Journal of Health Psychology http://hpq.sagepub.

com/

Just Thinking about It : Social Capital and Suicide Ideation among Homeless Persons
Kevin M. Fitzpatrick, Jessica Irwin, Mark LaGory and Ferris Ritchey J Health Psychol 2007 12: 750 DOI: 10.1177/1359105307080604 The online version of this article can be found at: http://hpq.sagepub.com/content/12/5/750

Published by:
http://www.sagepublications.com

Additional services and information for Journal of Health Psychology can be found at: Email Alerts: http://hpq.sagepub.com/cgi/alerts Subscriptions: http://hpq.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://hpq.sagepub.com/content/12/5/750.refs.html

>> Version of Record - Sep 12, 2007 What is This?

Downloaded from hpq.sagepub.com by RAVI BABU BUNGA on October 29, 2011

Just Thinking about It

Social Capital and Suicide Ideation among Homeless Persons
Journal of Health Psychology Copyright 2007 SAGE Publications Los Angeles, London, New Delhi and Singapore www.sagepublications.com Vol 12(5) 750760 DOI: 10.1177/1359105307080604

KEVIN M. FITZPATRICK
University of Arkansas, USA

Abstract
Suicide ideation is a sensitive indicator of personal well-being. While ideation occurs in roughly 3 percent of the US population annually, in this study rates are 10 times higher. This article explores the role of social capital in mediating negative life circumstances on ideation for a sample of 161 homeless adults in a mid-sized Southern US metropolitan area. Our results imply that social capital does not function the same way for homeless persons as it does for the general population. This finding supports growing evidence that social capitals much touted benefits for personal well-being may not apply to disadvantaged populations.

JESSICA IRWIN, MARK LAGORY, & FERRIS RITCHEY

University of Alabama at Birmingham, USA

COMPETING INTERESTS: ADDRESS.

None declared.

Keywords

Correspondence should be directed to: PhD, Department of Sociology, University of Arkansas, Fayetteville, AR 72701, USA. [email: kfitzpa@uark.edu]
KEVIN M. FITZPATRICK,

homelessness social capital suicide ideation

750

Downloaded from hpq.sagepub.com by RAVI BABU BUNGA on October 29, 2011

FITZPATRICK ET AL.: JUST THINKING ABOUT IT

HOMELESSNESS

is an undesirable life circumstance associated with negative life events and significant distress. Homeless persons, however, generally have limited personal and social assets available to address these challenging events and circumstances. Consequently, depressive symptomatology is prevalent among the homeless, with estimates ranging as high as 80 percent suffering the symptoms of clinical depression (Fitzpatrick & LaGory, 2000; LaGory, Ritchey, & Mullis, 1990; Ritchey, LaGory, Fitzpatrick, & Mullis, 1990). The very high prevalence rates of depression coupled with the depressing circumstances of homelessness itself, suggest the inappropriateness of designating this as mental illness and medicalizing the condition. Some of the symptoms of depression (difficulty sleeping, problems eating, difficulty with others) may actually reflect the physical circumstances of homelessness itself. Indeed the association between homelessness and depression can better be viewed as evidence of the psychological suffering normally associated with the homeless condition. In short, it is not surprising that homeless persons are depressed; it is after all a depressing condition. The depths of this suffering, however, have had only minor attention in the empirical literature on homelessness. In particular, almost no information exists on suicide ideation and its etiology among the homeless, yet such data provide a reliable barometer of the level of felt distress, its potential consequences, as well as the resources necessary to alleviate it (Schutt, Meschede, & Rierdan, 1994). Data over the last decade suggest that about 3 percent of the US population aged 18 to 54 report suicide ideation annually. The few studies that report ideation among the homeless indicate rates that are at least four to five times greater than the general population (Rossi, 1989; Schutt et al., 1994). Suicide ideation represents a heightened state of emotional distress; one that better assesses a homeless individuals degree of maladjustment to their precarious situation. While depression can in fact be a normal response to a particularly dehumanizing experience, suicide ideation is the result of feelings of hopelessness and despondency, an emotional reaction highly indicative of pathology. Most contemporary sociological discussions of suicide begin with Durkheims (1888/1957) seminal observation that the extent of social connections in a community or society affects the probability of suicide. Focusing on social integration and cohesion, Durkheim recognized the important role that such ties played as protection against the general chaos and

