Personality and Individual Dierences 43 (2007) 21592170

www.elsevier.com/locate/paid

How social connectedness aects guilt and shame:

Mediation by hope and dierentiation of self
a,*
Ian Williamson , Steve J. Sandage b, Richard M. Lee c

a
New Mexico Highlands University, 205 Hewett Hall, Las Vegas, NM 87701, United States
b
Bethel Seminary, SEM A204, PO Box 7029, St. Paul, MN 55112, United States
c
University of Minnesota, 75 East River Road, Minneapolis, MN 55455, United States

Received 25 August 2006; received in revised form 16 June 2007; accepted 28 June 2007
Available online 20 August 2007

Abstract

We tested a mediation model of social connectedness and guilt and shame. Social connectedness was
hypothesized to be related to greater hope and dierentiation of self. Hope and dierentiation of self, in
turn, were expected to be related to less shame and more prosocial guilt. The results found that hope med-
iated the relationship between social connectedness and guilt, and dierentiation of self mediated the rela-
tionship between social connectedness and shame. Alternative explanatory models were explored and taken
into consideration in the interpretation of the results.
2007 Elsevier Ltd. All rights reserved.

Keywords: Social connectedness; Dierentiation of self; Hope; Guilt; Shame

1. Introduction

Why is it that some people are able to handle interpersonal conicts constructively, as expressed
by guilt, and not let these problems overwhelm them with feelings of shame? We examine three
individual dierence variables social connectedness, hope, and dierentiation of self and their

*
Corresponding author. Tel.: +1 505 454 3342; fax: +1 505 454 3331.
E-mail address: iwilliamson@nmhu.edu (I. Williamson).

0191-8869/$ - see front matter 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.paid.2007.06.026
2160 I. Williamson et al. / Personality and Individual Dierences 43 (2007) 21592170

relationships with adaptive feelings of guilt and maladaptive feelings of shame. Using Kohuts
(1977, 1984) theory of self psychology as a guiding framework, we draw upon theory and research
in each area (Lee & Robbins, 1995, 1998; Skowron & Friedlander, 1998; Snyder et al., 1991) to
suggest that social connectedness and feelings of guilt and shame are mediated by self-dierenti-
ation and hope. Specically, people with a secure sense of connectedness in their lives are able to
manage interpersonal conict because they have a more dierentiated sense of self and a greater
sense of hope in life.

2. Social connectedness, hope, and dierentiation of self

Kohut (1984) described the need for belonging, and its internalized representation of social con-
nectedness, as a core interpersonal aspect of the self. Building upon this denition, Lee and Rob-
bins (1995, 1998) dened social connectedness as an individuals sense of enduring interpersonal
closeness with the social world in toto (Lee, Draper, & Lee, 2001). When developmental needs for
belongingness are successfully negotiated in childhood and adolescence a person is likely to form
an internalized sense of self as worthy and capable of forming and maintaining social connection
(Kohut, 1984). Conversely, individuals whose developmental needs for belongingness are frus-
trated experience narcissistic wounds to the self that impede the formation of an internalized sense
of social connectedness in adulthood. These socially disconnected people are prone to feel shame-
ful, socially anxious and alienated from others. Unfortunately, socially disconnected people have
diculty resolving these problems because the conicts overwhelm them emotionally and they feel
powerless to resolve the interpersonal rifts in their lives. Empirical studies of social connectedness
have found negative correlations with anxiety, interpersonal problems, loneliness, and psycholog-
ical distress and positive correlations with self-esteem and social support (Lee et al., 2001; Lee &
Robbins, 1998).
Kohut (1984) further stated that a sense of connectedness serves as the foundation for goal di-
rected behavior. Highly socially connected people have a secure internalized base that allows them
to believe they can achieve life goals. Theoretically then, social connectedness should facilitate
hope, which is comprised of two main components: (a) agency thinking or the belief that one
can reach desired goals, and (b) pathways thinking or the perceived capacity to generate workable
routes to those goals (Snyder, Michael, & Cheavens, 1999). Snyder and colleagues theorized that
social connectedness and hope should be positively related throughout human development (Sny-
der, Cheavens, & Sympson, 1997). However, this hypothesis has received limited empirical testing.
Bolland, Lian, and Formicella (2005) found in a community sample that various types of social
disruption and disconnectedness lead to increased hopelessness and despair. Narcissism was not
correlated with hope in an unpublished study of college students (Snyder & Cheavens, 1997, as
reported in; Snyder et al., 1997), yet these ndings do not address the theorized positive relation-
ship between social connectedness and hope.
Kohut (1977, 1984) also suggested that people high in social connectedness exhibit a greater
capacity for relational independence, and described the socially connected self as an independent
center of initiative (Kohut, 1977, p. 94). In other words, connected selves have the self-cohe-
sion necessary for dierentiation (Banai, Mikulincer, & Shaver, 2005). Dierentiation of self is
described by Bowen (1978) as a trait of mature relational selfhood. Individuals high in
 I. Williamson et al. / Personality and Individual Dierences 43 (2007) 21592170 2161

