Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2011 Apr 28.
Published in final edited form as: J Soc Clin Psychol. 2006 Apr 1;25(4):404–428. doi: 10.1521/jscp.2006.25.4.404

Psychological and behavioral implications of connectedness to communities with opposing values and beliefs

Debra Mashek 1, Jeffrey Stuewig 1, Emi Furukawa 1, June Tangney 1
PMCID: PMC3084011  NIHMSID: NIHMS288543  PMID: 21532983

Abstract

Without a doubt, people can feel simultaneously connected to multiple communities (e.g., Deaux, 1993; Roccas & Brewer, 2002). But, to what degree can people feel simultaneously connected to communities with opposing beliefs and values? And, more importantly, what are the psychological implications of being dually connected to these communities? Capitalizing on a sample of individuals positioned to potentially feel connected to two very distinct communities, we examined jail inmates’ (N = 256) sense of connectedness to the criminal community and to the community at large. Results indicated that (a) connectedness to the community at large is orthogonal to connectedness to the criminal community, supporting the supposition that it is possible to be dually connected to opposing communities; and (b) connectedness to the community at large moderated the relationship between criminal connectedness and indicators of psychological distress, suggesting that connectedness to the criminal community is especially problematic when it occurs in tandem with connectedness to the community at large. These findings are consistent with predictions from the self-expansion model.

Keywords: Connectedness, Psychological distress, Self-expansion

How do individuals psychologically experience multiple connections when those connections are in conflict with one another? The current paper explores this question, deriving predictions from the self-expansion model (Aron & Aron, 1986). By applying the theoretical construct of inclusion of others in the self, we suggest that community connectedness can be appropriately conceptualized as the inclusion of community in self. Moreover, we argue that the community domain offers an apropos venue for testing predictions about the psychological consequences of including contrasting others in the self.

The jail setting offers an ideal opportunity to evaluate this basic social psychological question. Jail inmates are positioned to feel connected to two very distinct and potentially conflicting communities: the community at large and the criminal community. In addition to evaluating the extent to which jail inmates can simultaneously include both of these communities in the self, we examined whether particular patterns of inclusion were predictive of elevated psychological distress and self-reported problematic behaviors. The relationship between connectedness and indicators of distress is both theoretically and clinically important. On a theoretical level, although we all include many targets in the self (e.g., our children, our partners, our ethnic ingroups), little is known about how people negotiate the inclusion in the self of multiple conflicting targets. Clinically, understanding the psychological toll of conflicting connections may inform treatment and interventions for the 2.1 million Americans behind bars (Bureau of Justice Statistics, 2003a), especially given the profound, interconnecting links between psychological distress, drug use, and crime.

Community Connectedness as the Inclusion of Community in Self

Our social connections are significant and integral features of the self (Tajfel, 1981; Turner & Oakes, 1989). Who we are is largely a function of those around us and, as stated by Hogg (2003), social relationships “influence the type of people we are, the things we do, the attitudes and values we hold, and the way we perceive and react to people around us” (pg. 462).

The theoretical framework provided by the self-expansion model (Aron & Aron, 1986) speaks directly to this issue of self-definition through connection. This model, in conjunction with mounting empirical evidence (see Aron, Mashek & Aron, 2004 for a review), suggests that through relationships we come to take on the resources, perspectives, and identities of others (Aron & Aron, 1986).

Building on strategies used by Lewin (1948), Levinger and Snoek (1972), and others, Aron, Aron, and Smollan (1992) captured the intuitive essence of interconnected selves with their Inclusion of Others in Self (IOS) scale. The IOS scale consists of seven pairs of overlapping circles, each pair overlapping slightly more than the preceding pair. When participants are asked to indicate which pair of circles best describes their relationships with other individuals (e.g., parent, romantic partner), they do so easily and in a meaningful way. In fact, when asked to respond to an open-ended question about the subjective meaning of the circle diagrams, 86% of the respondents in a validation study generated connectedness-themed descriptors (see Study 1 of the Subjective Meaning of the IOS Scale section, Aron et al., 1992).

Given that these overlapping circle diagrams “portray so vividly the sense of interpersonal connectedness” (Aron et al., 1992, p. 610), it is not surprising that researchers have successfully adapted this measurement tool, and the underlying theoretical notion of including others in the self, to other research domains. Uleman, Rhee, Bardoliwalla, Semin, and Toyama (2000) adapted the IOS Scale to assess relationships with immediate family, other relatives, and close friends, comparing responses of U.S., Dutch, Turkish, and Japanese respondents. Blanchard, Perreault, and Vallerand (1998) developed a version for use in the sports domain, in which respondents indicated their relationship with their coach, their team, and the game in general (and showed reliable links of responses to various sports-relevant variables). In perhaps the broadest adaptation, Schultz (2002) had respondents indicate their relationships with the natural environment, which has proven to have substantial relationships with ecological concerns.

Work done by Tropp and Wright (2001) on the inclusion of ingroups in the self represents the adaptation most relevant to the current paper. These researchers found convincing evidence to support their hypothesis that the degree to which an ingroup is included in the self underlies varied definitions of ingroup identification including interdependent self-construals, racial and ethnic identity, collective identity, and collective self-esteem. Echoing Aron et al. (1992), Tropp and Wright argue that “the visual nature of the [measurement tool] might make it an especially effective measure of ingroup identification because the overlapping circles reflect the essence of connectedness to an ingroup” (p. 588).

Consistent with both the intuitive appeal of what it means to feel connected to others and the growing evidence that the inclusion of others in the self is at the root of other relational phenomenon (e.g., closeness, identification), we propose that community connectedness is the inclusion of community in self.

We believe that this definition of community connectedness captures the areas of commonality among other relevant social and community constructs. Although we are unaware of other data that speak directly and specifically to the inclusion of community in self, extant work on other social and community constructs may help elucidate the importance of community connectedness.

