You are on page 1of 6

Personality and Individual Differences 111 (2017) 291–296

Contents lists available at ScienceDirect

Personality and Individual Differences

journal homepage: www.elsevier.com/locate/paid

Attachment insecurity and openness to diversity: The roles of self-esteem


and trust☆
Suejung Han ⁎
Department of Psychology, Illinois State University, United States

a r t i c l e i n f o a b s t r a c t

Article history: Openness to diversity is useful to adjustment in the diversified modern society, but diversity can be perceived as a
Received 3 January 2017 threat to one's identity and cultural values. Thus, tendency to feel secure and not easily threatened—attachment
Received in revised form 10 February 2017 security—may be associated with openness to diversity. A path model hypothesized the associations between at-
Accepted 13 February 2017
tachment insecurity (i.e., attachment anxiety and avoidance) and universal-diverse orientation (UDO; i.e., open
Available online 20 February 2017
attitude toward diversity) both directly and indirectly through low self-esteem and interpersonal trust. A struc-
Keywords:
tural equation modeling analysis using two samples (N = 338 and 350, respectively) of U.S. undergraduates re-
Attachment vealed that attachment anxiety was associated with the UDO cognitive component Relativistic Appreciation
Self-esteem positively but its emotional component Comfort with Differences negatively, suggesting ambivalent attitudes
Trust to diversity. Indirect effects were significant between attachment insecurity and UDO through trust, although
Universal-diverse orientation self-esteem was associated with UDO only through its association with trust. It was suggested that mistrust in
Openness to diversity others' good intention might be why insecurely attached feel uncomfortable with the culturally different people.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction secure and not easily threatened may be more open to diversity. Thus,
this study examined whether attachment security was associated with
In the increasingly diversified and globalized modern society, openness to diversity because of its association with self-esteem and
successful adjustment entails interacting with diverse individuals and trust.
negotiating diverse worldviews effectively (Jackson & Ruderman,
1995). Such diversity competence may be fostered by open attitude to- 1.1. Attachment theory
ward diversity (Chang, 2001).
However, close-mindedness to differences may be an immediate re- Attachment theory proposes that attachment system—a biologically
sponse because it had been useful for human survival. From an evolu- prewired behavioral systems—activates upon perceiving threats, and
tionary perspective (Confer et al., 2010), differences were equated serves to ensure our survival by guiding physical proximity to attach-
with potential danger in prehistoric times; it was adaptive to defend ment figures for protection (Bowlby, 1988). The attachment system de-
oneself by categorizing different looking people as an enemy, feeling activates when the attachment figure provides protection and soothing.
fearful or angry, and distancing from or harming the person. As such, If attachment concern is active, other behavioral system functioning
human evolution facilitated close-mindedness to diversity (Cole & (e.g., exploratory system function for exploring environments and nov-
Teboul, 2004). elty; Elliot & Reis, 2003) becomes compromised (Bowlby, 1988).
Then, individual difference factors associated with openness to Individual differences in the attachment quality develop from the
diversity may counter such closed-mindedness to diversity. Because di- early childhood, depending on the attachment figure response to the
versity can be perceived as a threat (Strauss, Connerley, & Ammerman, proximity seeking (Bowlby, 1988). This attachment security versus in-
2003), and felt threat can decrease tolerance toward different cultural security persist into adulthood as a stable trait (Fraley, 2002), internal-
groups (Stephan, Ybarra, & Rios Morrison, 2009), individuals who feel ized in the form of internal working model (IWM; Bowlby, 1988).
IMW is a schema that consists of mental representations of self,
others, and relationships. The child may internalize positive self- (e.g.,
☆ This study was supported by Faculty Diversity Research Award granted to the author “I am worthwhile”) and other-representation (e.g., “others are trust-
by the University of Wisconsin System Institute on Race and Ethnicity. I thank Dr. M. Carole
worthy”) and develop attachment security, if the attachment figure is
Pistole for being the research mentor for this award.
⁎ Corresponding author at: Department of Psychology, Illinois State University, Campus
perceived as consistently responsive and available (Baldwin, 1992).
Box 4620, Normal, IL 61790-4620, United States. Conversely, perceived unpredictable and inconsistent responses of the
E-mail address: shan3@ilstu.edu. attachment figure may lead to negative self-representation (e.g., “I

http://dx.doi.org/10.1016/j.paid.2017.02.033
0191-8869/© 2017 Elsevier Ltd. All rights reserved.
292 S. Han / Personality and Individual Differences 111 (2017) 291–296