disorder of a changing society (Durkheim, 1888/1957). While never using the actual terminology, Durkheim argued that social capital was paramount to a populations health and well-being, and thus in circumstances where social capital was limited or deteriorating, suicide rates would be high. Social capital is a critical asset, a key ingredient in personal well-being. Much contemporary social science research supports Durkheims central assumption that social integration improves wellbeing (Berkman, Glass, Brissette, & Seeman, 2000; Campbell, 1999; Kawachi, Kennedy, Lochner, & Prothrow-Stith, 1997). Putnam (2000), in a review of research on social ties and well-being, concludes that personal happiness is best predicted by the breadth and depth of social connections. Indeed, he claims that the effects of these ties on overall happiness are roughly equivalent to receiving a college education and being married. Little is known, however, about the social capital of homeless persons. To this point no research has looked at the range of forms and sources of social capital available to homeless persons or how they work in mediating the homeless condition. Previous research on the homeless (LaGory, Ritchey, & Fitzpatrick, 1991; Schutt et al., 1994), however, has explored one aspect of social capitalsocial support. This research indicated that homeless persons have social support and ties, but it was unclear whether such support was sufficient to address the extraordinarily negative circumstances of homelessness, improve well-being and lower the overall risk for suicide. The homeless suffer significant deficits in fiscal and human capital (lack of income, wealth, housing, etc.), and their social ties are quantitatively and qualitatively different from the general population. These circumstances lead us to ask several questions: are homeless persons able to capitalize on the limited social assets they have? Are these assets meaningful to people in such a resource-deprived state? Are they sufficient to reduce the immense suffering that can accompany homelessness? The intention of this article is to focus on the relationship between social capital and well-being by examining suicide ideation among a particularly vulnerable population, the homeless. Using a randomized, representative cluster sample of 161 homeless adults living in a large metropolitan area in the Southeastern United States, it specifically addresses several critical questions: (1) What is the extent of suicide ideation among homeless persons, and how does its prevalence compare with the general population?
751

Downloaded from hpq.sagepub.com by RAVI BABU BUNGA on October 29, 2011

JOURNAL OF HEALTH PSYCHOLOGY 12(5)

(2) What are the determinants of suicide ideation for the homeless? (3) To what extent do the homeless have access to and use varied forms of social capital? (4) Does social capital mediate and/or buffer negative life circumstances for homeless persons in the same way it does for the general population by improving well-being outcomes such as suicide ideation?

Social capital and suicide ideation

Research consistently suggests a variety of factors associated with suicide ideation, including: social support; undesirable life events and circumstances; physical health; and depression (Gliatto & Rai, 1999; Hovey, 1999; Mazza & Reynolds, 1998; Reifman & Windle, 1995; Rudd, 1990; Schutt et al., 1994; Vilhjalmsson, Sveinbjarnardottir, & Kristjansdottir, 1998; Whatley & Clopton, 1993). While stressors generally deteriorate an individuals social resources and thus increase distress, the impact of this distress on suicide ideation has been shown to be buffered or moderated by these resources (Schutt et al., 1994). In essence, people with high distress and high levels of social resources have low suicide ideation. Generally, however, the social resources explored have been limited to the individuals perceived social support (Ensel & Lin, 1991). Given the current explosion of interest in the broader concept of social capital, little is known about the impact of social assets in general on suicide ideation, and its mediating and/or buffering role. Although the use of the term social capital is fairly recent (Bourdieu, 1986; Coleman, 1988; Lin, 2001; Putnam, Leonardi, & Nanetti, 1993), the notion that social ties are a personal asset is fundamental to sociology itself (Portes, 1998). With the work of Robert Putnam, however, the term has breathed new life, new questions and even a degree of confusion, into research on the link between social ties and well-being. Putnam defines social capital as social networks and the associated norms of reciprocity and trust that arise from them. He describes two basic forms of social capital, bonding and bridging. Bonding social capital are ties between socially similar individuals or groups; these ties help people get by. Bridging social capital, on the other hand, represents connections between dissimilar individuals and groups. Such ties help people get ahead because they unlock assets in the networks that would have been unavailable to them without the connection.
752