Research on a construct closely related to, but more general than, community connectedness suggests the benefit of feeling a part of the social world. Social connectedness captures a sense of “interpersonal closeness to the social world in toto” (Lee, Draper, & Lee, 2001, p. 310; emphasis added) – a social world that includes family, friends, peers, strangers, community, and society. Studies have shown social connectedness to be a powerful predictor of lower trait anxiety, and this is true even after one controls for such resiliency factors as perceived social support and collective self-esteem (Lee & Robins, 2000). Moreover, college students with higher social connectedness perceive less stress in their daily lives (Lee, Keough & Sexton, 2002). Social connectedness is also an important precursor of perceived social support, and this relationship is especially notable in light of findings that social support is significantly positively related to both physical and psychological health (e.g., Uchino, Cacioppo & Kiecolt-Glaser, 1996). In addition to its direct and important implications for psychological well-being, social connectedness has been shown to have inverse correlations with dysfunctional interpersonal behaviors (e.g., being domineering and stand-offish), which in turn are associated with such psychological dysfunction as depressive symptoms, hostility, and social discomfort (Lee at al., 2001).

Research on other related constructs likewise suggests that a sense of social connection is a desirable thing. For example, psychological sense of community is related to active participation in both religious and neighborhood functions, as well as residence in areas with large percentages of registered voters and high participation in neighborhood events (Brodsky, O’Campo, & Aronson, 1999). Furthermore, individuals are more likely to engage in prosocial helping behaviors when they feel a connection to the beneficiary of that helping (e.g., Cialdini, Brown, Lewis, Luce, & Neuberg, 1997; West, 2003).

Connectedness to multiple communities

Although people can feel a simultaneous connection to multiple others (e.g., Pretty, Conroy, Dugay, Fowler, & Williams, 1996; Roccas & Brewer, 2002; Sidanius, Feshbach, Levin, & Pratto, 1997), little research has examined the psychological implications of being simultaneously connected to others typified by opposing beliefs and values. This paper focuses on the case of simultaneous connectedness to communities with opposing beliefs and values.

Theoretically, simultaneous connectedness to opposing communities should be associated with a less than favorable constellation of psychological functioning. After all, when one includes two targets in the self, one takes on the resources, perspectives, and identities of both targets (Aron & Aron, 1986; Tajfel, 1981). Both positive and negative attributes of the other should become integrated into the self. If the attributes central to one included target are in opposition to those of the other, then the self comes to include incompatible values and beliefs. Dissonance theory (Festinger, 1957) and other self-consistency approaches (e.g., Heider, 1946; Higgins, 1999) assert that the internal struggle to negotiate juxtaposed realities of the self should weigh heavily on the psyche, inducing psychological distress and subsequent coping behaviors. Of note is the observation that these self-consistency perspectives apply only once one assumes that communities and their values become part of the self as cognitive elements that can be dissonant.

Extant research offers initial support for this prediction. For example, individuals negotiating multiple cultural identities often report elevated psychological distress (for a review, see LaFromboise, Coleman, & Gerton, 1993). In their recent work on bicultural identity integration, Benet-Martinez and her colleagues found that bicultural individuals who perceive their two cultures as being incompatible (i.e., low bicultural identity integration) reported elevated neuroticism, depression, and anxiety (Benet-Martinez & Haritatos, 2004). Other researchers report similar identity conflict when individuals attempt to reconcile particular religious identities (e.g., Jewish, Catholic) with gay identities (Coyle & Rafalin, 2000; Toman, 1997).

Present Research

Thus, there are theoretical and empirical reasons for predicting that dual connectedness to communities with opposing beliefs and values should be associated with elevated psychological distress. We build on both the conceptual and methodological traditions of the self-expansion model to examine two questions about these connections. First, to what degree can people feel simultaneously connected to communities with opposing beliefs and values? (Question 1.) Second, and more importantly, what are the psychological implications of being dually connected to conflicting communities? (Question 2.) The first question was necessary to answer before moving to the second question, which served as a key theoretical thrust of this paper. Based on the self-expansion model’s notion of including others in the self, we predicted (Hypothesis 1) that simultaneous connectedness with communities associated with incompatible beliefs and values would correspond with elevated psychological distress and problematic behaviors.

We drew on data from an ongoing study of jail inmates, individuals positioned to feel connected to two very distinct—and potentially conflicting—communities: the community at large and the criminal community. Although the primary focus of this ongoing study is on the moral emotions of incarcerated offenders, appropriate data were nevertheless available to allow adequate exploration of the questions at the heart of this paper.

We evaluated three indicators of psychological distress: anxiety, depression, and self-esteem. We evaluate anxiety and depression because (a) methodologically, these constructs can by reliably measured via self-report inventories, (b) they are among the very top reasons individuals seek treatment, both in and out of the correctional setting, and (c) Clark & Watson’s (1991) tripartite model of emotion suggests that there is a general distress factor (“negative affect”) that is associated with both anxiety and depression. We included self-esteem as a possible indicator of a lack of psychological distress owing to the idea that self-esteem is associated with fewer negative attributions following negative events. For example, high self-esteem would protect individuals from negative self-appraisals, cognitive distortions, and other processes that can contribute to anxiety and depression. We further examined self-reports of drug and alcohol abuse as behavioral corollaries of distress.

Methods

Participants

These data were extracted from an on-going study of moral emotions and criminal recidivism. Participants were a non-random sample of 258 male1 inmates recruited from the “general population” (e.g., not people who are severely mentally ill or in solitary confinement) of a county-administered Adult Detention Center (ADC). The ADC, located outside Washington DC, houses approximately 1,000 pretrial and post-trial individuals convicted of a range of misdemeanor and felony offenses. Participants in this study were deemed likely to remain in the jail’s custody for at least four months (typically held on a felony charge and with at least a $7,000 bond). Participants who volunteered for the study received assurance that the information they provided would be kept strictly confidential, protected by a Certificate of Confidentially from the Department of Health and Human Services. Participants received a $15 honorarium2.

Participants were racially diverse (44.2% African American, 33.7% Caucasian, 11.2% Hispanic, 3.5% mixed race, 3.5% Asian/ Pacific Island, 0.8% Middle Eastern, 0.8% Native American, and 2.3% “other”) and varied in age from 18 to 60 years (M = 31.1, SD = 9.3).