may not be worthy of love”), developing attachment anxiety. Perceived negative self-evaluation (e.g., self-esteem; Mikulincer & Shaver, 2007a)
consistent rejection or unavailability of the attachment figure may lead and become motivated to protect felt-security by defensive measures
to negative other-representation (e.g., “others are not trustworthy”), (e.g., devaluing the culturally different) (Mikulincer & Shaver, 2001,
developing attachment avoidance. These mental representations beget 2007b), which would be associated with low UDO; a person with
evaluative judgment of self (e.g., self-esteem) and others (e.g., trust) attachment avoidance may feel threatened due to the negative other-
(Wearden, Peters, Berry, Barrowclough, & Liversidge, 2008) and guide evaluation (e.g., trust; Mikulincer & Shaver, 2007a) and become moti-
the person with certain expectations and motivations in life tasks and vated to protect felt-security by distancing from them, which would
relationships, thereby affecting adjustment (Mikulincer & Shaver, be associated with low UDO.
2007a) including how one manages perceived threats. Moreover, due to unaddressed attachment concerns, attachment
system is chronically activated (attachment anxiety) or excessively
1.2. Attachment and diversity as threat suppressed (attachment avoidance) (Bowlby, 1988). Therefore, explor-
atory system functioning such as exploring diversity may be compro-
Cultural diversity may be perceived as threat to our survival (Confer mised (Elliot & Reis, 2003), resulting in low openness to diversity.
et al., 2010), desire to maintain positive self-concept (Tsui, Egan, & Thus, attachment anxiety and avoidance may also be associated with
O'Reilly, 1992), and the sense of sharedness within one's cultural low UDO directly.
group (Stephan et al., 2009). Such felt threat can create negative evalu-
ation and behaviors toward culturally different people (Stephan et al.,
2009). 2. Hypothesis and alternative models
Individuals with attachment insecurity versus security may address
such felt threat differently, developing more negative versus positive at- With these rationales, I proposed a path model that attachment anx-
titude toward diversity. Despite the dearth of research on the attach- iety and avoidance were associated with UDO directly and indirectly
ment and openness to diversity associations, literature on attachment through low self-esteem and low trust respectively (Fig. 1). This
and prejudice/intergroup relations is relevant. For example, experimen- model (Model 1) was tested and compared with three theoretically
tally primed attachment security (e.g., subliminal presentation of secu- plausible alternative models. Model 2 added to Model 1 paths from at-
rity-related words or images) reduced (a) negative evaluations of tachment anxiety to trust and attachment avoidance to self-esteem.
outgroups (Boag & Carnelley, 2015; Mikulincer & Shaver, 2001), (b) ag- With mixed findings on the valence of other-representation in attach-
gressive behaviors toward outgroup members (Mikulincer & Shaver, ment anxiety (Pietromonaco & Barrett, 2000) and self-representation
2007b) and immigrants (Saleem et al., 2015), and (c) discriminatory in- in attachment avoidance (Hazan & Shaver, 1987), it is not clear whether
tention and behaviors toward Muslims (Boag & Carnelley, 2012). These self-esteem and trust are common mediating paths for both attachment
results suggested that enhanced attachment security reduced a sense of anxiety and avoidance versus distinctive paths (as hypothesized) in re-
felt threat, thereby rendering unnecessary the natural defensive efforts lation to UDO, which was explored by Model 2. Model 3 added to Model
to devalue or harm outgroup members (Mikulincer & Shaver, 2007b). 1 a path from self-esteem to trust. IWM includes representations of self,