dierentiation of self are able to exibly integrate both thoughts and feelings, maintain their sense
of self in the presence of others, and independently function when needed (see Schnarch, 1997;
Skowron & Friedlander, 1998). Poorly dierentiated people are more susceptible to negative emo-
tional contagion because they struggle with emotional boundaries in close relationships. Dieren-
tiation of self has been positively correlated with psychological well-being, ethnic belongingness,
self-control, and marital adjustment and negatively correlated with chronic anxiety, psychiatric
symptoms, fears of abandonment, and desire for merger (Murdock & Gore, 2004; Skowron &
Friedlander, 1998; Tuason & Friedlander, 2000).

3. Hope and dierentiation of self in relation to shame and guilt

According to Tangney and her colleagues, highly shame-prone individuals tend to make global
self-evaluations (e.g., I am awful.) following a transgression, while highly guilt-prone individ-
uals tend to distinguish self and behavior (e.g., What I did was awful.) (for a review, see Tang-
ney & Dearing, 2002). Shame activates feelings of worthlessness during conict and generates
hostile defensiveness. In contrast, guilt-proneness appears to be less destructive to self-esteem
and is positively correlated with empathy, apology, seeking forgiveness from others, forgiving oth-
ers, and other reparative behaviors following relational conict.
To our knowledge, the relationships between hope, shame, and guilt have not been empirically
tested. Kierkegaard (1847/1995) suggested that shame can arise in individuals when their self-in-
vested hopes fail to materialize and they experience this as a disappointing failure of the self. Ko-
hut (1977) echoed these sentiments in his description of what happens when an individuals
ambitions and goals are frustrated. By contrast, Tangney and Dearing (2002) described guilt as
a hopeful, future-oriented moral-emotional experience (p. 119). Hopeful people believe that
the future oers an opportunity for improvement and repair, which is more akin to guilt than
shame.
The relationships between dierentiation of self and shame and guilt also appear largely un-
tested. Schnarch (1997) oered the clinical observation that couples low in dierentiation of self
struggle with high levels of shame, which perpetuates relational conicts and prevents connected-
ness. Consistent with this clinical view, Lewis (1971) tested a broader dierentiation construct
in relation to shame and guilt.. In this study, less dierentiated participants tended to be more
shame-prone, while better dierentiated participants tended to be guilt-prone.