Two participants were removed from all analyses because of an apparent response set on the majority of the measurement scales, in addition to their being extreme univariate and multivariate outliers on many of the distributions and scatterplots. One additional participant was a univariate outlier on measures of anxiety and depression. In addition, this individual was an obvious bivariate outlier on scatterplots of these variables with (a) connectedness to the community at large, and (b) connectedness to the criminal community. Examination of item responses and previously documented validity concerns did not reveal an obvious reason to discard this individual’s data. To limit the influence of this one case on future analyses, a decision was made a priori to recode this participant’s scores on the anxiety and depression scales such that he remained the highest scoring participant in the distribution, with the degree of difference from the next lowest participant reduced (Tabachnick & Fidell, 1998). Thus, the final N for most analyses was 2563.

Materials

Connectedness

Community connectedness was measured using the Inclusion of Community in Self (ICS) scale (Mashek, Cannaday & Tangney, 2005). The ICS is a single-item pictorial measure consisting of six pairs of overlapping circles, with each pair of same-sized circles overlapping slightly more than the preceding pair (see Figure 1). Participants are told that each circle on the left of the pair represents themselves, while the circle on the right represents the target (i.e. community at large, criminal community). Connectedness to the community at large was assessed by asking participants to “circle the picture that best describes your relationship with the community at large.” Connectedness to the criminal community was assessed by asking participants to “circle the picture that best describes your relationship with the criminal community.” If participants asked something akin to “What do you mean by community at large,” we said that the community at large refers to all the people in your town, city or county; people in general; people who live on the outside and who do not commit crimes. We defined the criminal community as people who commit crimes whether they are in jail, prison, or living on the outside. Our definitions were intentionally vague to allow idiographic interpretations of the labels “community at large” and “criminal community.” Any noise introduced by this approach should work against our ability to identify statistical effects. The ICS scale is an explicit derivative of Aron et al.’s (1992) Inclusion of Other in Self Scale (IOS), which has excellent test-retest reliability (.83 over two weeks), alternate form validity (r = .93), convergent validity (correlating .25 with “behaving close” and .26 with “feeling close”), predictive validity (r = .45 with whether or not close relationships dissolved in the 3 months after participants completed the measure), and has been shown to be minimally associated with impression management (r = .05).

Figure 1.

Figure 1

Open in a new tab

Inclusion of Community in Self Scale (where ‘S’ = Self and ‘C’ = Community).

The community at large item of the ICS appears to be construct valid in college samples (Mashek et al., 2005), demonstrating the expected moderate correlations with relevant facets of psychological sense of community as assessed by Obst, Smith, and Zinkiewicz (2002). The ICS correlated .47 with the conscious identity items (e.g., “Belonging to my community is an important part of who I am”) used by Obst et al., (2002). Further, the ICS correlated .30 with “ties and friendship” (e.g., “If I need a little company, I can contact a neighbor I know”), .30 with “support” (e.g., “If there was a serious problem in my local community, people who live in it could get together and solve it”), and .28 with “belongingness” (e.g., “I plan to remain a resident of my local community for a number of years”). The ICS did not much correlate with “perceived influence” (r = .10; e.g., “I have almost no influence over what my community is like”). Discriminant validity was evidenced by the ICS’s minimal correlations with agreeableness (r = .09; Saucier, 1994) and impression management (r = −.02; Paulhus, 1988). Furthermore, the test-retest reliability of the ICS over a two-week period was acceptable (r = .78; n = 50). The college sample showed inadequate variance on the criminal connectedness item, thus precluding analysis of the validity of that item.

Although reliance on a single-item measure of any construct is less than optimal, we proceeded with this approach in the current study for a number of reasons. First, connectedness was not a focus of the larger parent study; thus, it was not feasible for us to add longer measures to the already packed protocol. Second, as reported above, the overlapping circles approach has been demonstrated in other research both to capture the sense of connectedness we hoped to tap and to prospectively predict important relationship outcomes (Aron et al., 1992); in fact, the overlapping circles measure outperformed substantiated, multi-item measures of closeness in predicting relationship breakup prospectively over the three months following administration. Third, the community at large item of the ICS corresponds in expected ways with complex measures of psychological sense of community, a construct related to connectedness as we define it (Mashek et al., 2005).

Values and Beliefs

Preliminary to our evaluation of Question 1, we sought to confirm our assumption that the values and beliefs associated with connectedness to the community at large were different than those associated with connectedness to the criminal community. Our confirmation focused on both negative and positive attributes, namely criminogenic beliefs and character strengths.

We assessed criminogenic beliefs using a preliminary version of the Criminogenic Beliefs and Assumptions Scale, which was designed for the larger project (CBAS; Tangney, Meyer, Furukawa, & Cosby, 2002). The CBAS contains 40 items rated on a four-point Likert scale where “1” equals “strongly disagree” and “4” equals “strongly agree.” CBAS items were created based on the experience of trained clinicians who work with jail inmates. The CBAS taps five types of criminogenic beliefs: failure to accept responsibility (e.g., “I feel like what happens in my life is mostly determined by powerful people”), notions of entitlement (e.g., “I deserve more than other people.”), negative attitudes toward authority (e.g., “People in position of authority generally take advantage of others.”), short-term orientation (e.g., “Why plan to save for something if you can have it now”), and insensitivity to impact of crime (e.g., “My crime did not really harm anyone.”). Each subscale contains 8 items. The internal consistencies of the five subscales were favorable, ranging from .58 to .80. A second, more positive, set of beliefs and values were assessed using the Values in Action Inventory of Strengths (VIA-IS; Peterson & Seligman, 2001), recently developed based on current theory and research in positive psychology. The nine selected subscales assessed originality, judgment, integrity/honesty, optimism/hope, industry and perseverance, gratitude, spirituality, generosity/kindness, and forgiveness. We asked four to eight items from each subscale, based on the appropriateness for an incarcerated sample. Participants indicated on a 1 to 5 scale whether each statement was “Not at all like me” “Unlike me,” “Sometimes like me,” “Like me,” or “Very much like me.”