Attachment insecurity as an enduring trait—not experimentally others, and their relations (Baldwin, 1992). It is possible that self-in-re-
boosted state—has been examined in relation to prejudice and inter- lation-to others representation may be more relevant than the indepen-
group contacts in a few studies. Some reported attachment anxiety, dent self-representation in predicting UDO (Pietromonaco & Barrett,
but not avoidance, was associated with outgroup devaluation 2000), because one may feel personal threat from diversity, rather
(Mikulincer & Shaver, 2001) and negative emotions toward Arabs than viewing diversity as a threat in general, which was explored by
among U.S. college students (Saleem et al., 2015). Another study Model 3. Model 4 combined Models 2 and 3. The best fitting model
found that both anxiously and avoidantly attached people showed then was chosen and cross-validated in a different sample.
higher prejudice to outgroup members (Hofstra, van Oudenhoven, &
Buunk, 2005). Also, attachment security, but not attachment anxiety 3. Method
and avoidance, was associated with direct contacts with immigrants, al-
though attachment avoidance was associated with outgroup devalua- 3.1. Participants
tion (Boccato, Capozza, Trifiletti, & Di Bernardo, 2015). Thus, it is not
conclusive whether and which attachment insecurity dimensions are Participants were undergraduates enrolled at a regional Midwestern
associated with prejudice or negative behaviors toward the culturally university in the U.S. Sixty out of 398 responses in the main sample and
different, let alone with openness to diversity. 63 out of 413 in the validation sample with b 50% survey completion
were deleted. The main sample (N = 338) included 136 (40.2%) men
1.3. Attachment and universal-diverse orientation and 198 (58.6%) women with a mean age 21.18 (SD = 5.13, range
18–54). The validation sample (N = 350) included 139 (39.7%) men
In this study, openness to diversity was operationalized as universal- and 209 (59.7%) women with a mean age 21.34 (SD = 5.78, range
diverse orientation (UDO)—openness to and appreciation of human 18–46). Both the samples were mostly European Americans (88.8%,
similarities and cultural and individual differences of others (Miville et 89.7% respectively).
al., 1999). UDO consists of a behavioral Diversity of Contacts—one's
past and intended future behaviors regarding diversity, a cognitive Rel-
ativistic Appreciation—appreciation of value of diversity, and an emo- 3.2. Instruments
tional Comfort with Differences—a sense of connectedness and
comfort with diverse individuals—components (Miville et al., 1999). 3.2.1. Attachment
UDO was associated negatively with prejudices such as homophobia The 12-item Experiences in Close Relationship-Short Form (Wei,
and personality such as dogmatism (e.g., Miville et al., 1999), but no Russell, Mallinckrodt, & Vogel, 2007) measured attachment anxiety
previous studies have examined systematic differential predictions of (6 items, e.g., “I need a lot of reassurance that I am loved by my partner”)
the UDO components, nor investigated their association with and attachment avoidance (6 items, e.g., “I am nervous when partners
attachment. get too close to me”) on a 7-point Likert-type scale from 1 (strongly dis-
As a tendency to feel threatened easily, however, attachment insecu- agree) to 7 (strongly agree). Internal consistency coefficients were 0.78
rity may be associated with UDO through negative IMWs. Encountering and 0.78 for attachment anxiety and 0.86 and 0.84 for attachment
cultural diversity, the anxiously attached may feel threatened due to the avoidance in each sample.
S. Han / Personality and Individual Differences 111 (2017) 291–296 293