Besides making this distinction between two basic types of social capital, Putnam also outlines a number of measures representing a wide range of forms, including: political, civic and religious participation; workplace connections; informal social connections; volunteering and philanthropy; as well as reciprocity and trust. This scatter shot approach to social capital has been critiqued by a number of scholars (Boggs, 2001; Portes, 1998). In spite of the concepts growing use, it has not been defined adequately, according to Portes (1998). No concrete boundaries exist as to what constitutes social capital. Measurement of social capital is highly variable, leaving the concept open to the critique that it is a catch-all term for a variety of social connections and normative circumstances. This critique highlights the importance of investigating different forms of social capital and their roles in social outcomes. In addition, Portes (1998) bemoans the fact that social capital has become something of an analytical quick fix for the problems affecting society, be it at a community, national or international level, hinting to its ever-growing application within research. The hidden assumption in Putnams approach to social capital seems to be that this asset has the same uniform spending power as other forms of capital (physical and human capital). Hence possessing a unit of social capital is likely to have the same consequences for all persons no matter what their status within the social structure. Lin (2001), however, attempts to qualify this notion and to provide a more structurally and resource-conscious understanding of social capital. For him, social capital focuses on the resources embedded in the individuals social networks. In this perspective the social asset is dependent on the resources contained within it. Not all social asset units are equal, rather they are dependent on the value of the resources embedded within them. This view suggests that social capitals impact on personal well-being will be variable across social positionin essence it reflects the social condition in which the poor generally pay more for less. Thus, it is important to explore the impact of social capital on well-being among groups such as the homeless. Of particular interest within the social capital literature is the relationship between social assets and both physical and mental health (Cattell, 2001; Fitzpatrick, Piko, Wright, & LaGory, 2005; Harpham, Grant, & Rodrigues, 2004; Mitchell & LaGory, 2002; Putnam, 2000; Ziersch, Baum, MacDougall, & Putland, 2005). In Bowling alone (2000), Putnam claims that social integration is the best measure of

Downloaded from hpq.sagepub.com by RAVI BABU BUNGA on October 29, 2011

FITZPATRICK ET AL.: JUST THINKING ABOUT IT

individual and community well-being. Specifically, those with more connections are healthier. Furthermore, these benefits have been found with a multitude of social capital measures: family ties; friendship; social group participation; as well as religious and civic affiliations; and a variety of health outcomes. Regarding mental health, Putnam contends that individuals with more social ties are less likely to experience sadness, loneliness and low self-esteem, all indicators of depressive disorders. Furthermore, according to Putnam, the extent of an individuals social ties is the best predictor of his or her level of happiness. Kawachi and Berkman (2001) investigate the link between social capital and mental health by employing assistance from their earlier work on social networks and find similar results. Moreover, a consistent relationship has been found between social capital and mental health (Lin, 2001; McKenzie, Whitley, & Weich, 2002). While the link between social capital and wellbeing has generated a great deal of research recently, the literature is by no means clear or complete. Some research actually shows negative associations between some forms of social capital and healthrelated outcomes. Specifically, research conducted on low-income populations and communities finds that individual social capital may not always benefit those who possess it (Caughy, OCampo, & Muntaner, 2003; Mitchell & LaGory, 2002). Work done by Mitchell and LaGory (2002), for example, concludes that social capital in the form of bonding capital actually increases an individuals level of mental distress within impoverished communities. Similarly, in a study of childhood behavioral problems, Caughy, OCampo and Muntaner (2003) find that social capital in poorer neighborhoods (as measured by how well parents knew their neighbors) is negatively related to preschool behavioral problems. Specifically related to suicide, Kushner and Sterk (2005) point out that disadvantaged communities that are very homogeneous may not benefit from social capital between its members. The same may also be true of physical health outcomes. A recent study on the impact of social capital on health outcomes found that socioeconomic factors are more important than social capital in determining health differences (Ziersch et al., 2005). These findings lend support to Lins notion that social capital may not impact persons in different positions in the social structure in the same manner. Most notably the research evidence indicates that disadvantaged populations do not always benefit in

