. For example, reassurance of worth support provides
affirming feedback about competence and abilities and was
most closely related to older workers' successful coping with
job loss (MaUinckrodt & Fretz, 1988) and to preventing
burnout among nurses (Constable & Russell, 1986) and class-
room teachers (Russell, Altmaier, & Van Velzen, 1987). How-
ever, this form of support was not as helpful to first-time
mothers as was guidance support, which involves the availa-
bility of a confidant who may provide authoritative advice
(Cutrona, 1984). Reliable alliance support, the perception
that someone can always be counted on to provide assistance
in an emergency, was crucial for elderly persons living in rural
areas (Chwalisz, Russell, Cutrona, & Mallinckrodt, 1988).
Research has also suggested that social support varies in
effectiveness depending on its source (Gottlieb, 1978; Wilcox
& Birkel, 1983). For example, support from co-workers or
supervisors was found to moderate the effects of occupational
stress, but support from spouses or friends outside the work
setting was much less effective (La Rocco & Jones, 1978).
These findings are relevant in considering sources of social
support available to members of therapy groups, who ex-
change mutual support but, particularly in general process
groups, are also encouraged to gain interpersonal skills to help
them mobilize support from sources outside the group (Ya-
lom, 1985). In fact, the most important long term therapeutic
gains may be related to increased support from these sources
in the natural social environment (Brown, Brady, Lent, Wol-
SOCIAL SUPPORT AND GROUP THERAPY 171
fert, & Hall, 1987; Rook, 1984). Conversely, for members of
theme groups who have experienced the same life stresses,
mutually exchanged support may be more effective than the
same type of support provided by persons outside the group
who have not experienced the stressor.
Information about which types and sources of social sup-
port are most beneficial to therapy group members may help
to increase therapeutic effectiveness of the groups, but no
previous study that investigated these factors could be located.
Thus, the purpose of this study was to identify the specific
source of support, group members versus sources outside the
group, and specific types of support that are most closely
associated with positive changes in psychological symptoms
of stress, including depression, self-esteem, and global psycho-
logical symptoms. In addition, because support processes may
differ depending on the nature of the therapy group, this study
examined differences in support for members of theme-ori-
ented and general process therapy groups.
Participants and Procedure
Clients assigned to group therapy at a university counseling center
located in the Rocky Mountain states participated in the study. Their
cooperation was solicited by group leaders during the first therapy
session. Pretest materials were distributed by group leaders with a
request to complete the packets at home and return them to the
counseling center receptionist before the next group meeting. Clients
were assured that their decision about participation would not affect
the counseling services they received and that group leaders and the
researcher would remain unaware of which clients chose not to
participate. Of 48 clients who were members of the six therapy groups
run during the data collection period, 37 (77%) chose to complete
Of the original 48 clients, 38 (79%) continued in therapy until
their groups were terminated. During the last session of each group,
posttest materials were distributed with a request to return materials
by mail in prestamped envelopes. Clients labeled materials for both
pre- and posttest packets only with their mothers' maiden names. In
this way anonymity was maintained while allowing the matching of
pre- and posttest data. Unfortunately, this procedure made it impos-
sible to identify clients who did not complete the posttest packets
after the initial distribution and to mail follow-up materials.
Of the original 37 clients who completed pretests, 26 (70%) chose
to return posttests. Procedures used to assure anonymity made it
impossible to determine how many of the 11 clients who did not
complete posttest materials were, in fact, among the 10 clients no
longer in therapy at posttest. For the 26 pre-post clients, the mean
interval between completion dates of the pretest and posttest materials
was 7.8 weeks, Clients in theme groups comprised 62% of the pretest
participants (23 from a total of 37) and 62% of the pre-post partici-
pants ( 16 from a total of 26). Pre-post participants included 18 (69%)
women and 8 (31%) men. Their mean age was 24.9 years (SD = 5.5).
Two clients reported a Hispanic ethnic identification, and the re-
maining 24 (92%) indicated that they were White.
To sample the widest possible range, all six therapy groups offered
by the counseling center during the semester of data collection were
included in the study. All were closed membership, time-limited,
semester-length interventions that met once per week for approxi-
mately 90 rain. None of the groups were structured interventions,
although some theme groups involved occasional didactic instruction.