Fig. 1. The hypothesized (Model 1) and Alternative models (Models 2–4). Model 1 specified the solid lines. Model 2 added the dashed lines to Model 1. Model 3 added the dotted line to
Model 1. Model 4 added both the dashed and dotted lines to Model 1.

3.2.2. Self-esteem any UDO components in the validation sample. In both samples, attach-
The 10-item Rosenberg's (1965) Self-Esteem Scale measured global ment anxiety was associated with Relativistic Appreciation positively
self-esteem (e.g., “On the whole, I am satisfied with myself.”) on a 4- but Comfort with Differences negatively. Neither attachment anxiety
point scale, from 1 (strongly disagree) to 4 (strongly agree). Internal nor avoidance was associated with Diversity of Contact. Both attach-
consistency coefficients were 0.87 and 0.90 in each sample. ment anxiety and avoidance were associated with self-esteem and
trust negatively. Self-esteem was associated with Comfort with Differ-
3.2.3. Trust ences only; trust was associated with all the UDO components.
The 20-item subscale of General Trust of the 40-item Trust Inventory
(Couch, Adams, & Jones, 1996) measured trust in generalized others
(e.g., “Most people are trustworthy.”) on a 5-point Likert-type scale 4.2. Data analysis plan for hypothesis testing
from 1 (strongly disagree) to 5 (strongly agree). Internal consistency co-
efficients were 0.87 and 0.88 in each sample. A structural equation modeling (SEM; Schumacker & Lomax, 2010)
analysis tested the hypothesized path model using AMOS 22.0. In con-
structing the model, Diversity of Contact was not included due to its
3.2.4. Universal-diverse orientation non-significant associations with attachment anxiety and avoidance.
The 15-item Miville-Guzman Universality-Diversity Scale-Short Relativistic Appreciation and Comfort with Differences were specified
(Fuertes, Miville, Mohr, Sedlacek, & Gretchen, 2000) measured the as separate latent constructs because attachment anxiety and avoidance
UDO (Miville et al., 1999) on a Likert-type scale from 1 (strongly dis- were associated with Relativistic Appreciation in the opposite direc-
agree) to 6 (strongly agree). The three subscales include (a) Diversity tions. Three parcels for attachment anxiety and avoidance respectively
of Contact (5 items; e.g., “I attend events where I might get to know peo- and two parcels for self-esteem, trust, Relativistic Appreciation, and
ple from different racial backgrounds”), (b) Relativistic Appreciation (5 Comfort with Differences were created using the item-to-construct bal-
items; e.g., “Persons with disabilities can teach me things I could not ance method (Little, Cunningham, Shahar, & Widaman, 2002). The sam-
learn elsewhere”), and (c) Comfort with Differences (5 items; e.g., “I ple sizes (N = 338 & 350) were sufficient for detecting a small effect size
am only at ease with people of my own race” [reverse coded]). Internal with 6 latent variables and 14 indicators (minimum required sample
consistency coefficients were 0.79 and 0.81 (Diversity Contact) 0.71 and size = 323, Westland, 2010). Model fit was evaluated by Comparative
0.73 (Relativistic Appreciation), and 0.81 and 0.82 (Comfort with Differ- Fit Index (CFI) value of 0.95 or above, Root Mean Square Error of Ap-
ences) in each sample. proximation (RMSEA) value of 0.08 or below, and Standardized Root
Mean Square Residual (SRMR) value of 0.08 or below (Schumacker &
4. Results Lomax, 2010). Chi-square statistics (χ2) was used to compare nested
models fit; Expected Cross-Validation Index (ECVI; smaller scores indi-
4.1. Preliminary analyses cate better fit) for non-nested models fit. The Expectation Maximization
imputation for missing data was used to allow AMOS to calculate SRMR
Mean, standard deviations, and correlations among the study vari- and conduct bootstrapping analyses to test the significance of indirect
ables are presented in Table 1. In both samples, skewness and kurtosis effects, because missing data was at complete random, Little's MCAR
statistics demonstrated univariate normality of the variables. Attach- χ2 (188, N = 338) = 162.42, p = 0.91 in the main sample, χ2 (321,
ment avoidance was negatively associated with Relativistic Apprecia- N = 350) = 243.53, p = 1.00 in the validation sample (Schlomer,
tion and Comfort with Differences in the main sample but not with Bauman, & Card, 2010).
294 S. Han / Personality and Individual Differences 111 (2017) 291–296

Table 1
Means, standard deviations, and correlations.

Range Main Validation 1 2 3 4 5 6 7

M SD M SD

1 Anxiety 1–7 3.76 1.13 3.73 1.24 − 0.18⁎⁎ −0.32⁎⁎⁎ −0.23⁎⁎⁎ 0.02 0.11⁎ −0.27⁎⁎⁎
2 Avoidance 1–7 2.74 1.14 2.73 1.19 0.28⁎⁎⁎ − −0.24⁎⁎⁎ −0.24⁎⁎⁎ −0.04 −0.18⁎⁎ −0.16⁎⁎
3 Self-esteem 1–4 3.05 0.46 3.06 0.52 −0.43⁎⁎⁎ −0.34⁎⁎ − 0.34⁎⁎⁎ −0.03 0.04 0.19⁎⁎
4 Trust 1–5 3.30 0.59 3.23 0.63 −0.31⁎⁎⁎ −0.35⁎⁎⁎ 0.40⁎⁎⁎ − 0.17⁎⁎ 0.30⁎⁎⁎ 0.35⁎⁎⁎
5 Behavioral 1–6 3.72 1.00 3.67 1.03 0.03 0.01 0.07 0.22⁎⁎⁎ − 0.49⁎⁎⁎ 0.32⁎⁎⁎
6 Cognitive 1–6 4.49 0.68 4.45 0.74 0.20⁎⁎⁎ 0.01 0.05 0.24⁎⁎⁎ 0.51⁎⁎⁎ − 0.32⁎⁎⁎
7 Emotional 1–6 4.57 0.89 4.51 0.94 −0.21⁎⁎⁎ −0.06 0.23⁎⁎⁎ 0.27⁎⁎⁎ 0.28⁎⁎⁎ 0.23⁎⁎⁎ −

Note. N = 338 for main sample, 350 for validation sample; 1 attachment anxiety, 2 attachment avoidance; 5 UDO Diversity of Contact, 6 UDO Relativistic Appreciation, 7 UDO Comfort with
Differences.
Above diagonal = main sample; below diagonal = validation sample.
⁎ p b 0.05.
⁎⁎ p b 0.01.
⁎⁎⁎ p b 0.001.