meaningful ways from social capital. As Lin (2001) suggests, social capital involves the resources embedded in social networks. Thus, for bonding social capital, if the resource base is shallow and the resources are concentrated or highly segregated, such capital may at best help people get by. At worst, such ties can actually become imbalanced exchanges with deleterious consequences. In such cases, resources either deplete quickly or the social assets become networks of obligation rather than exchange (Thoits, 1995). Thus, for the disadvantaged, bonding capital may have limited benefit for well-being generally. Bridging capital, on the other hand, offers an individual access to resources and opportunities not normally available. It expands the depth and breadth of assets, in so doing it can improve individual well-being. Because the homeless represent a group of highly disadvantaged persons, they are an ideal population to assess the impact of social capital on distress and suicide ideation. Can bonding capital offer meaningful mediation of the devastating circumstances of homelessness? Are the conditions of the homeless so challenging that not even bridging capital can offer assistance in alleviating the psychological devastation of a homeless life? This study represents a critical test of the value of social assets to personal well-being.

Data and methods Sample

The data for this research come from a 2005 survey of homeless persons in a mid-sized, southern Metropolitan Statistical Area (MSA) in the United States. Using procedures from an earlier study and specific criteria developed by the Department of Housing and Urban Development for assessing homelessness, a census of street and sheltered homeless people was conducted prior to the survey. The enumeration included residents of, and visitors to, the 35 shelters and soup kitchens serving homeless persons in the MSA, as well as persons without residence sleeping in public places in a 360-square-block area of the downtown. The count took place over a 24-hour period with several mechanisms in place to reduce mistaking homed as homeless persons. The purpose of the census was twofold: (1) to provide an accurate count of street and shelter-based homeless persons in the MSA; (2) to establish sampling quotas for site, sex and race that served as the basis for the sample design used to select survey respondents.
753

Downloaded from hpq.sagepub.com by RAVI BABU BUNGA on October 29, 2011

JOURNAL OF HEALTH PSYCHOLOGY 12(5)

Less than one month after the census, an intensive interview (questionnaire format) was conducted with 161 homeless persons, a number that represents nearly 12 percent of the estimated street and shelter population on any given day. Interviews averaged one hour with a response rate of 88 percent. In order to encourage participation, respondents were given $10.00 in cash after completing the interview. Interviews took place in designated shelter or service provider locations, and were private with only respondents and interviewers present. The interviewers were trained, experienced professional interviewers with considerable field experience in face-to-face interviewing. The survey population was a random probability sample with quotas for the variables race, sex and geographic site. Quotas were filled by randomly selecting persons from each shelter or public location. The general demographic characteristics of the current sample match earlier studies, suggesting that the present sample is fairly representative of the homeless population in mid-sized metropolitan areas in the Southeastern United States

Measurement
The dependent variable is dichotomous and is based on respondent answers to the following question: Since youve been homeless, have you ever thought about killing yourself?; nearly one-third (31%) of the homeless responded yes to this question. Two categorical socio-demographic control variables, gender (1 = Male) and race (1 = White), are used in the current analysis. In addition, an intervallevel variable measuring the time (in days) since respondents lived in permanent housing is included. Besides these controls, we examine individual and group effects for three sets of variables: life circumstances; health symptomatology; and social capital. These variable sets are chosen not only because of their importance for understanding suicide ideation, but because they routinely have been shown to be significant predictors of other mental health symptomatology among the homeless. Additionally, we are particularly interested in the role social capital plays both independently and as a mediator/buffer in predicting suicide ideation among the homeless. Two life circumstance variables are used in the analysis. The first one is an interval-level variable, life events, which asks homeless respondents whether or not they ever experienced any of 12 major life events (job loss, eviction, time in jail/prison, death of close friend, spouse or child, trouble getting
754