4. Summary of expected relationships

In this research, we test an exploratory model of the relationship between social connectedness,
dierentiation of self, hope, and the emotions of shame and guilt. Our empirical model relates so-
cial connectedness directly to positive psychological functioning. Social connectedness is hypoth-
esized to encourage dierentiation of self and a sense of hope, two indicators of healthy selfhood.
We expect hope to encourage guilt and to discourage shame in response to interpersonal trans-
gressions. Dierentiation of self is expected to have a negative relationship with shame. However,
the relationship between dierentiation of self and guilt is not so clear. Maturely separating
2162 I. Williamson et al. / Personality and Individual Dierences 43 (2007) 21592170

oneself from negative emotional contagion may free emotional reserves that can best be put to-
ward repairing conicts (i.e. guilty responding). If dierentiation of self implies too much distance
from others, oenders high in dierentiation of self may respond in an aloof manner, entirely fail-
ing to show guilt or its positive interpersonal manifestations.

5. Method

5.1. Participants

A total of 226 participants (57 men and 169 women) participated in the questionnaire study for
extra course credit at a small Midwestern university. Participants ranged in age from 17 to 23
years, with a median age of 18 years. Participants described themselves mainly as White (93%),
as Asian-American (5%), and only a few from other ethnicities (2%).

5.2. Measures

5.2.1. Social connectedness

To measure peoples subjective sense of being in a close relationship with people, we used the
social connectedness scale-revised (Lee & Robbins, 1995, 1998). Previous studies have shown high
internal consistency (Lee & Robbins, 1995, 1998), and for our sample, the alpha reliability was
high (a = .92). Half of the items on the scale are phrased to directly indicate social connectedness
(e.g. I feel close to people), and the other half are reverse-coded (e.g. I feel disconnected from
the world around me). Participants showed their relative agreement using 6-point Likert scales
(1 = Strongly Disagree, 6 = Strongly Agree).

5.3. Dierentiation of self

Skowron and Friedlander (1998) developed the Dierentiation of Self Inventory1 to capture
four related constructs that underlie Bowens (1978) original concept. The 43-item inventory is
designed to measure the capability to experience emotional intimacy with others, while simulta-
neously remaining independent of others in important ways. The 6-point Likert scales measure
perceptions of how well each statement reects the individual (1 = Not at all true of me,
6 = Very true of me). The alpha reliability for the overall scale was good (a = .82).

5.3.1. Hope
To capture peoples varying degrees of hopefulness, we used Snyder et al. (1997) Hope Scale.
The 8-item scale measures two aspects of hope, pathways thinking (e.g. There are lots of ways
around any problem) or the extent to which people believe they can circumvent numerous bar-
riers in the pursuit of their goals, and agentic thinking (e.g. I energetically pursue my goals) or

1
Our use of the Dierentiation of Self Inventory as a single scale is consistent with past research on the topic (for just
a few examples, see Murdock & Gore, 2004; Peleg, 2005; Skrwron, 2005).
 I. Williamson et al. / Personality and Individual Dierences 43 (2007) 21592170 2163

the extent to which people proactively pursue the goals in their lives. Alpha reliability for the en-
tire scale was .68.2

5.3.2. Guilt and shame

We measured these two constructs using the test of self-conscious aect (TOSCA, Tangney &
Dearing, 2002). For this measure, participants read 15 short hypothetical scenarios which were
gleaned from previous written accounts of personal shame and guilt experiences in the lives of
people of college age or older. After reading these scenarios, participants rated their likely emo-
tional reactions to these incidents on 5-point Likert Scales. For all scenarios, there was a question
measuring guilt and a question measuring shame. For instance, one scenario reads, You make
plans to meet a friend for lunch. At 5 oclock, you realize you stood him up. A response indi-
cating shame is, You would think: Im inconsiderate. A response indicating guilt is, You
would try to make it up to him as soon as possible. Another scenario reads, You make a mis-
take at work and nd out a co-worker is blamed for the error. The shame response for this sce-
nario is You would keep quiet and avoid the co-worker. The guilt response is You would feel
unhappy and eager to correct the situation. The alpha reliability for the shame questions was .72,
and .65 for the guilt questions. Also, as found in previous studies, the correlation between these
two measures was substantial (r = .50).

5.4. Procedure

Participants completed a demographic sheet indicating their age, gender, year in college, and
race or ethnicity, and then completed the measures of social connectedness, dierentiation of self,
hope, guilt, and shame.