All groups were led by two cofacilitators (two female-female pairs,
one male-male pair, and three female-male pairs). Three persons led
two different groups. Of the nine counselors who served as coleaders,
two were senior staff members, two were interns, and five were
advanced graduate students. Four of the six groups were theme
oriented; that is, clients were assigned on the basis of a cluster of
similar presenting concerns. In one group the clients were nontradi-
tional-age students; in another the clients had experienced the ending
of a romantic relationship; another was a women's support group;
and the fourth was an eating-disorders group for women. The re-
maining two were general process groups in which clients had a
variety of concerns.
At both pre- and posttest the participants completed a measure of
perceived social support and three symptom measures of self-esteem,
depression, and psychological symptoms of stress. These particular
symptoms were selected because they represented a broad range of
potential client changes in the diverse groups studied.
Social support. The Social Provisions Scale (SPS; Cutrona &
Russell, 1987; Russell & Cutrona, 1984) is a 24-item measure of
perceived social support. Clients respond on a 4-point Likert scale (1
= strongly disagree to 4 = strongly agree). Six subscales, composed
of four items each, are used to assess different social support provi-
sions. Labels for the six types of support, the functional psychological
needs each is believed to satisfy, and representative items from the
SPS subscale are: (a) attachment, feelings of safety and security in a
close emotional bond, "I lack a feeling of intimacy with another
person"; (b) social integration, interests and concerns are shared by
others, "I feel a part of a group of people who share my attitudes and
beliefs"; (c) reassurance of worth, having skills and abilities acknowl-
edged, "there are people who admire my talents and abilities"; (d)
reliable alliance, assurance that one can count on assistance being
available if needed, "there are people I can count on in an emergency";
(e) guidance, availability of confidants or authoritative others to
provide advice, "there is no one I feel comfortable talking about
problems with"; and, (f) opportunity for nurturance, the sense of being
needed in vital ways by other persons, "there is no one who really
relies on me for their well-being."
Russell and Cutrona (1984) reported internal consistency (coeffi-
cient alpha) for each of the subscales ranging from .76 to .84 in a
sample of older adults and from .61 to .76 in a sample of teachers.
They reported test-retest reliabilities ranging from .37 to .66 for the
subscales and a total scale test-retest reliability of.59. A confirmatory
factor analysis resulted in a goodness-of-fit index of .86 for the
subscales, which indicates a fairly good fit of the data to Russell and
Cutrona's six-factor model.
At pretest the clients in this study were instructed to consider "all
current relationships" in completing the SPS. At posttest clients
completed one copy to assess their perceptions of support only from
group members and a second copy to assess support from "all current
relationships excluding group leaders and group members."
Self-esteem. The Rosenberg Self-Esteem Scale (Rosenberg, 1965)
consists of 10 items scored on a 4-point Likert scale (1 = strongly
disagree to 4 = strongly agree). Internal consistency (coefficient alpha)
of .81 has been reported (Mallinckrodt & Fretz, 1988). A test-retest
reliability of.85 was obtained after a 2-week interval (Silber & Tippett,
1965), and the scale correlated well with other measures of self-esteem
and with clinical assessments (correlations ranging from .56 to .83).
172 BRENT MALLINCKRODT
The Beck Depression Inventory (BDI) is a widely
used self-report measure of depression consisting of 21 items, each
item containing four self-report statements representing a cluster of
symptoms (Beck, Ward, Mendelson, Mock, & Erbaugh, 1961). The
statements reflect different cognitive-bahavioral levels of depression
and are scored on a 0-3 scale, with higher scores indicating more
The Bell Global Psychopathology Scale
(BGPS; J. J. Schwab, Bell, Warheit, & Schwab, 1979) consists of 33
self-report items yielding eight subscale scores of psychological symp-
toms of stress (e.g., depression, anxiety, obsessive thoughts, phobias,
and hallucinations). Although J. J. Schwab et al. analyzed each of the
BGPS subscales individually, in order to limit the number of analyses
in this study, only the total BGPS score was used as a global indicator
of psychological distress. Participants respond considering their "ex-
perience of the past month or so" using 5-point (1 =
to 5 =
or 2-point scales (1 =
or 2 =
The BGPS has
demonstrated internal consistency (coefficient alpha) and test-retest
reliability correlations all greater than .80, and the total BGPS score
was significantly correlated (r = .74) with a widely used measure of
psychopathology, the Health Opinion Survey (J. J. Schwab et al.,
A one-way multivariate analysis of variance
used to compare the 26 pre-post clients and the 11 pretest
only clients regarding pretest levels of each of the research
variables. No significant differences were found, F(1, 35) =
2.05, p > .05. Consequently, all 37 clients were included in
pretests analyses as the most representative sample for deter-
mining the relation of social support to symptom variables at
the point clients begin therapy. Table 1 presents a correlation
matrix of the relations among the six types of support and
three symptom variables. Results indicated that different types
of support varied in the magnitude of their negative relation
to symptom levels and that reliable alliance, attachment, and
reassurance of worth seemed to exhibit the highest negative
correlations with symptom measures.