4.3. Hypothetical and alternative models testing 4.4. Validation of the model

In the main sample, the measurement model fit the data well, χ2 (62, The validation sample was used to cross-validate Model 4 and the in-
N = 338) = 147.22, p b 0.001, CFI = 0.96, RMSEA = 0.064 (90% CI = direct effects. The absolute and relative fits of the measurement model
0.051, 0.077), SRMR = 0.044. Standardized factor loadings ranged and Models 1–4 were similar to those from the main sample
from 0.66 to 0.93, p b 0.001, suggesting that the observed indicators rep- (Appendix 1). Model 4 path coefficients and directions were the same
resented latent constructs well. All the latent variables were correlated as those identified in the main sample (see Fig. 2), except for the non-
with one another significantly, except for self-esteem and Relativistic significant attachment anxiety-Comfort with Differences and attach-
Appreciation, r = 0.06, p = 0.40, and attachment anxiety and Relativis- ment avoidance-Relativistic Appreciation associations. A bootstrapping
tic Appreciation, r = 0.14, p = 0.06. analysis with 5000 resamples revealed that all the indirect effects were
The hypothesized model (Model 1) fit to the data was marginally ad- significant (Table 2). In summary, Model 4 that specified common me-
equate, χ2 (65, N = 338) = 191.14, p b 0.001, CFI = 0.95, RMSEA = diating paths by trust and self-esteem × trust between attachment anx-
0.076 (90% CI = 0.064, 0.088), SRMR = 0.089, ECVI = 0.888. The iety and avoidance and Relativistic Appreciation and Comfort with
paths coefficients were significant in the expected direction, except for Differences fit the data best in the main sample and was cross-validated
the positive association between attachment anxiety and Relativistic in the validation sample.
Appreciation, β = 0.29, p b 0.001, and non-significant associations be-
tween self-esteem and both Relativistic Appreciation, β = − 0.004, 5. Discussion
p = 0.95, and Comfort with Difference, β = 0.11, p = 0.08, and attach-
ment avoidance and Comfort with Difference, β = −0.05, p = 0.41. 5.1. Indirect effects by self-esteem and trust
Model 2, adding to Model 1 paths from attachment anxiety to trust
and attachment avoidance to self-esteem, and Model 3, adding to The best-fitting alternative model suggested that both attachment
Model 1 path from self-esteem to trust, fit the data well (see anxiety and avoidance were associated with Relativistic Appreciation
Appendix 1 for fit indices) and better than Model 1, Δ χ2 (2, N =
338) = 19.42, p b 0.001 for Model 2, and Δ χ2 (1, N = 338) = 31.37, Table 2
p b 0.001 for Model 3. ECVI was smaller in Model 3 than in Model 2, sug- Standardized bias-corrected indirect effects with 95% confidence intervals (CIs) from
gesting Model 3 was a better fitting model than Model 2. The model Model 4.

paths coefficients were the same in their significance and direction as Indirect effect through trust & self-esteem × trust Estimate 95% CI
in Model 1. The added paths in both Models 2 and 3 were significant, Main sample
β = − 0.29, p b 0.001 for attachment anxiety- trust, β = − 0.15, p = Attachment anxiety → Relativistic Appreciation −0.101⁎ (−0.168,
0.008 for attachment avoidance-self-esteem (Model 2), and β = 0.35, −0.030)
p b 0.001 for self-esteem-trust (Model 3). Attachment anxiety → Comfort with −0.095⁎⁎ (−0.161,
Differences −0.050)
Model 4, combining Models 2 and 3, fit the data well, χ2 (62, N =
Attachment avoidance → Relativistic −0.097⁎ (−0.162,
338) = 147.22, p b 0.001, CFI = 0.96, RMSEA = 0.064 (90% CI = Appreciation −0.045)
0.051, 0.077), SRMR = 0.044, ECVI = 0.775 and better than Model 1, Attachment avoidance → Comfort with −0.076⁎⁎ (−0.124,
Δ χ2 (3, N = 338) = 43.92, p b 0.001, Model 2, Δ χ2 (1, N = 338) = Differences −0.042)
17.64, p b 0.001, and Model 3, Δ χ2 (2, N = 338) = 12.55, p b 0.001, Validation sample
Attachment anxiety → Relativistic Appreciation −0.166⁎⁎ (−0.268,
thus was chosen as the best fitting model. The model paths coefficients
−0.081)
were the same in their significance and direction as in Model 1. The Attachment anxiety → Comfort with −0.123⁎⁎⁎ (−0.217,
added paths were all significant. Differences −0.053)
The indirect effects by self-esteem × trust (i.e., specified as self-es- Attachment avoidance → Relativistic −0.158⁎⁎ (−0.244,
Appreciation −0.090)
teem to trust path) and trust on the attachment-UDO associations
Attachment avoidance → Comfort with −0.115⁎⁎⁎ (−0.198,
were tested using a bootstrapping analysis with 5000 resamples. As Differences −0.058)
displayed in the bias-corrected estimates of the standardized indirect
Note. The indirect effects are not calculated separately for multiple mediators in AMOS.
effects and their 95% confidence intervals (Table 2), all the indirect ef- ⁎ p b 0.05.
fects were significant. Separate indirect effects by self-esteem × trust ⁎⁎ p b 0.01.
and trust were not calculated in AMOS. ⁎⁎⁎ p b 0.001.
S. Han / Personality and Individual Differences 111 (2017) 291–296 295