6. Results

6.1. Demographic analyses

A few gender dierences emerged. Men scored higher than women on the dierentiation of self
(Men: M = 3.72, SD = .45; Women: M = 3.55, SD = .45, t (224) = 2.34, p < .02, d = .36). Wo-
men, however, scored higher on guilt (Women: M = 4.07, SD = 3.72; Men: M = 3.73,
SD = .43, t = 5.54, p < .001, d = .87), shame (Women: M = 3.18, SD = .48; Men: M = 2.83,
SD = .48, t = 4.34, p < .001, d = .74), and social connectedness (Women: M = 4.55, SD = .74;
Men: M = 4.32, SD = . 72, t = 2.05, p < .05, d = .31), in comparison to men.3 Age and year in
college correlated with only shame (rs = 22 and 14, respectively, ps < .05), and no other
variables.4

2
Treating hope as a single construct is also quite common (e.g. Michael & Snyder, 2005; Peterson, Gerhardt, & Rode,
2006.
3
All four gender dierences are consistent with Carol Gilligans (1982) theory which claims that women, as compared
to men, tend to favor an aective, relational approach to morality, more commonly referred to as an ethic of care.
4
Ethnicity was not analyzed because the low cell sizes make the test highly unreliable.
2164 I. Williamson et al. / Personality and Individual Dierences 43 (2007) 21592170

Table 1
Bivariate correlations between measures
Hope DOS Social Conn. Guilt Shame
Hope
DOS .21**
Social Conn. .41** .33**
Guilt .12 .11 .13*
Shame .18** .37** .13* .50**
*
Indicates p < .05.
**
Indicates p < .01.

6.2. Inter-correlations

We next examined both bivariate and part correlations (see Table 1). The bivariate correlations
show the moderate to strong relationships between the social connectedness, hope, and the dier-
entiation of self (.21 6 rs 6.41), and the strong relationship of shame to guilt (r = .50). Hope and
dierentiation of self are also correlated with shame. None of the bivariate relationships to guilt
are very strong. Table 2 presents the part correlations between the individual dierence measures
and guilt, parsing out the common variance associated with shame, as well as the part correlations
with shame, extracting the common variance associated with guilt. Tangney and colleagues have
consistently found that guilt-free shame is associated with more negative predictors and conse-
quences, and shame-free guilt is associated with more positive predictors and consequences.
Thus, these relationships are obscured when only looking at bivariate correlations (see Tangney
& Dearing, 2002 for a review). As expected, the relationships of hope and social connectedness to
shame-free guilt are strengthened by partialling out shame.

6.3. Structural modeling

We proposed a latent variable model to explain the correlational relationships between all the
measured variables in the data (see Fig. 1). To create the latent variables, we used three item par-
cels as the indicators for all of the latent variables. Item parcels are created by simply taking the
mean of a certain number of items on a scale. For instance, because hope is measured with a 12-
item scale, we created three item parcels for the hope variable by taking the mean of each set of
four items. These three means then served as indicators for the latent construct. Item parcels are
used to increase the stability of the indicators in the model (Joreskog & Sorbom, 1986; MacCal-
lum, 1986).5
Previous research has shown that social connectedness underlies much positive individual psy-
chological functioning, improving self-esteem (Baumeister & Leary, 1995; Leary, Haupt, Straus-
ser, & Chokel, 1998), and mitigating against anxiety (Lee et al., 2001). For this reason, we posited

5
Because of the controversial nature of using item parcels as observed variables (e.g. Little, Cunningham, Shahar, &
Widaman, 2002), we also used a modeling technique in which the means of the whole scales and reliability scores were
used in lieu of item parcels. This model also t the data well (CFI = .99, RMSEA = .03), with the only path change
being that dierentiation of self signicantly, rather than marginally, predicted guilt.
 I. Williamson et al. / Personality and Individual Dierences 43 (2007) 21592170 2165

Table 2
Part correlations with guilt and shame
Guilt Shame
Social connectedness .23 (.001) .23 (.001)
Dierentiation of self .08 (.14) .36 (<.001)
Hope .24 (.001) .28 (<.001)
Note: p-values in parentheses.