A one-way, repeated measures MANOVA was used to com-
pare pre- and posttest levels of symptoms and social support
from sources outside the group for the 26 pre-post clients.
Significant overall pre-post differences were obtained, F(8,
18) = 304.7, p < .001. The results of repeated measures t tests
used as univariate follow-ups are shown in Table 2. During
the course of therapy clients exhibited significant gains in
total support and several specific types of support from sources
outside the group. There were also significant improvements
in clients' self-esteem, depression, and BGPS scores.
The effects of support from different sources, group mem-
bers versus persons outside the group, are compared in Table
3. Changes in symptoms are represented by partial correla-
tions of support with posttest symptoms, controlling for pre-
test symptom levels. In general, support from outside sources
appeared to be more closely associated with positive changes
in symptoms than support from group members. For exam-
ple, total support from persons outside the group was related
to improvement in self-esteem (partial r = .64, p < .01) and
to decreased levels of depression (partial r = -.48, p < .01),
whereas total support from group members was not (partial
rs = . 11 and. 17, respectively). An unexpected finding was
that opportunity for nurturance support from other group
members was positively correlated with depression.
Finally, a one-way MANOVA was used to compare members
of theme groups with members of general process groups
regarding support from group members and from sources
outside the group. The analysis indicated significant overall
differences, F(6, 19) = 6.14, p < .001. Univariate follow-ups
indicated no differences in support from outside sources, but
two types of support from other group members were more
available to clients in theme groups, guidance support, F(1,
24) = 10.67, p < .01, and reliable alliance support, F(1, 24)
= 8.57, p < .01, with a trend toward more available attach-
ment support, F(1, 24) = 3.91, p < .06. Tests of demographic
differences indicated that there was no difference in gender
distribution between the two types of groups, but members of
theme groups were significantly older (M = 26.6 years,
Pretest Social Support and Symptom Intercorrelations
Variable 1 2 3 4 5 6 7 8 9 10
1. Total support -- .88 .90 .89 .70 .71 .67 -.23 -.74 .69
2. Reliable alliance -- .81 .73 .52 .61 .51 -.22 -.68 .57
3. Attachment -- .77 .66 .48 .48 -.29 -.77 .73
4. Guidance -- .76 .49 .43 -.14 -.57 .56
5. Opportunity for nurtur-
ance -- .18 .16 -.08 -.37 .30
6. Social interaction -- .72 -.15 -.58 .43
7. Reassurance of worth -- -.20 -.57 .64
8. BGPS --
9. Depression -- -.81
M 72.8 13.3 11.4 13.2 10.2 12.7 12.1 44.2 14.9 27.7
11.3 1.9 3.3 2.5 2.2 2.1 2.1 16.7 9.5 6.2
n = 37. BGPS = Bell Global Psychopathology Scale. For correlations greater in absolute value than .28, p < .05; for correlations greater
in absolute value than .42, p < .01.
Higher scores indicate more positive self-esteem.
SOCIAL SUPPORT AND GROUP THERAPY 173
Pretest to Posttest Changes in Symptoms and Support
M SD M SD t a
27.4 6.5 29.2
16.0 9.7 12.5
43.3 12.0 37.2
Social support from sources outside the group
Reliable alliance 13.2 1.7 13.8 1.8 1.93t
Attachment 10.9 3.3 12.1 3.4 4.95**
Guidance 12.8 2.5 13.3 2.8 1.73
Opportunity for nurturance 9.5 2.1 10.3 2.2 2.30*
Social integration 12.8 1.9 13.3 1.5 1.86
Reassurance of worth 12.1 2.3 12.7 2.0 2.42*
Total social support 71.4 11.0 75.5 11.4 4.10"*
Note. n = 26. BGPS = Bell Global Psychopathology Scale.
df = 25, repeated measures t tests, b Higher scores indicate more
*p < .05. **p < .01.
t P < .07.
6.2) than general process groups (M = 22.2 years, SD = 2.8),
t(24) = 2.11, p < .05.