Because some of the original nine subscales appeared closely linked conceptually, we a priori derived six indicators of character from these scales: (a) judgment (alpha = .82; e.g, “I always examine both sides of an issue.”), (b) originality (alpha = .81; e.g, “I like to think of new ways to do things”), (c) integrity (alpha = .79; e.g, “I always keep my promises.”), (d) optimism/perseverance (alpha = .88; e.g, “I know that I will succeed with the goals I set for myself”), (e) gratitude/spirituality (alpha = .87; e.g, “I feel thankful for what I have received in life.”), and (f) generosity/forgiveness (alpha = .84; e.g, “I am usually willing to give someone another chance.”). For example, generosity and forgiveness scales both appeared to measure tolerance and warm interpersonal strength. Similarly, optimism and perseverance scales seemed to assess a character strengths that help individuals to work toward future goals persistently.

Psychological Distress and Problem Behaviors

Question 2 focused on the interactive properties of dual connectedness in terms of psychological distress. We assessed three general indicators of psychological distress: level of self-esteem, depression, and anxiety. We also assessed drug and alcohol problems as behavioral indicators of psychological distress.

Level of self-esteem was assessed using the Rosenberg Self-Esteem Scale (Rosenberg, 1965). The level of self-esteem scale contains 10 items to which participants respond using a 5-point likert scale where “1” equals “always false” and “5” equals “always true” (alpha = .87

Indicators of anxiety and depression were drawn from the Personality Assessment Inventory (PAI; Morey, 1991) which assesses dimensions of personality and symptoms of psychopathology. For each of the 24 items on each of the scales, participants are asked to indicate whether the statement is “false, not at all true,” “slightly true,” “mainly true,” or “very true.” Sample items from the anxiety scale include “I usually worry about things more than I should” and “Sometimes I am afraid for no reason.” Sample items from the depression scale include “I feel that I’ve let everyone down” and “Much of the time I’m sad for no real reason.” Based on data from the larger study (N = 357), internal consistencies were .89 for anxiety and .89 for depression. We employed community-based T-scores throughout this report to aid in the comparison of these data with those drawn from other samples.

We were also interested in whether particular problem behaviors might be associated with the interaction between connectedness to the community at large and connectedness to the criminal community. Using PAI scales, we examined participants’ alcohol- and drug- related problem behaviors. The scales for alcohol problems and drug problems each consist of 12 items, which focus directly on problematic consequences of substance use and features of dependence (e.g., “My drinking seems to cause problems in my relationships with others.” “I’ve had health problems because of my drug use.”) The internal consistencies of the alcohol and drug scales were .92 and .90, respectively.

Procedure

Participants individually completed all assessments, which were facilitated by a trained research assistant. Ratings of criminogenic beliefs, character strengths, psychological distress, and problem behaviors were obtained in the second and third of four one to two-hour Phase I sessions4. All the scales, except for the ICS, were completed using a touch-screen laptop computer5. A pre-recorded voice accompanied each on-screen item, making it possible for the participant to both hear and read the questions and response options. Ratings of connectedness were obtained in the fourth of the Phase I sessions. Trained clinical students administered the ICS measure after completing an in-depth semi-structured social-clinical interview.

Results

Means, standard deviations, Ns and ranges are shown in Table 1.

Table 1.

Sample size, means, standard deviations, possible range, and observed range for study variables

N M SD Possible
Range 		Observed
Range
Community Connectedness
 Connectedness to the community at large 256 2.61 1.59 1-6 1-6
 Connectedness to the criminal community 256 2.86 1.78 1-6 1-6
Criminogenic Beliefs
 Failure to accept responsibility 255 2.16 0.53 1-4 1.00-3.83
 Notions of entitlement 255 2.36 0.41 1-4 1.38-3.63
 Negative attitudes toward authority 255 2.37 0.52 1-4 1.13-3.75
 Short-term orientation 255 2.20 0.45 1-4 1.00-3.43
 Insensitivity to impact of crime 254 2.02 0.42 1-4 1.00-3.00
Character Strengths and Values
 Originality 255 3.84 0.65 1-5 1.71-5.00
 Judgment 255 3.68 0.62 1-5 1.75-5.00
 Integrity and Honesty 255 4.09 0.57 1-5 2.33-5.00
 Gratitude and Spirituality 255 3.86 0.80 1-5 1.50-5.00
 Hope and Perseverance 255 3.90 0.55 1-5 2.06-5.00
 Generosity and Forgiveness 255 3.90 0.57 1-5 2.30-5.00
Psychological Distress
 Level of Self-Esteem 255 3.88 0.69 1-5 2.20-5.00
 Anxiety 251 21.09 11.28 0-72 1.00-61.00
 Depression 251 21.37 12.09 0-72 0.00-60.00
Problem Behaviors
 Alcohol problems 249 11.85 10.07 0-36 0.00-36.00
 Drug problems 250 14.31 9.82 0-36 0.00-36.00
Open in a new tab

Question 1: To what degree can people feel simultaneously connected to communities with opposing beliefs and values?

Question 1 assumes that the communities of interest are in fact associated with opposing beliefs and values. Thus, before answering Question 1, it was first useful to establish empirically that connectedness to the community at large is associated with beliefs and values unlike those associated with connectedness to the criminal community. We were particularly interested in whether connectedness to different communities might be differentially associated with criminogenic beliefs and character strengths.

As shown in Table 2, connectedness to the community at large was (a) not much associated with criminogenic beliefs (the correlations ranged from .00 to −.16, with an average effect of −.05) and, (b) positively associated with character strengths (the correlations ranged from .13 to .26, with an average effect of .19). Connectedness to the criminal community, on the other hand, was (a) positively associated with criminogenic beliefs (correlations ranged from .10 to .29, with an average effect of .18), and (b) typically negatively associated with character strengths (−.03 to −24, with an average effect of −.14).