Hope
Guilt
Social
Connectedness

Shame
DOS

Fig. 1. Proposed latent model underlying the data.

that peoples perceived sense of connectedness with others would be associated with greater dif-
ferentiation of self and with increased hope. We expected that dierentiation of self would be
associated with lower shame and with greater guilt. Both hope and dierentiation of self were ex-
pected to mediate the relationship between social connectedness and shame and guilt.
Though not represented in the diagram, we allowed the structural error terms of these two con-
structs to correlate in order to partial out their mutual eect on each other. This is consistent with
prior research on shame and guilt which typically examines relationships of these constructs to
others after partialling out their mutual eect upon one another (Tangney & Dearing, 2002).

6.3.1. Fit of the proposed model

The overall t statistics for the proposed mediation model were acceptable (CFI = .93,
RMSEA = .056) (see Fig. 2).6 All of the regression weights for the drawn paths were signicant
(all ps < .01), with two exceptions: the path between dierentiation of self and guilt was margin-
ally signicant (t = 1.88. p = .06), and the path between hope and shame was non-signicant
(t = 1.20, ns). As expected, social connectedness positively relates to both dierentiation of self
and hope. Furthermore, hope is positively related to guilt, and dierentiation of self is negatively
related to shame. The two somewhat surprising ndings are that dierentiation of self was nega-
tively related to the expression of guilt, and hope had no relation to shame. Recall that we made
no specic prediction about the relation between dierentiation of self and guilt because we sug-
gested that both a positive and negative relation could be theoretically supported.
Structural models can be used to demonstrate mediational models just as well as regression
(Maruyama, 1996). We have already shown three of the four steps to full mediation as outlined
by Baron & Kenny (1986) because (a) the independent variable correlates with the dependent

6
When gender was added as an exogenous, observed variable predicting dierentiation of self, social connectedness,
guilt, and shame, the overall t statistics remained acceptable (CFI = .92, RMSEA = .06). Moreover, hope and
dierentiation still mediated the relationship of social connectedness to guilt and shame, respectively.
2166 I. Williamson et al. / Personality and Individual Dierences 43 (2007) 21592170

.31
.55 Hope
Guilt
Social
Connectedness -.18
.40 -.11
DOS -.44
Shame

Fig. 2. Original model. Notes: All paths signicant at p 6 .01, except the path from hope to shame (ns), and the path
from dierentiation of self to guilt (p = .06).

variables, (b) the paths from the independent variable to the mediators are signicant, and c) the
paths from the mediators to the dependent variables are signicant. To demonstrate full media-
tion (step 4), we need to show that the paths from the independent variable to the dependent vari-
ables drops to non-signicance with the mediators present in the model. Indeed, the paths from
social connectedness to guilt (b = .13, p = .29) and shame (b = .12, p = .27) are not signicant
when added to the model.
In conclusion, the mediated relationship between social connectedness and guilt is best ex-
plained by hope, whereas the mediated relationship between social connectedness and shame is
best explained by the dierentiation of self. As with multiple regression, structural equation mod-
eling allows you to examine the squared multiple correlation (R2) to get a sense of the explanatory
power of the model. The R2 statistics for each of the endogenous latent variables are as follows:
Dierentiation of Self = .16, Hope = .31, Guilt = .11, Shame = .24. The standardized indirect
relationship between social connectedness and shame was .24, and the standardized indirect rela-
tionship between social connectedness and guilt was .10.

6.3.2. Alternative Model #1

According to Bowen (1978), people learn the dierentiation of self largely through constructive
or destructive relationships through family socialization. Therefore, it is theoretically plausible
that dierentiation of self paths could be associated with hope and social connectedness, which
would then each relate to shame and guilt. This model did not t the data as well (CFI = .89;
RMSEA = .07), and the direct paths from social connectedness to guilt and shame were not sig-
nicant (ps P .19).