(UDO cognitive component) and Comfort with Difference (UDO emo- were negative and significant, even after self-esteem and trust ex-
tional component) via low trust, highlighting the centrality of mistrust plained the variance of these associations. This result suggests that
in the attachment insecurity—UDO association. Both anxiously and chronically activated (attachment anxiety) or overly suppressed (at-
avoidantly attached might not be open to diversity because they may tachment avoidance) attachment system (Bowlby, 1988) may compro-
not trust the good nature or intention of the culturally different. This re- mise an exploratory functioning of approaching and appreciating
sult is consistent with the finding that agreeableness—the tendency to cultural diversity. Specifically, attachment anxiety may hinder explora-
be trustful and amicable—was a strong predictor of attitude toward di- tion of diversity due to emotional discomfort, whereas attachment
versity (Strauss et al., 2003). Contrary to the hypothesis, both attach- avoidance may do so with cognitive dismissal of value of diversity. How-
ment anxiety and avoidance (versus avoidance only) were associated ever, these direct paths were not significant (i.e., full mediation by self-
with trust. This result contradicts the theoretical notion that attachment esteem and trust) in the validation sample, warranting further testing.
anxiety IWM includes negative self- and positive other-representations One unexpected yet notable finding was the positive direct associa-
and attachment avoidance IWM includes defensively boosted positive tion between attachment anxiety and Relativistic Appreciation, along
self- and negative other-representations (Bartholomew & Horowitz, with their concurrent negative indirect association. This contradiction
1991) but may be explained by mixed findings on valence of other-rep- may highlight ambivalence toward diversity of the anxiously attached
resentation in attachment anxiety (for a review, see Pietromonaco & (van Oudenhoven & Hofstra, 2006), attachment anxiety possibly being
Barrett, 2000). It is possible that attachment anxiety other-representa- a positive social orientation—at least at a cognitive level—with trust is-
tion may be positive about others' competence in providing protection sues being controlled; they may appreciate the value of diversity but
(Hazan & Shaver, 1987), but negative about their willingness to support not trust good nature of the culturally different people. Also, attachment
(Sirois, Millings, & Hirsch, 2016), whereas attachment avoidance other- anxiety negative association with Comfort with Differences suggested
representation is negative about others' goodwill (Mikulincer & Shaver, internal dissonance of cognitive appreciation but emotional discomfort
2007b). Low trust in this study may have captured negative other-rep- with diversity.
resentation that others are not well-intentioned in general (attachment Noteworthy was that neither attachment anxiety nor avoidance was
avoidance) or in supporting self (attachment anxiety). associated with Diversity of Contacts, UDO behavioral component. One
On the other hand, self-esteem mediated the attachment-UDO asso- could argue that attachment security, defined as low attachment anxi-
ciation only through trust. This result may suggest that self-in-relation- ety and avoidance (Mikulincer & Shaver, 2007b), may not be relevant
to-others representation (Pietromonaco & Barrett, 2000), such as self- to approach behaviors toward diversity. However, Boccato et al.
efficacy in interacting with others, may be more relevant in predicting (2015) found that attachment security, when measured separately
UDO, rather than independent self-representation operationalized as from attachment anxiety and avoidance, was associated with direct
self-esteem. This result is consistent with the findings that self-esteem contacts (i.e., behaviors) with immigrants, whereas attachment anxiety
was not associated with attitude toward diversity (Strauss et al., and avoidance were not. It is possible that attachment security versus
2003), but is contradicted by higher sensitivity to threat among individ- insecurity are independent, and thus are associated with different com-
uals low in self-esteem (Greenberg et al., 1990). Perhaps, diversity may ponents of openness to diversity distinctively.
not be a threat to one's global self-esteem but possibly to self-efficacy in
effective social interactions with diverse people. 5.3. Limitations and future research

5.2. Direct effects This study has limitations. First, this study tested a path model with
cross-sectional data and is correlational. Any causal interpretations
As predicted, the direct attachment anxiety-Comfort with Differ- among the variables should not be made. Longitudinal design such as
ences and attachment avoidance-Relativistic Appreciation associations daily diaries could elucidate the causal relations. Second, generalizability