along with people, expulsion from school, physical abuse, sexual abuse, legal problems and major health problems). A second life circumstance variable is an ordinal variable that assesses crime victimization and asks the respondent whether they had been victimized in the last six months (robbed, beaten, raped and/or attacked with a weapon). Three health-related variables are examined as predictors of suicide ideationdiagnosed mental health problem, self-assessed physical health and depressive symptomatology. Diagnosed mental health problem is a dichotomous variable that indicates whether respondents were ever told by a physician that they had a mental illness. Self-assessed physical health is measured by a single question asking respondents how they describe their health at the time of the interview with the ordinal categories 1 = poor, 2 = fair, 3 = good and 4 = excellent. Depressive symptomatology is assessed using the 20-item Center for Epidemiological Studies Depression Scale (CES-D), a reliable instrument used widely to assess self-reported depressive symptoms (Radloff, 1977; Weissman, Sholomskas, Pottenger, Prusoff, & Locke, 1977). The scale contains 20 items that reflect the six major dimensions of depressive symptomatology: mood; feelings of guilt and worthlessness; helplessness and hopelessness; loss of appetite; sleeplessness; and psychomotor retardation. The scale ranges from 060 based on frequency of symptoms in the past week and responses range from 0 = never to 3 = most or all the time. In studies using the CES-D, a score of 16+ is designated as a cutoff for possible clinical caseness, and a score of 21+ as a cutoff for probable clinical caseness. The selfreport scale is reliable with a Cronbachs alpha = .89. The last set of variables examines different aspects of social capital and follows closely the measurement employed by Putnam (2000) in his research. Specifically, we use proxies to assess two distinct elements of the social capital construct bonding and bridging social capital. The bonding form, which promotes homogeneity and group exclusivity, is examined using four social affiliation measurements: religious social capital; group participation social capital; trust; and strength of social ties. Religious social capital is the sum of six sets of responses assessing an individuals level of religious participation. The questions for this variable are as follows: (a) How often do you attend church? (b) Is religion important in your life? (c) Do you turn to someone spiritual when you have problems? (d) Do you turn to someone spiritual when you are

Downloaded from hpq.sagepub.com by RAVI BABU BUNGA on October 29, 2011

FITZPATRICK ET AL.: JUST THINKING ABOUT IT

lonely? (e) Are you a member of a church? (f) Are you involved in church activities? Group participation is created using seven responses concerning: (a) veterans group participation; (b) political group participation; (c) trade group participation; (d) support group participation; (e) homeless group participation; (f) other group participation; and (g) if the respondent voted in the last presidential election. Trust is measured by adding four measures of social trust: (a) trust in others generally; (b) trust in other homeless; (c) trust in community leaders; and (d) trust in homeless service providers. Lastly, the strong ties variable is the sum of three sets of responses pertaining to how often respondents felt bothered by: (a) not having a close companion; (b) not having enough friendships; and (c) not getting to see the people they are close to, over the last six months. The responses to the questions range from 1 = most or all of the time (57 days per week) to 5 = never. The second capital variable assesses bridging social capitalthe kind of social capital that promotes heterogeneity and group inclusivity. The variable is operationalized using a four-item scale that asks respondents whether or not they have close friends who are different from them in terms of their race, educational background, if the person owns their own business and whether or not they are seen as a community leader.

Table 1. Descriptive statistics for socio-demographic, life circumstance, health, social capital and suicide ideation among 161 homeless persons Variables Mean (SD) Percent

Sociodemographics Race (Nonwhite) Gender (Female) Time w/o permanent 365 (947)a housing (in days) Life circumstances Life events 6.3 (2.50) Crime victimization 2.4 (.95) Physical and mental health assessment CES-Depression 23.1 (12.20) Health assessment 2.4 (.86) Told by physician of mental illness (yes) Social capital: Bonding Religious participation 3.5 (4.0) Group participation 1.3 (1.0) Trust 1.1 (1.1) Strong social ties 7.9 (3.5) Bridging Bridging social capital 2.1 (1.9) Suicide ideation Considered suicide since homeless (yes)