6.3.3. Alternative Model #2

It is common practice to test a model in which the dependent measures and the mediating vari-
ables are switched. In this case, social connectedness would be associated with a tendency toward

1.12
.23 Guilt
Hope
Social
Connectedness -1.13
-.21 .61
Shame -.95
DOS

Fig. 3. Alternative Model #2. Notes: All paths signicant at p 6 .02.

 I. Williamson et al. / Personality and Individual Dierences 43 (2007) 21592170 2167

guilt and away from shame. Guilt, in response to transgression would be related to greater dier-
entiation of self, and greater hopefulness. Shame would do the reverse. This model is consistent
with Leary et al. (1998) sociometer hypothesis in which abstracted feelings about ourselves (i.e.
hope and dierentiation of self) are derived primarily through relational processes, especially in
the context of feeling rejected (e.g. social connectedness, shame, and guilt). Klein (1975) is one
inuential theorist who suggested that guilt mediates the relationship between social connection
and hope. This model t the data quite well (CFI = .92, RMSEA = .058), in which all of the paths
in the model are signicant. Please see Fig. 3. This model is a potential alternative to the proposed
model.

7. Discussion

Social connectedness, insofar as it implies an active involvement and connection in the rela-
tional sphere, provides psychological resources to people who want to make reparations for their
past. Our proposed model shows that people who have greater social connectedness are more
hopeful, and are better at dierentiating themselves. Hope is associated with proactive guilt,
and self-dierentiation is associated with personal shame.
Caution is warranted about likely causal pathways, however, because an equally plausible alter-
native model shows guilt and shame mediating the relationship between social connectedness and
hope and guilt. Though it is true that the causal ordering in this study is not altogether clear, this
does not support the claim that we can not draw any conclusions based on support for our ori-
ginal model. Structural modeling still allows us to narrow the range of plausible models (Maruy-
ama, 1996). Our proposed model is a better theoretical t with self psychology than alternative
model number 2. Kohut related dierentiated self-cohesion with low shame and hope with an
internalized relation to empathic others which facilitates reparative guilt. Alternative model 3 is
plausible and ts Kleins (1975) theoretical focus on internalized relational conicts and guilt.
Klein suggested that moral emotions of adaptive and neurotic guilt (with the latter similar to
Tangneys shame construct) mediate the relationship between internal security about connecten-
dess with others and hopeful and dierentiated functioning. Overall, our results also t with
theoretical and clinical advances in contemporary relational psychoanalysis that posit a complex,
co-occurance of the dynamics of hope and guilt rather than a linear developmental sequence as
earlier debated by Kohutians and Kleinians (Mitchell, 1993).
Prior studies of social connectedness and dierentiation of self have already implicated these
relational self-concepts in our understanding of mental health and well-being (Lee & Robbins,
1995, 1998; Skowron & Friedlander, 1998), as well as marital adjustment (Skowron, 2000). Our
research shows that some of these ndings may be linked to the experience of guilt and shame.
People low in social connectedness and dierentiation may want to hide or lash out (i.e. experi-
ence shame) instead of being motivated to take corrective or reparative action (i.e. experience
guilt).
In the past, hope has typically been correlated with indicators of achievement and well-being
(e.g. Snyder, Rand, & Sigmon, 2002). We add to this knowledge by demonstrating how hope is
directly associated with social connectedness, dierentiation of self, and oenders proactive guilty
responses to interpersonal transgressions. We found support in our structural model that those
2168 I. Williamson et al. / Personality and Individual Dierences 43 (2007) 21592170