Results of this study provide support for some elements of
the specificity hypothesis. Ten of the 15 correlations among
the six types of support were less than .70 at pretest, which
suggests a multidimensional structure for the social support
construct. At pretest different types of support also varied in
their relation to different symptoms, which suggests that each
provides somewhat different coping resources.
Furthermore, pre-post comparisons indicated significant
improvements in all three symptom measures for therapy
group participants, although these changes cannot be unam-
biguously attributed to effects of the therapy, because there
was no untreated waiting-list group of clients available for
comparison. Interestingly, support from persons outside the
groups also improved significantly from pretest to posttest.
Most notably, attachment support (close emotional bonding)
increased, and it was the lack of this support that was most
closely associated with depression and low self-esteem at
pretest. Perhaps a significant proportion of the improvement
in symptoms was due to gains in beneficial support from
sources outside the group. If so, this finding would support
the therapeutic strategy of helping clients to acquire interper-
sonal skills for mobilizing support in their natural social
environment (Brown et al., 1987; Rook, 1984).
The primary purpose of this study was to identify the
sources and types of social support most closely related to
positive changes in specific stress symptoms. In general, it
appears that social support from persons outside the therapy
group may be more beneficial than support from other group
members. Of the 18 correlations between the six types of
outside support and the three symptoms measures, l0 were
significant, whereas only three of the 18 correlations for group
member support were significant, and opportunity for nur-
turance support from group members was positively corre-
lated with depression (see Table 3).
Thus, these findings suggest different effects for specific
types of support depending on the source. For example,
attachment support (close emotional bonding) and reliable
alliance (others can be counted on for help in an emergency)
may be related to improvement in self-esteem and depression
but only if this support is provided through relationships with
persons outside the therapy group. However, reassurance of
worth support (affirmation of competencies and abilities)
seems to have a beneficial effect on self-esteem if provided by
either source. Only one support-symptom relation showed a
stronger negative correlation for group sources than for out-
side sources, namely reliable alliance and BGPS scores.
Among the symptoms measured by the BGPS are anxiety,
phobias, and obsessive thoughts. Perhaps the perception that
group members can be counted on to provide help in an
emergency may be more effective in reducing these symptoms
because clients believe that other group members are,more
able (compared with those outside the group) to understand
their anxieties and help them cope with potential crises.
It is difficult to explain the finding that opportunity for
nurturance support from group members was positively cor-
related with depression. Cutrona and Russell (1987) pointed
out that "strictly speaking, this [opportunity for nurturance]
cannot be considered social support, in that the individual is
the provider rather than the recipient of assistance. However,
... giving and receiving help may enhance health through
Correlations of Support From Therapy Group Members
(Group) and From Relationships Outside the Group
(Outside) with Change in Symptoms
Type of support and source Self-esteem a Depression BGPS
Group .11 .17 -.32
Outside .64** -.48** -.26
Group -.04 .23 -.57**
Outside .53** -.36* -.33
Group .02 .19 -.25
Outside .60** -.39* -. 16
Group -. 13 -. 14 -.29
Outside .52** -.44* -.31
Opportunity for nurturance
Group -.27 .38* .10
Outside -.20 -.23 -.07
Group .57** -.10 -.01
Outside .61"* -.10 -.50**
Reassurance of worth
Group .67** .07 -.31
Outside .80** -.66** -.16
Note. n = 26. BGPS = Bell Global Psychopathology Scale. Partial
correlations are of social support with posttest symptoms, controlling
for level of symptoms at pretest.
Higher scores indicate more positive self-esteem.
*p < .05. **p< .01.
174 BRENT MALLINCKRODT
some of the same cognitive mechanisms" (p. 42). Research
suggests that the opportunity to nurture others does indeed
furnish the provider of this assistance with the same types of
psychological coping benefits as other, more direct, forms of
social support (Mallinckrodt & Fretz, 1988). However, in this
study depression was correlated with increased perceptions by
clients that other group members were dependent on them
for their sense of well-being. Perhaps some clients perceived
the dependency needs of some other group members as over-
whelming. Some clients, especially women, may also invest
increased efforts in nurturing others while their own psycho-
logical well-being suffers (Gilligan, 1982; Norwood, 1985). In
any case, this finding suggests that not all types of support
from group members may be beneficial.
The final purpose of this study was to compare support
perceived by members of theme-oriented and general process
therapy groups. The results suggest that clients in both types
of groups had equivalent perceptions of support from sources
outside the group. However, regarding support from group
members, participants in theme groups apparently perceived
greater availability of guidance support (available confidants
or authoritative advice) and reliable alliance support (others
can be counted to help in an emergency), with a trend toward
greater attachment support.