63.4 36.0

40.4

31.0

Results
This articles primary interest is in the factors associated with suicide ideation among the homeless. The results show that suicide ideation is prevalent among the homeless. As seen in Table 1, 31 percent of respondents report thoughts of committing suicide since becoming homeless. This figure is approximately 10 times higher than the annual percentage of the general US population reporting such thoughts (Gliatto & Rai, 1999; Kessler, Berglund, Gorges, Nock, & Wang, 2005). The dramatic difference is suggestive of the significant suffering that accompanies the homeless life circumstance. Table 1 includes descriptive statistics on the other factors believed to be associated with suicide ideation. The sample of 161 homeless adults is 64 percent male and 63 percent nonwhite. With regard to life circumstance variables, more than 90 percent of the sample report experiencing two or more significant life events in their lifetime, though the average is considerably higher. One-third of the sample

Standard deviations are in parentheses. Categories may not add up to 100 percent because of either rounding or missing cases a Median days spent without permanent home

report being a victim of violence, with the majority of those reporting having been victimized at least once in the last six months. The health and wellbeing variables indicate a sample of homeless that on average are clinically depressed with over twothirds of the sample meeting the minimum criteria of possible clinical caseness (+16). Interestingly, the majority of homeless see themselves as either in excellent or good health. Forty percent of the sample report being told by a physician that they have a mental illness. The final sets of independent variables assess two different types of social capital. Bonding social capital is assessed using four separate variables: religious participation; group participation; trust; and strong social ties. As suggested earlier, the homeless are not socially isolated and actually are well connected
755

Downloaded from hpq.sagepub.com by RAVI BABU BUNGA on October 29, 2011

JOURNAL OF HEALTH PSYCHOLOGY 12(5)

through a variety of social and organizational mechanisms. Bridging social capital looks at the extent to which friendships and social relationships take place with the non-homeless. Only 10 percent of the sample reported having no such contact and over half of the sample report having more than two bridging friendships. Particularly striking is the fact that 90 percent of those sampled described religion as important or very important in their lives, a response that is high even for a Southeastern United States sample of the general population. The major focus of the analysis is to examine the cross-sectional relationships between selected predictors and suicide ideation and whether the presence of social capital makes any difference in lowering the odds of suicide ideation among homeless adults. Logistic regression is used to estimate the odds of having considered suicide since becoming homeless, taking into account multiple factors contributing to that decision. The analysis provides odds ratios and estimated effects for each of the predictor variables, a measure of statistical significance for the entire model, a summary measure of the models fit (pseudo R2), statistical significance for each of the variables and the significance for each variable set as it is added to the model. As variable sets are added, one can also note whether changes occur in the specific odds ratios for variables entered earlier. This will be particularly important for the social capital variables, as they are generally believed to mediate the distress process. Finally, we test the buffering hypothesis to see whether or not persons with increased distress coupled with resource capital have lower odds of suicide ideation. The logistic regression results are presented in Table 2. By examining the odds ratios, those that are less than one indicate lower odds of suicide ideation compared to odds ratios greater than one, where odds increase ones chances of ideation. Each column in the table represents the addition of a variable or variable set. The first column introduces control variables and the change in R2 between the model with just the constant and the addition of these variables. In the first model, the effects of the control variables are examined; being female significantly decreases ones odds of having thought about suicide since becoming homeless, while race and time spent without permanent housing have no significant effect on these odds. The control variables as a block are significant and when life circumstance variables are added, gender remains significant throughout the remaining models. In the second model, life events and crime victimization significantly increase the odds of a person
756

having thought about committing suicide since becoming homeless. Being a victim nearly doubles the odds of ideation among this sample of homeless adults. Again this block of variables significantly improves the overall model fit by nearly 20 percent. The next step adds the health assessment variables CES-D, self-assessed physical health and whether a respondent was told that they had a mental illness. As expected, CES-D (i.e. increasing symptomatology), significantly increases the odds that a person would have thoughts about committing suicide, however, self-assessed physical health does not. Knowledge that you are suffering from a mental illness actually decreased the odds of having thought about committing suicide; the model fit improves significantly with the addition of these variables by nearly 14 percent. The next set of variables examined in the regression equation are social capital variableseach group added separately to examine both their independent effect and whether or not they made a significant contribution to the overall explained variation of the model. As suggested earlier, bonding social capital while an important aspect of the overall social fabric of the homeless had no significant effect on lowering the odds of suicide ideation and no additional impact on the model fit. The group of variables had no impact on the model or the explained variation in the suicide ideation odds. Bridging social capital, however, has a different effect. Not only does it increase the explained variation in the model by 6 percent, it has a significant negative (lowering) effect on the odds that a person had considered suicide since being homeless. Clearly, social relationships outside the immediate circle of homeless friendships make a difference and can clearly impact ones health and well-being. This model was statistically significant at the .001 level (2 = 49.3, d.f. = 13) with a pseudo R2 = 42%. On further examination it appears as though social capital plays limited/no mediating effect in predicting the odds of suicide ideation among the homeless. No major changes are taking place in the effects of other variables with the addition of these mediators and the bonding variables themselves are not significant. Finally, in an effort to examine the capitalideation relationship more closely, we constructed interaction effects for all of the social capital measures (bonding and bridging) with depressionanticipating that this buffering effect would give us further insight into the heightened effect of capital on ideation. We tested each of the social capital X depression measures for statistical