Fig. 2. The path coefficients of Model 4. Note. Coefficients in regular font (main sample) and in italics (validation sample) *p b 0.05, **p b 0.01 ***p b 0.001.
296 S. Han / Personality and Individual Differences 111 (2017) 291–296

of the results is limited, due to using mostly European American sam- Confer, J. C., Easton, J. A., Fleischman, D. S., Goetz, C. D., Lewis, D. M. G., Perillous, C., & Buss,
D. (2010). Evolutionary psychology: Controversies, questions, prospects, and limita-
ples. Future research should address moderators such as race/ethnicity. tions. American Psychologist, 65, 110–126.
Third, attachment security was not measured directly. Future research Couch, L. L., Adams, J. M., & Jones, W. H. (1996). The assessment of trust orientation.
could examine attachment security association with UDO behavioral Journal of Personality Assessment, 67, 305–323.
Elliot, A. J., & Reis, H. T. (2003). Attachment and exploration in adulthood. Journal of
component (Boccato et al., 2015). Personality and Social Psychology, 85, 317–331.
Fraley, R. C. (2002). Attachment stability from infancy to adulthood: Meta-analysis and
dynamic modeling of developmental mechanisms. Personality and Social Psychology
Review, 6, 123–151.
6. Conclusion Fuertes, J. N., Miville, M. L., Mohr, J. J., Sedlacek, W. E., & Gretchen, D. (2000). Factor struc-
ture and short form of the Miville-Guzman Universality-Diversity scale. Measurement
This study applied attachment theory to openness to diversity, UDO. and Evaluation in Counseling and Development, 33, 157–169.
Greenberg, J., Pyszczynski, T., Solomon, S., Rosenblatt, A., Veeder, M., Kirkland, S., & Lyon,
It was suggested that attachment insecurity was associated with mis-
D. (1990). Evidence for terror management II: The effects of mortality salience on re-
trust of others' good intention and sense of vulnerability in relation to actions to those who threaten or bolster the cultural worldviews. Journal of
others, which then associated with low UDO. Future research could ex- Personality and Social Psychology, 58, 308–318.
amine whether addressing these negative IWMs associated with attach- Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process.
Journal of Personality and Social Psychology, 52, 511–524.
ment insecurities could promote openness to diversity. Hofstra, J., Van Oudenhoven, J. P., & Buunk, B. P. (2005). Attachment patterns and majority
members' attitudes towards adaptation strategies of immigrants. International Journal
of Intercultural Relations, 29, 601–619.
Appendix 1. Structural equation model (SEM) fit indices Jackson, S. E., & Ruderman, M. N. (1995). Introduction: Perspectives for understanding di-
verse work teams. In S. E. Jackson, & M. N. Ruderman (Eds.), Diversity in work teams:
Research paradigms for a changing workplace (pp. 17–46). Washington: American Psy-
chological Association.
Sample Model χ2 df CFI RMSEA (90% CI) SRMR ECVI Little, T. D., Cunningham, W. A., Shahar, G., & Widaman, K. F. (2002). To parcel or not to
parcel: Exploring the question and weighing the merits. Structural Equation
Main Measurement 147.22 62 0.96 0.064 (0.051, 0.044 0.775 Modeling, 9, 151–173.
sample 0.077) Mikulincer, M., & Shaver, P. R. (2001). Attachment theory and intergroup bias: Evidence
Model 1 191.14 65 0.95 0.076 (0.064, 0.089 0.888 that priming the secure base schema attenuates negative reactions to outgroups.
0.088) Journal of Personality and Social Psychology, 81, 97–115.
Model 2 164.86 63 0.96 0.069 (0.056, 0.057 0.822 Mikulincer, M., & Shaver, P. R. (2007a). Attachment in adulthood: Structure, dynamics, and
0.082) change. New York: Guildford Press.
Model 3 159.77 64 0.96 0.067 (0.054, 0.057 0.801 Mikulincer, M., & Shaver, P. R. (2007b). Boosting attachment security to promote mental
health, prosocial values, and intergroup tolerance. Psychological Inquiry, 18, 139–156.
0.080)
Miville, M. L., Gelso, C. J., Pannu, R., Liu, W., Touradji, P., Holloway, P., et al. (1999). Appre-
Model 4 147.22 62 0.96 0.064 (0.051, 0.044 0.775
ciating similarities and valuing differences: The Miville-Guzman Universality-Diversity
0.077) Scale. Journal of Counseling Psychology, 46, 291–307.