high in social connectedness will be more hopeful. Furthermore, hope is associated with better
remedial strategies for resolving relational culpability.
Though in our study we do not study guilt and shame responses in real-world settings, there are
multiple recent studies which make use and establish the validity of the TOSCA as a proxy mea-
sure (Fedewa, Burns, & Gomez, 2005). Though Tangney and Dearing (2002) acknowledge the
limits of scenario-based measures, they support the use of their measure for the following reasons:
Scenario-based measures t with the contextual framework of much recent work on shame and
guilt. Specically, contextual factors can cue emotional responses in a manner that is dierent
from asking participants to respond to global adjectives. The TOSCA scenarios were subject-
generated (Tangney & Dearing, 2002, p. 42) through studies with large samples of participants
to facilitate the ecological validity of the measure. The scenarios selected for the measure include a
wide range of situations across numerous domains (e.g., work, personal relationships) that most
people could encounter in everyday life. The scenario approach is also thought to hold the indirect
advantage of getting past the potential defensive response if participants were asked more directly
about feeling guilty or ashamed. Perhaps most importantly, TOSCA has been validated by
correlating it with the wide range of behaviors (e.g. avoidance responses) and psychopathologies
(e.g. anxiety, depression) to which we refer in the manuscript.
A suggested future study involves examining how social connectedness, dierentiation of self,
and hope relate to peoples shameful and guilty responses to actual transgressions. Such a study
would add more conrmation to the relationships between these individual dierence measures
and responses to culpability in everyday life. Another avenue for further research would involve
counseling interventions focused on improving aspects of the relational self (i.e. social connected-
ness and dierentiation of self), and measuring the eect on hope, shame, and guilt with clinical
populations. Finally, future studies could investigate the relationships between social connected-
ness, dierentiation of self, hope, and forgiving others following conicts.

Acknowledgement

We acknowledge the work of the anonymous reviewers, who carefully commented on our work
in an earlier draft.

References

Banai, E., Mikulincer, M., & Shaver, P. R. (2005). Selfobject needs in Kohuts Self Psychology: Links with attachment,
self-cohesion, aect regulation, and adjustment. Psychoanalytic Psychology, 22, 224260.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research:
Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 11731182.
Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental
human motivation. Psychology Bulletin, 117, 497529.
Bolland, J. M., Lian, B. E., & Formicella, C. M. (2005). The origins of hopelessness among inner-city African
American adolescents. American Journal of Community Psychology, 36, 293305.
Bowen, M. (1978). Family therapy in clinical practice. New York: Joshua Aronson Inc.
Fedewa, B. A., Burns, L. R., & Gomez, A. A. (2005). Positive and negative perfectionism and the shame/guilt
distinction: adaptive and maladaptive characteristics. Personality and Individual Dierences, 38, 16091619.
 I. Williamson et al. / Personality and Individual Dierences 43 (2007) 21592170 2169