Downloaded from hpq.sagepub.com by RAVI BABU BUNGA on October 29, 2011

FITZPATRICK ET AL.: JUST THINKING ABOUT IT Table 2. Estimated effects and odds ratios for socio-demographic, life circumstance, health and social capital variables (bonding and bridging) on suicide ideation among homeless persons OR (B) .40* (.90) 1.15 (.14) 1.00 (.01) OR (B) .37* (.99) .80 (.23) 1.00 (.01) 1.28** (.25) 1.76** (.56) OR (B) .37* (.99) .61 (.50) 1.00 (.01) 1.25** (.22) 1.63** (.48) 1.06** (.06) .70 (.35) .41* (.89) OR (B) .36* (1.01) .54 (.61) 1.00 (.01) 1.24** (.21) 1.67** (.51) 1.06** (.06) .71 (.35) .38* (.98) 1.05 (.05) .79 (.24) .98 (.02) .94 (.07) OR (B) .32* (1.14) .60 (.52) 1.00 (.01) 1.35** (.30) 1.88** (.63) 1.07** (.07) .60 (.52) .42* (.87) 1.07 (.07) .86 (.15) .97 (.03) .94 (.06) .58** (.54) OR (B) .27* (1.32) .74 (.30) 1.00 (.01) 1.40** (.34) 1.85** (.62) 1.01 (.01) .56* (.58) .46 (.78) 1.03 (.03) .21* (1.58) .86 (.15) .95 (.05) .56** (.58) 1.05* (.05) 55.71** 14 47.00%a

Independent variables Gender (Male = 1) Race (White = 1) Time w/o permanent housing Life events Criminal victimization CES-Depression Self-assessed physical health Told by physician about mental illness Bonding social capital Religious participation Group participation Trust Strong social ties Bridging social capital Group participation X CES-Depression X 2. df. Nagelkerke R2 (pseudo R2)

5.83* 3 5.80%a

24.25** 5 22.80%a

40.26** 8 35.80%a

42.97 12 37.80%

49.27** 13 42.50%a

*p < .05; **p < .01; ap < .05; hierarchical change in R2 with the addition of variable(s)

significance and only one, social group participation X depression, was significant. That variable was entered in the final model in Table 2. That model was significant with an increase in R2 of almost 5 percent. The interaction effect suggests that higher bonding capital coupled with higher levels of distress have a combinative effect increasing the odds that someone would consider committing suicide during a homeless episode. A few changes occur in this modelfurther indicating that this buffering effect is important and may warrant a more careful investigation than we are unable to do in the current analysis.

Discussion
The results reported support the conceptual and practical value of studying suicide ideation among homeless persons. As demonstrated here, a significant number of homeless persons report having considered suicide during their homeless experience. While suicide ideation is strongly related to depressive symptomatology, there are much greater odds of being depressed than having thoughts of suicide. This suggests that suicide ideation may be a more sensitive measure of the depth of suffering experienced by individuals than more often-used
757