Table 2.

Bivariate correlations and t-tests for dependent means for criminogenic beliefs and character strengths

Bivariate Correlations with Connectedness
 Dependent t tests
N Community at Large Criminal Community t p
Criminogenic Beliefs
 Failure to accept responsibility 255 −.05 .23** 3.21 < .001
 Notions of entitlement 255 .00 .11 1.12 ns
 Negative attitudes toward authority 255 −.16* .29** 5.32 < .001
 Short-term orientation 255 −.02 .19** 2.38 < .05
 Insensitivity to impact of crime 254 −.03 .10 1.46 ns
Character Strengths and Values
 Originality 255 .13* −.03 −1.80 < .10
 Judgment 255 .18** −.15* −3.79 < .001
 Integrity and Honesty 255 .26** −.07 −3.82 < .001
 Gratitude and Spirituality 255 .19** −.20** −4.53 < .001
 Hope and Perseverance 255 . 24** −.14* −4.41 < .001
 Generosity and Forgiveness 255 .14* −.24** −4.41 < .001
Open in a new tab
*

p < .05

**

p < .01

Note: df for all t-tests = n – 3

In addition to testing whether the correlations were different from zero we also tested whether the correlates varied as a function of type of community. For example, was the correlation between connectedness to the community at large and failure to accept responsibility significantly different from the correlation between connectedness to the criminal community and a failure to accept responsibility? Results from t-tests for differences between dependent correlations (Cohen & Cohen, 1983) are presented in the last two columns of Table 2. With the exception of sense of entitlement and insensitivity toward the impact of crime, criminogenic beliefs showed a significantly different pattern of relationships with connectedness to the community at large as compared to connectedness to the criminal community.

For character strengths there was a similar differential pattern (see Table 2). Once again the t-test for dependent correlations showed that correlations with the character strengths were significantly different (except originality, which trended toward significance) between the two types of connectedness. Specifically, connectedness to the community at large correlated significantly positively with character strengths, whereas connectedness to the criminal community correlated negatively with these same strengths.

Thus, in general, connectedness to the community at large was associated with beliefs and values unlike those associated with connectedness to the criminal community. Having established this preliminary point, we returned to Question 1: To what degree can people feel simultaneously connected to communities with opposing beliefs and values? A simple bivariate correlation between connectedness to the community at large and connectedness to the criminal community showed a near-zero relationship between the two (r = −.01). Furthermore, an inspection of the scatterplot between the two variables revealed no discernable pattern. In this sample, a person’s connectedness to one community had no bearing on whether one felt connected to the other. In other words, it is possible to be connected to both communities, just one community, or neither community.

Question 2: What are the psychological implications of being dually connected to conflicting communities?

Having demonstrated that (a) connectedness to the community at large is associated with different beliefs and values than connectedness to the criminal community, and (b) connectedness to one community is orthogonal to connectedness to the other community, we were in a position to examine Question 2: What are the psychological implications of being dually connected to conflicting communities? Based on the self-expansion model, Hypothesis 1 predicted that dual connectedness to these two communities would be associated with elevated psychological distress and problem behaviors.

Following the steps outlined by Aiken and West (1991), we evaluated whether connectedness to the community at large moderated the relationship between connectedness to the criminal community and the dependent variables of interest (i.e., anxiety, depression, self-esteem, drug problems, and alcohol problems). In predicting each DV separately, the centered scores of connectedness to the community at large and connectedness to the criminal community were entered in Step 1 and the product of the two were entered in Step 2.

In terms of main effects, high levels of connectedness to the community at large was significantly associated with lower drug problems (beta = −.14, p = .03) and marginally significantly associated with lower depression (beta = −.11, p = .09). In contrast, high levels of connectedness to the criminal community was significantly associated with higher drug problems (beta = .22, p = .001) and marginally significantly associated with higher depression (beta = .11, p = .08) and alcohol problems (beta = .10, p = .10). Table 3 shows the R2, B, SE, and beta values for the psychological distress and behavior problem variables.

Table 3.

Regression information for significant interactions between connectedness to the community at large (CAL) and connectedness to the criminal community (CC)

Change in R2 B SE Beta
Level of Self-Esteem
 Step 1: .01
  CAL .03 .03 .07
  CC −.02 .03 −.06
 Step 2: .02*
  CAL X CC −.04 .02 −.15*
Depression
 Step 1: .02*
  CAL −.81 .47 −.11†
  CC .76 .43 .11†
 Step 2: .01†
  CAL X CC .46 .26 .11†
Anxiety
 Step 1: .005
  CAL −.29 .45 −.04
  CC .38 .40 .06
 Step 2: .005
  CAL X CC .28 .25 .07
Drug Problems
 Step 1: .07***
  CAL −.82 .38 −.14*
  CC 1.22 .34 .22***
 Step 2: .01*
  CAL X CC .41 .21 .12*
Alcohol Problems
 Step 1: .01
  CAL −.09 .40 −.01
  CC .59 .36 .10†
 Step 2: .005
  CAL X CC .26 .21 .07
Open in a new tab

Note: Information in table is based on regressions with centered variables; CAL = Connectedness to the community at large; CC = Connectedness to the criminal community; Step 1 statistics are for the simple effects, not controlling for the interaction.

†

p < .10

*

p < .05

**

p < .01

***

p < .001.

The main effects, however, were moderated by a number of interactions. Connectedness to the community at large moderated the relationship between criminal connectedness and depression, self-esteem, and drug use. Figures 2, 3, and 4 show plots of the significant interactions. Specifically, when individuals were not connected to the community at large, being connected to the criminal community was not significantly associated with important indicators of psychological distress such as depression (simple slopes beta = .01, p = .88) and self-esteem (simple slopes beta = .08, p = .36), nor with problematic drug use (simple slopes beta = .11, p = .18). However, when individuals were connected to the community at large, also being connected to the criminal community was associated with low self-esteem (simple slopes beta = −.22, p = .02), high depression (simple slopes beta = .23, p = .01), and elevated drug problems (simple slopes beta = .34, p = .001). Connectedness to the community at large did not moderate the relationship between anxiety or alcohol problems and connectedness to the criminal community.