Guidance and reliable alliance support often involve direct
offers of advice or help. Research has shown that such offers,
no matter how well intentioned, may sometimes have a
negative impact on persons experiencing a traumatic life
event, especially if the offer is unsolicited and made by persons
who have not experienced a similar trauma (Lehman, Ellard,
& Wortman, 1986). However, direct advice and other forms
of tangible assistance may be welcomed and sought from
persons who have experienced the same stressor, even though
individuals vary greatly in their ability to actually profit from
this assistance (Gottlieb, 1983). Relative to members of gen-
eral process groups, perhaps because of assumed shared ex-
periences, members of theme groups may have felt a greater
willingness to confide and accept direct advice from one
another, or they may have perceived more helpful potential
The trend toward higher levels of attachment support (a
sense of close emotional bonds) between members of theme
groups may reflect what Parson (1985) terms "post-traumatic
accelerated cohesion," the rapid and premature development
of cohesion in groups of persons who have experienced the
same traumatic event. Parson believes this premature cohe-
sion is often countertherapeutic because it is based only on
superficial characteristics of group members. This early sense
of"we-ness" tends to delay the healthy exploration of conflict
and anxiety and may interfere with the development bonds
based on deeper aspects of personality. Whatever the thera-
peutic effects, the trend observed in this study suggests that
members of theme groups may have perceived a closer at-
tachment bond with one another than members of general
Differences in theme versus process groups may have also
been due to the occasional didactic instruction given in theme
groups, although in rating support within the groups, clients
were instructed to exclude support from group leaders. A
further complicating factor may have been contacts between
group members that occurred outside the group sessions.
Leaders of general process groups warned clients against such
contacts, whereas leaders of theme groups were mixed in
whether or not they gave such a warning. In several instances
outside contacts were known to have occurred between mem-
bers of theme groups.
Findings of this study must be interpreted cautiously and
regarded only as a pilot exploration of social support in
therapy groups because several factors may limit the general-
izability of results. In studying actual clients in therapy,
compromises in design were required to ensure the quality of
clients' therapeutic experience, to protect clients from undue
pressure to participate in this study, and to maintain client
anonymity. For example, multiwave survey mailings and
vigorous follow-up tracking procedures to achieve higher re-
turn rates were not possible. Consequently, only 77% of the
clients chose to participate initially, and only 70% of these
completed posttest materials. Although a 20% attrition rate
is considered typical for many therapy groups, the subsample
in this study who completed pre- and posttest materials may
have been different from other clients in important respects.
Moreover, the final sample size of 26 is small and further
In analyses comparing general process groups to theme
groups data from very different theme groups were combined.
This was a practical necessity because individual groups con-
tained as few as 3--4 participating clients. Thus, in testing the
specificity hypothesis, only the requirement to use a multidi-
mensional measure of support was met. The requirement of
assessing specific stressors must await studies conducted over
several years to accumulate a larger number of participants
in the same theme group. These problems notwithstanding,
finding differences in social support from group members
when comparing very diverse combinations of theme and
general process groups suggests a robust phenomenon that
needs further exploration.
The findings of this study, if replicated in future research,
do suggest that both the source of social support, type of social
support, and type of support group all may have important
implications for therapeutic change. Not every type of support
from every source may be beneficial. Further empirical inves-
tigation of group member support processes may suggest ways
facilitators can maximize the availability of the most helpful
types of social support.
Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J.
( 1961). An inventory for measuring depression.
Archives of General
Brown, S. D., Brady, T., Lent, R. W., Wolfert, J., & Hall, S. (1987).
Perceived social support among college students: Three studies of
the psychometric characteristics and counseling uses of the Social
Support Inventory [Monograph].
Journal of Counseling Psychol-
Chwalisz, K., Russell, D. W., Cutrona, C. E., & Mallinckrodt, B.
Qualitative versus quantitative social support and depres-
sion in the elderly.
Unpublished manuscript, Division of Psycho-
SOCIAL SUPPORT AND GROUP THERAPY 175
logical and Quantitative Foundations, University of Iowa, Iowa
Cohen, S., & Wills, T. A. (1985). Stress, social support, and the
buffering hypothesis. Psychological Bulletin, 98, 310-357.
Constable, J. F., & Russell, D. (1986). The effect of social support
and the work environment upon burnout among nurses. Journal
of Human Stress, 12, 20-26.