Gilligan, C. (1982). In In a dierent voice: Psychological theory and womens development. Cambridge, MA: Harvard
University Press.
Joreskog, K. G., & Sorbom, D. (1986). LISREL IV: Analysis of linear structural relationships by the method of maximum
likelihood. Chicago: National Educational Resources.
Kierkegaard, S. (1995). In H. V. Hong & E. H. Hong (Eds.), Works of love. Princeton, NJ: Princeton University Press,
Original work published 1847.
Klein, M. (1975). Envy and gratitude and other works. The writings of Melanie Klein 19461963 (vol. 3). New York: The
Free Press, pp. 247263.
Kohut, H. (1977). The analysis of the self: A systematic approach to the psychoanalytic treatment of narcissistic
personality disorder. Madison, CT: International Universities Press.
Kohut, H. (1984). How does analysis cure? New York: International Universities Press.
Leary, M. R., Haupt, A. L., Strausser, K. S., & Chokel, J. T. (1998). Calibrating the sociometer: The relationship
between interpersonal appraisals and the state self-esteem. Journal of Personality & Social Psychology, 74,
12901299.
Lee, R. M., Draper, M., & Lee, S. (2001). Social connectedness, dysfunctional interpersonal behaviors, and
psychological distress testing a mediator model. Journal of Counseling Psychology, 48, 310318.
Lee, R. M., & Robbins, S. B. (1995). Measuring belongingness: The social connectedness and social assurance scales.
Journal of Counseling Psychology, 42, 232241.
Lee, R. M., & Robbins, S. B. (1998). The relationship between social connectedness and anxiety, self-esteem, and social
identity. Journal of Counseling Psychology, 45, 338345.
Lewis, H. B. (1971). Shame and guilt in neurosis. New York: International Universities Press.
Little, T. D., Cunningham, W. A., Shahar, G., & Widaman, K. F. (2002). To parcel or not to parcel: Exploring the
question, weighing the merits. Structural Equation Modeling, 9, 151173.
MacCallum, R. (1986). Specication searches in covariance structure modeling. Psychological Bulletin, 100, 107120.
Maruyama, G. M. (1996). Basics of structural equation modeling. Thousand Oaks, CA: Sage.
Michael, S. T., & Snyder, C. R. (2005). Getting unstuck: The roles of hope, nding meaning, and rumination in the
adjustment to bereavement among college students. Death Studies, 29, 435458.
Mitchell, S. A. (1993). Hope and dread in psychoanalysis. New York: Basic Books.
Murdock, N. L., & Gore, P. A. Jr., (2004). Stress, coping, dierentiation of self: A test of Bowen Theory. Contemporary
Family Therapy, 26, 319335.
Peleg, O. (2005). The relation between dierentiation and social anxiety: What can be learned from students and their
parents? The American Journal of Family Therapy, 33, 167183.
Peterson, S. J., Gerhardt, M. W., & Rode, J. C. (2006). Hope, learning goals, and task performance. Personality and
Individual Dierences, 40, 10991109.
Schnarch, D. M. (1997). Passionate marriage: Love, sex, and intimacy in emotionally committed relationships. New York:
W.W. Norton.
Skowron, E. A. (2000). The role of dierentiation of self in marital adjustment. Journal of Counseling Psychology, 47,
229237.
Skowron, E. A. (2005). Parent dierentiation of self and child competence in low-income urban families. Journal of
Counseling Psychology, 52, 337346.
Skowron, E. A., & Friedlander, M. L. (1998). The dierentiation of self inventory: Development and initial validation.
Journal of Counseling Psychology, 45, 235246.
Snyder, C. R., & Cheavens, J. (1997). Hope and social matters. Unpublished manuscript, University of Kansas,
Lawrence.
Snyder, C. R., Cheavens, J., & Sympson, S. C. (1997). Hope: An individual motive for social commerce. Group
Dynamics: Theory, Research, and Practice, 1, 107118.
Snyder, C. R., Harris, C., Anderson, J. R., Holleran, S. A., Irving, L. M., Sigmon, S. T., et al. (1991). The will and the
ways: Development and validation of an individual-dierences measure of hope. Journal of Personality and Social
Psychology, 60, 570585.
Snyder, C. R., Michael, S. T., & Cheavens, J. S. (1999). Hope as a psychotherapeutic foundation of common factors,
placebos, and expectancies. In M. A. Hubble, B. L. Duncan, & S. D. Miller (Eds.), The heart and soul of change:
What works in therapy (pp. 179200). Washington, DC: APA.
2170 I. Williamson et al. / Personality and Individual Dierences 43 (2007) 21592170

Snyder, C. R., Rand, K. L., & Sigmon, D. R. (2002). Hope theory: A member of the positive psychology family. In C.
R. Snyder & S. J. Lopez (Eds.), Handbook of positive psychology (pp. 257276). New York: Oxford University Press.
Tangney, J. P., & Dearing, R. L. (2002). Shame and guilt. New York: The Guilford Press.
Tuason, M. T., & Friedlander, M. L. (2000). Do parents dierentiation levels predict those of their adult children? and
other tests of Bowen theory in a Philippine sample. Journal of Counseling Psychology, 47, 2735.