Figure 2.

Figure 2

Open in a new tab

Interaction of connectedness to the community at large (CAL) and connectedness to the criminal community (CC) on level of self-esteem.

Figure 3.

Figure 3

Open in a new tab

Interaction of connectedness to the community at large (CAL) and connectedness to the criminal community (CC) on depression.

Figure 4.

Figure 4

Open in a new tab

Interaction of connectedness to the community at large (CAL) and connectedness to the criminal community (CC) on drug problems.

Discussion

Working within the theoretical framework offered by the self-expansion model (Aron & Aron, 1986), this study explored two questions regarding community connectedness. First, to what degree do people feel simultaneously connected to two very different types of communities? Second, and more importantly, what are the psychological implications of being dually connected to communities associated with conflicting beliefs and values? These issues were explored using data from a sample of jail inmates, individuals distinctively positioned to potentially feel connected to both the community at large and the criminal community.

Question 1: To what degree can people feel simultaneously connected to communities with opposing beliefs and values?

Given the interesting, although not surprising, finding that these two communities were indeed associated with contrasting values and beliefs, it is striking that connectedness with the community at large was unrelated to connectedness with the criminal community. In the context of the self-expansion model, these data are also the first direct evidence that individuals can simultaneously include in the self targets associated with disparate beliefs and values. This finding is both consistent with and an extension of research and theory from the social identity literature (e.g., Deaux, 1993; Roccas & Brewer, 2002; Sidanius et al., 1997) in that it (a) shows once again that social identities are complex and multi-dimensional, and (b) makes clear that seemingly incompatible social identities can be simultaneously present. In contrast, the pattern of findings do not seem compatible with theoretical perspectives that suggest individuals have a general propensity to either “connect” or “disconnect” (e.g., Hirschi, 1969).

Question 2: What are the psychological implications of being dually connected to conflicting communities?

The finding of orthogonal connectedness opened the door for exploring the important question concerning psychological implications of connectedness to opposing communities. The data provided support, modest in magnitude but evidenced across multiple outcomes, for our prediction that simultaneous connectedness with communities holding incompatible beliefs and values would correspond with elevated psychological distress and problematic behaviors including low self-esteem, high depression, and elevated drug problems. (It is notable that the pattern of effects was consistent across all of the indicators of psychological distress and problem behavior.) In short, it seems that individuals who are dually connected to discrepant communities may be at a somewhat elevated risk for psychological distress and problematic drug use.

These findings are consistent with predictions from the self-expansion model. The model argues that when one includes others in the self, one takes on the perspectives of those others. When one includes multiple targets, the attributes of the multiple targets come to be part of the self – even if the attributes of one target are in opposition to those of the other target.

From a cognitive dissonance perspective, inclusion of these opposing values and beliefs should contribute to psychological distress. Of course, psychological distress is multiply determined (e.g., by everything from genetics to diet, from environmental stressors to brain functioning). We only looked at one possible social psychological predictor of the clinical phenomenon of distress. The fact that the interaction of two single-item assessments of connectedness was able to predict any of the variance in self-reports of a multiply determined construct such as psychological distress is compelling (Cicchetti & Rogosch, 1996), especially when one considers the challenges of detecting interactions among continuous indicators (Cortina & DeShon, 1998).

Caveats and Future Directions

Although the generalizability of these results is necessarily limited by the use of a distinctive sample of incarcerated males, they nevertheless point to important theoretical and practical issues, as well as highlight areas for continued research. Of course, a replication of this pattern of findings with people possessing other kinds of opposing connections (e.g., individuals in committed and loving relationships who engage in extra-dyadic affairs; pacifists working in the Department of Defense) would aid in interpreting the generalizability of these results.

To the extent that these findings are generalizable, they may shed light on the challenges faced historically by people trying to connect to two or more communities with opposing sets of expectations, values, beliefs and assumptions. Consider, for example, the case of soldiers returning to the United States from the Vietnam war. Those individuals who tried to reintegrate into the then dovish “mainstream” society, while acknowledging their participation in the war effort, may have been the most likely to experience significant psychological angst upon their home coming.

Longitudinal research on the interplay between connectedness to different communities and the psychological and behavioral concomitants of this connectedness will offer insights on the processes involved in negotiating connections to opposing communities. For example, to what extent does the salience of one’s connections play a role in the experience of dual connectedness? Although the data reported here focus solely on inmates incarcerated during the previous two weeks, we anticipate evaluating longitudinal change in these connections and the corresponding psychological changes. It could be that inmates who are in regular contact with family, friends, and volunteers from the “outside” are apt to be especially aware of their connectedness with the community at large. Such saliency may enhance the relationship between dual connectedness and psychological distress (c.f., Fazio & Towles-Schwen, 1999). Of particular interest will be the accounts of ex-inmates. Once released from jail, do these individuals continue to feel dually connected? If so, do they remain psychologically distressed?

Theorists may advance scientific understanding of identity by considering more directly the intrapsychic properties of holding multiple connections. Particularly rich research questions may emerge from considering the intersection of identity structure (e.g., how does an individual perceive her or his different connections to be related to one another) and the implications of these connections (e.g., how do different identities interact to create psychological health or distress?). For example, identity structure (e.g., intersection, dominance, compartmentalization, and merger; Roccas & Brewer, 2002) may moderate the observed relationship between conflicted connectedness and psychological distress. Individuals who compartmentalize their identities may find conflicted connectedness minimally distressing, whereas individuals who merge their identities may be especially prone to psychological fall-out in the face of conflicted connectedness. Given that the current study was not designed to show how inmates managed their multiple connections, we do not yet know whether the observed negative correlates of dual connectedness are due to having dual connections to communities with incompatible values or to certain and perhaps less than adaptive ways of managing or structuring one’s identity.