Cutrona, C. E. (1984). Social support and stress in the transition to
parenthood. Journal of Abnormal Psychology, 93, 378-390.
Cutrona, C. E., & Russell, D. W. (1987). The provisions of social
relationships and adaptation to stress. In W. H. Jones & D. Perlman
(Eds.), Advances in personal relationships (Vol. l, pp. 37-67).
Greenwich, CT: JAI Press.
Gilligan, C. (1982). In a different voice." Psychological theory and
women "s development. Cambridge, MA: Harvard University Press.
Gottlieb, B. H. (1978). The development and application of a classi-
fication scheme of informal helping behaviors. Canadian Journal
of Behavioral Science, 10, 105-115.
Gottlieb, B. H. (1983). Social support strategies: Guidelines for mental
health practice. Beverly Hills, CA: Sage.
Gressard, C. F. (1986). Self-help groups for Vietnam veterans expe-
riencing posttraumatic stress disorder. Journal for Specialists in
Group Work, 11, 74-79.
Hall, R. P., Kassees, J. M., & Hoffman, C. (1986). Treatment for
survivors of incest. Journal for Specialists in Group Work, 11, 85-
La Rocco, J. M., & Jones, A. P. (1978). Co-worker and leader support
as moderators of stress-strain relationships in work situations.
Journal of Applied Psychology, 63, 629-634.
Leavy, R. L. (1983). Social support and psychological disorder: A
review. Journal of Community Psychology, 11, 3-21.
Lehman, D. R., Ellard, J. H., & Wortman, C. B. (1986). Social
support for the bereaved: Recipients' and providers' perspectives
on what is helpful. Journal of Consulting and Clinical Psychology,
Mallinckrodt, B., & Fretz, B. R. (1988). Social support and the impact
of job loss on older professionals. Journal of Counseling Psychology,
Norwood, R. (1985). Women who love too much. New York: Pocket
Parson, E. R. (1985). Post-traumatic accelerated cohesion: Its recog-
nition and management in group treatment of Vietnam veterans.
Group, 9, 10-23.
Rook, K. S. (1984). Promoting social bonding: Strategies for helping
the lonely and socially isolated. American Psychologist, 39, 1389-
Rosenberg, M. (1965). Society and the adolescent self-image. Prince-
ton, N J: Princeton University Press.
Russell, D. W., Altmaier, E., & Van Velzen, D. (1987). Job-related
stress, social support, and burnout among classroom teachers.
Journal of Applied Psychology, 72, 269-274.
Russell, D., & Cutrona, C. E. (1984, August). The provisions of social
relationships and adaptation to stress. Paper presented at the annual
meeting of the American Psychological Association, Toronto, On-
Russell, D., Cutrona, C. E., Rose, J., & Yurko, K. (1984). Social and
emotional loneliness: An examination of Weiss's typology of lone-
liness. Journal of Personality and Social Psychology, 6, 1313-1321.
Schwab, J. J., Bell, R. A., Warheit, G. J., & Schwab, R. B. (1979).
Social order and mental health: The Florida Health Study. New
Schwab, R. (1986). Support groups for the bereaved, Journal for
Specialists in Group Work, 11, 100-106.
Silber, E., & Tippett, J. (1965). Self-esteem: Clinical assessment and
measurement validation. PsychologicalReports, 16, 1017-1071.
Thoits, P. A. (1985). Social support and psychological well-being:
Theoretical possibilities. In I. G. Samson & B. R. Samson (Eds.),
Social support: Theory, research and applications (pp. 51-72).
Boston: Martinus Nijhoff.
Weiss, R. (1974). The provisions of social relationships. In Z. Rubin
(Ed.), Doing unto others (pp. 17-26). Englewood Cliffs, N J: Prentice
Wilcox, B. L., & Birkel, R. C. (1983). Social networks and the help-
seeking process: A structural perspective. In A. Nadler, J. D. Fisher,
& B. M. DePaulo (Eds.), New directions in helping: 1Iol. 3. Applied
perspectives on help-seeking and -receiving (pp. 235-253). New
York: Academic Press.
Wilcox, B. L., & Vernberg, E. M. (1985). Conceptual and theoretical
dilemmas facing social support. In I. G. Samson & B. R. Sarason
(Eds.), Social support: Theory, research and applications (pp. 3-
20). Boston: Martinus Nijhoff.
Yalom, I. D. (1985). The theory and practice of group psychotherapy
(3rd ed.). New York: Basic Books.
Received March 10, 1988
Revision received June 20, 1988
Accepted July 26, 1988 •