Validation Measurement 136.80 62 0.97 0.059 (0.045, 0.043 0.719 Pietromonaco, P. R., & Barrett, L. F. (2000). The internal working models concept: What do
sample 0.072) we really know about the self in relation to others? Journal of General Psychology, 4,
Model 1 202.46 65 0.95 0.078 (0.066, 0.099 0.890 155–175.
0.090) Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton Uni-
Model 2 153.46 63 0.97 0.064 (0.051, 0.053 0.761 versity Press.
0.077) Saleem, M., Prot, S., Cikara, M., Lam, B. C. P., Anderson, C. A., & Jelie, M. (2015). Cutting
Model 3 166.18 64 0.96 0.068 (0.055, 0.066 0.791 Gordian knots: Reducing prejudice through attachment security. Personality and
0.080) Social Psychology Bulletin, 1–15.
Model 4 136.80 62 0.97 0.059 (0.045, 0.043 0.719 Schlomer, G. L., Bauman, S., & Card, N. A. (2010). Best practices for missing data manage-
0.072) ment in counseling psychology. Journal of Counseling Psychology, 57, 1–10.
Schumacker, R. E., & Lomax, R. G. (2010). A beginner' guide to structural equation modeling
Note. df = degrees of freedom; CFI = Comparative Fit Index; RMSEA = Root Mean Square (3rd ed.). Mahwah, NJ: Lawrence Erlbaum Associates.
Error of Approximation (RMSEA); CI = confidence interval; SRMR = Standardized Root Sirois, F. M., Millings, A., & Hirsch, J. K. (2016). Insecure attachment orientation and well-
Mean Square Residual; ECVI = Expected Cross-Validation Index. being in emerging adults: The roles of perceived social support and fatigue.
Personality and Individual Differences, 101, 318–321.
Stephan, W. G., Ybarra, O., & Rios Morrison, K. (2009). Intergroup threat theory. In T. D.
Nelson (Ed.), Handbook of prejudice, stereotyping, and discrimination (pp. 43–60).
References Mahwah, NJ: Lawrence Erlbaum Associates.
Strauss, J. P., Connerley, M. L., & Ammerman, P. A. (2003). The “threat hypothesis,” per-
Baldwin, M. W. (1992). Relational schemas and the processing of social information. sonality, and attitudes toward diversity. The Journal of Applied Behavioral Science, 39,
Psychological Bulletin, 112, 461–484. 32–52.
Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles among young adults: A test Tsui, A. S., Egan, T. D., & O'Reilly, C. A., III (1992). Being different: Relational demography
of a four category model. Journal of Personality and Social Psychology, 61, 226–244. and organizational attachment. Administrative Science Quarterly, 37, 549–579.
Boag, E. M., & Carnelley, K. B. (2012). Self-reported discrimination and discriminatory be- Van Oudenhoven, J. P., & Hofstra, J. (2006). Personal reactions to ‘strange’ situations: At-
havior: The role of attachment security. British Journal of Psychology, 51, 391–403. tachment styles and acculturation attitudes of immigrants and majority members.
Boag, E. M., & Carnelley, K. B. (2015). Attachment and prejudice: The mediating role of International Journal of Intercultural Relations, 30, 783–798.
empathy. British Journal of Social Psychology, 55, 337–356. Wearden, A., Peters, I., Berry, K., Barrowclough, C., & Liversidge, T. (2008). Adult attach-
Boccato, G., Capozza, D., Trifiletti, E., & Di Bernardo, G. A. (2015). Attachment security and ment, parenting experiences, and core beliefs about self and others. Personality and
intergroup contact. Journal of Applied Social Psychology, 45, 629–647. Individual Differences, 44, 1246–1257.
Bowlby, J. (1988). A secure base: Parent-child attachment and healthy human development. Wei, M., Russell, D. W., Mallinckrodt, B., & Vogel, D. L. (2007). The experiences in close re-
New York: Basic Books. lationship scale (ECR)-short form: Reliability, validity, and factor structure. Journal of
Chang, M. J. (2001). Is it more than about getting along?: The broader educational impli- Personality Assessment, 88, 187–204.
cations of reducing students' racial biases. Journal of College Student Development, 42, Westland, J. C. (2010). Lower bounds on sample size in structural equation modeling.
93–105. Electronic Commerce Research and Applications, 9, 476–487.
Cole, T., & Teboul, J. C. B. (2004). Non-zero-sum collaboration, reciprocity, and the prefer-
ence for similarity: Developing an adaptive model of close relational functioning.
Personal Relationships, 11, 135–160.

You might also like