Clinical and Policy Implications

When study participants were not connected to the community at large, being connected to the criminal community was not significantly associated with important indicators of psychological distress. Do these results suggest that being connected to the criminal community is a desirable state? Probably not. After all, criminal connectedness was consistently associated with endorsement of criminogenic thinking, including antisocial assumptions such as believing that crime does not hurt others. Further, those especially connected to the criminal community showed a suspicious absence of oft-desired character strengths such as gratitude and generosity.

Practically, it is important for practitioners in therapeutic and correctional settings to consider directly the problems associated with dual connectedness. Mounting evidence indicates that social connections play powerful roles in promoting desistance from crime and rehabilitation among ex-offenders. A robust finding in the criminology and sociology literatures is that the relationships of ex-inmates, both during incarceration and after, influence their prospects for recidivism, as well as rehabilitation (for a review, see Hairston, 1991). Although most of the extant literature focuses on familial ties, we expect community connections to likewise be of value. Interventions that strengthen one’s connectedness to the community at large, while failing to also degrade connectedness to the criminal community, might inadvertently precipitate elevated psychological distress – an outcome that could potentially derail individual efforts to desist from crime and contribute fully to the world beyond the confines of jail. On the other hand, interventions that strengthen one’s connectedness to the community at large, while also degrading connectedness to the criminal community, might be especially effective at reducing psychological distress, potentially decreasing the likelihood of recidivism and facilitating post-incarceration re-entry into the community.

Acknowledgments

This study was supported by a grant awarded to June Tangney from the National Institute of Drug Abuse (RO1 DA14694-01).

We are grateful to the members of the Human Emotions Research Lab for assistance with study design and data collection.

Footnotes

1

Data from only 19 women were available at the time of analysis. Upon visual inspection of the data, it was apparent women were outliers on a number of key variables. Ideally, we would have liked to check statistically whether gender moderated any of the observed effects. However, due to the small sample size, we’re not yet in a position to explore this possibility. Thus, the decision was made to restrict the analyses to a homogenous sample of males.

2

The data we report here were obtained during Phase I of a five-phase longitudinal study of criminal recidivism.

3

We also conducted all analyses with the dropped cases and without attenuating the extreme scores. The pattern of results, magnitude of effects, and significance levels were not altered.

4

In addition to the measures discussed in this paper, participants provided data on other constructs that are not relevant to the present study and are thus not discussed here.

5

Thank you to Zac Elston for developing the program used for data collection.

References

  1. Aiken LS, West SG. Multiple regression: Testing and Interpreting interactions. Sage; London: 1991. [Google Scholar]
  2. Aron A, Aron EN. Love and the expansion of self: Understanding attraction and satisfaction. Hemisphere; New York: 1986. [Google Scholar]
  3. Aron A, Aron EN, Smollan D. Inclusion of other in the Self Scale and the structure of interpersonal closeness. Journal of Personality & Social Psychology. 1992;63:596–612. [Google Scholar]
  4. Aron A, Mashek D, Aron EN. Closeness as including other in the self. In: Mashek D, Aron A, editors. The Handbook of Closeness and Intimacy. Erlbaum; Mahwah, NJ: 2004. pp. 27–41. [Google Scholar]
  5. Baumeister RF, Leary MR. The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin. 1995;117(3):497–529. [PubMed] [Google Scholar]
  6. Benet-Martinez V, Haritatos J. Bicultural Identity Integration (BII): Components, dynamics, and psychosocial correlates; Poster presentation at the annual meetings of the Society of Personality and Social Psychology; Austin, TX. 2004, January. [Google Scholar]
  7. Blanchard C, Perreault S, Vallerand RJ. Participation in team sport: A self-expansion perspective. International Journal of Sport Psychoology. 1998;29:289–302. [Google Scholar]
  8. Brewer MB. The many faces of social identity: Implications for political psychology. Political Psychology. 2001;22:115–125. [Google Scholar]
  9. Brodsky AE, O’Campo PJ, Aronson RE. PSOC in community context: Multi-level correlates of a measure of psychological sense of community in low-income, urban neighborhoods. Journal of Community Psychology. 1999;27:659–679. [Google Scholar]
  10. Bureau of Justice Statistics Prisoners in 2002. (NCJ 200248) 2003a Also available at http://www.ojp.usdoj.gov/bjs/prisons.htm.
  11. Cialdini RB, Brown SL, Lewis BP, Luce C, Neuberg SL. Reinterpreting the empathy-altruism relationship: When one into one equals oneness. Journal of Personality and Social Psychology. 1997;73:481–494. [PubMed] [Google Scholar]
  12. Clark LA, Watson D. General affective dispositions in physical and psychological health. In: Snyder CR, Forsyth DR, editors. Handbook of social and clinical psychology: The health perspective. Pergamon; Elmsford, NY: 1991. pp. 27–41. [Google Scholar]
  13. Cohen J, Cohen P. Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences. 2nd ed. Erlbaum; Hillsdale, NJ: 1983. [Google Scholar]
  14. Cortina JM, DeShon RP. Determining relative importance of predictors with the observational design. Journal of Applied Psychology. 1998;83:798–804. [Google Scholar]
  15. Coyle A, Rafalin D. Jewish gay men’s accounts of negotiating cultural, religious, and sexual identity: A qualitative study. Journal of Psychology and Human Sexuality. 2000;12(4):21–48. [Google Scholar]
  16. Deaux K. Reconstructing Social Identity. Personality and Social Psychology Bulletin. 1993;19:4–12. [Google Scholar]
  17. Fazio RH, Towles-Schwen T. The MODE model of attitude-behavior processes. In: Chaiken S, Trope Y, editors. Dual-process theories in social psychology. Guilford Press; New York: 1999. pp. 97–116. [Google Scholar]
  18. Festinger L. A theory of cognitive dissonance. Stanford University; Palo Alto, CA: 1957. [Google Scholar]
  19. Hairston CF. Family ties during imprisonment: Important to whom and for what? Journal of Sociology and Social Welfare. 1991;18:87–104. [Google Scholar]
  20. Heider F. Attitudes and cognitive organization. Journal of Psychology. 1946;21:107–112. doi: 10.1080/00223980.1946.9917275. [DOI] [PubMed] [Google Scholar]
  21. Higgins ET. Self-discrepancy: A theory relating self and affect. In: Baumeister RF, editor. The self in social psychology. Key readings in social psychology. Psychology Press; New York: 1999. pp. 150–181. [Google Scholar]
  22. Hirschi T. Causes of Delinquency. University of California Press; Berkeley, CA: 1969. [Google Scholar]
  23. Hogg MA. Social Identity. In: Leary MR, Tangney JP, editors. Handbook of Self and Identity. Guilford; New York: 2003. pp. 462–479. [Google Scholar]
  24. LaFromboise T, Coleman HLK, Gerton J. Psychological impact of biculturalism: Evidence and theory. Psychological Bulletin. 1993;114:395–412. doi: 10.1037/0033-2909.114.3.395. [DOI] [PubMed] [Google Scholar]
  25. Lee RM, Draper M, Lee S. Social connectedness, dysfunctional interpersonal behaviors, and psychological distress: Testing a mediator model. Journal of Counseling Psychology. 2001;48:310–318. [Google Scholar]
  26. Lee RM, Keough KA, Sexton JD. Social connectedness, social appraisal, and stress in college women and men. Journal of Counseling & Development. 2002;80:355–361. [Google Scholar]
  27. Lee RM, Robbins SB. Understanding social connectedness in college women and men. Journal of Counseling & Development. 2000;78:484–491. [Google Scholar]
  28. Levinger G, Snoek JD. Attraction in relationships: A new look at interpersonal attraction. General Learning Press; Morristown, NJ: 1972. [Google Scholar]
  29. Lewin K. The background of conflict in marriage. In: Lewin G, editor. Resolving social conflicts: Selected papers on group dynamics. Harper; New York: 1948. pp. 84–102. [Google Scholar]
  30. Mashek D, Cannaday L, Tangney J. Validating a pictorial measure of community connectedness. 2005 Unpublished manuscript. [Google Scholar]
  31. Morey LC. The Personality Assessment Inventory Professional Manual. Psychological Assessment Resources; Odessa, FL: 1991. [Google Scholar]
  32. Obst P, Smith SG, Zinkiewicz L. An exploration of sense of community, Part 3: Dimensions and predictors of psychological sense of community in geographical communities. Journal of Community Psychology. 2002;30:119–133. [Google Scholar]
  33. Paulhus DL. Assessing self deception and impression management in self-reports: the Balanced Inventory of Desirable Responding. Department of Psychology, University of British Columbia; Vancouver, BC, Canada: 1988. [Google Scholar]
  34. Peterson C, Seligman MEP. VIA Inventory of Strengths (VIA-IS) University of Pennsylvania; 2001. [Google Scholar]
  35. Pretty GMH, Conroy C, Dugay J, Fowler K, Williams D. Sense of community and its relevance to adolescents of all ages. Journal of Community Psychology. 1996;24:365–379. [Google Scholar]
  36. Roccas S, Brewer M. Social identity complexity. Personality & Social Psychology Review. 2002;6:88–106. [Google Scholar]
  37. Rosenberg M. Society and the adolescent self-image. Princeton University Press; Princeton, NJ: 1965. [Google Scholar]
  38. Saucier G. Mini-markers: A brief version of Goldberg’s unipolar big-five markers. Journal of Personality Assessment. 1994;63:506–516. doi: 10.1207/s15327752jpa6303_8. [DOI] [PubMed] [Google Scholar]
  39. Schultz PW. Inclusion with nature: Understanding the psychology of human-nature interactions. In: Schmuck P, Schultz PW, editors. The psychology of sustainable development. Kluver; New York: 2002. pp. 61–78. [Google Scholar]
  40. Sidanius J, Feshbach S, Levin S, Pratto F. The interface between ethnic and national attachment: Ethnic pluralism or ethnic dominance? Public Opinion Quarterly. 1997;61:102–133. [Google Scholar]
  41. Tabachnick BG, Fidell LS. Using Multivariate Statistics. 3rd ed. Harper Collins; New York: 1996. [Google Scholar]
  42. Tajfel H. Human groups and social categories. Cambridge University Press; Cambridge, England: 1981. [Google Scholar]
  43. Tajfel H, Turner JC. An integrative theory of intergroup conflict. In: Austin WG, Worchel S, editors. The social psychology of intergroup relations. Brooks/Cole; Monterey, CA: 1979. pp. 33–47. [Google Scholar]
  44. Tangney JP, Meyer P, Furukawa E, Cosby B. Criminogenic Beliefs and Attitudes Scale. George Mason University; Fairfax, VA: 2002. [Google Scholar]
  45. Toman JA. Unpublished master’s thesis. Cleveland State University; 1997. Dual identity: Being Catholic and gay. [Google Scholar]
  46. Tropp LR, Wright SC. Ingroup identification as inclusion of ingroup in the self. Personality and Social Psychology Bulletin. 2001;27:585–600. [Google Scholar]
  47. Turner JC, Oakes PJ. Self-categorization theory and social influence. In: Paulus PB, editor. Psychology of group influence. 2nd ed. Erlbaum; Mahwah, NJ: 1989. pp. 233–275. [Google Scholar]
  48. Uchino BN, Cacioppo JT, Kiecolt-Glaser JK. The relationship between social support and physiological processes: A review with emphasis on underlying mechanisms and implications for health. Psychological Bulletin. 1996;119:488–531. doi: 10.1037/0033-2909.119.3.488. [DOI] [PubMed] [Google Scholar]
  49. Uleman JS, Rhee E, Bardoliwalla N, Semin G, Toyama M. The relational self: Closeness to ingroups depends on who they are, culture, and the type of closeness. Asian Journal of Social Psychology. 2000;3:1–17. [Google Scholar]
  50. West T. Unpublished doctoral dissertation. State University of New York, Stony Brook; 2003. The role of lay theories about the malleability of human attributes in helping. [Google Scholar]

RESOURCES