Downloaded from hpq.sagepub.com by RAVI BABU BUNGA on October 29, 2011

JOURNAL OF HEALTH PSYCHOLOGY 12(5)

measures of possible or probable clinical caseness for depression. We would urge more study of suicide ideation among the homeless for that reason alone. In addition, however, there are very practical reasons for further exploration of this variable. Suicide ideation requires detection and intervention. It indicates a crisis situation for the individual that may require a serious institutional response. Greater understanding of the circumstances surrounding suicide ideation will aid in targeting specific individuals for intervention programs. Perhaps of greater significance for the social and behavioral sciences, however, these results help to clarify the link between social capital and wellbeing among the impoverished. Thus, the results serve a more general sociological purpose. Due to the ever-increasing application of the social capital concept, this work points to the importance of examining the intricate and complex nature of social affiliation. Future work, especially work examining disadvantaged populations, can benefit by investigating the many types of social capital available instead of focusing solely on a broad generalized measurement. Clearly, our findings demonstrate that social capital, while important, functions differently for the homeless than the general population. Bonding capital does not reduce the odds of suicide ideation since becoming homeless. That is, social ties in general do not allow the homeless individual to feel better about themselves or their life situation. Only social capital that links the disadvantaged to socially dissimilar individuals provides this benefit. This finding is particularly important to the literature on mental health because it further clarifies the long-standing notion that social support has a generally positive effect on mental health and wellbeing. In fact, the limited test of the buffering hypothesis suggests that some capital coupled with higher levels of depression may actually increase the odds that the homeless would consider committing suicide. In addition, while bridging social capital impacts suicide ideation, negative life events and criminal victimization remain extremely important for the homeless. Bridging social capital, while significant, does not appear to mediate the effects of such undesirable life circumstances. This too represents a difference in the way social capital functions for the general population. These findings, however, do support research that indicates the absence of an inverse relationship between bonding social capital
758

and distress in a population of high poverty ghetto residents (Mitchell & LaGory, 2002). This may point to a broader generalization about social capital among the severely disadvantaged (homeless, high poverty ghetto residents, etc.). Bridging social capital appears to give the disadvantaged individual critical access to resources not available within their own social circle. It does not, however, provide a mediating role, instead acting along with a number of other circumstances and events to affect health and well-being. The policy and practical implications of these findings are not entirely clear. While bonding social capital generally has little or no impact on suicide ideation, support from dissimilar others affects suicide ideation. This fact points to the critical role of formal systems of support and case management in affecting the quality of life of the individual in severe poverty. One effective source of such support in the United States is the local church connected with larger denominations. Such churches usually consist of a significant membership comprised of diverse populations that can promote links between its heterogeneous members. In the Southeastern United States particularly, religion and church ties are often reported to be important or very important in peoples lives (Putnam, 2000), a fact that is obviously true for this sample. Not only do homeless respondents find this source of support important, the United States government has made it increasingly easier for religious groups and agencies to seek funding for supportive services; this is also true for many of the state funding agencies as well. What emerges from our findings in this article is how important we believe it is to develop more systematic approaches to understanding social capital and its potential impact on psychosocial outcomes among homeless and other disadvantaged subpopulations. There are, however, some important limitations that should be noted.

Study limitations
The data used here to examine suicide ideation are cross-sectional; in order to establish causality, longitudinal data are necessary. As we might suspect, some of these relationships may require refinement with regard to respecification (changing the causal sequencing) of relationships. We do not know how many of these persons had suicidal thoughts before being homeless, nor do we know much about their capital before becoming homelessboth factors are clearly important to refining and retesting models in future research.

Downloaded from hpq.sagepub.com by RAVI BABU BUNGA on October 29, 2011

FITZPATRICK ET AL.: JUST THINKING ABOUT IT

In addition, we examined only one sample of homeless persons from the United States, thus making general statements about the homeless population in the United States or any other place in the world requires some care. While we urge the reader to be cautious in drawing conclusions about the results, we believe that the findings suggest an important leading conclusion: social capital may not be what it claims to be, particularly for societies most vulnerable. Homeless persons have extraordinary needs that may require extraordinary forms of supportwe believe those may actually supersede cultural specificity. It is hard to know whether the findings apply to a population of US homeless, let alone those in Europe. A hint at the difficulty of generalizing for policy purposes lies in the significant difference between European and American samples in terms of the importance of religion in everyday life. We also note that all the data collected are based on homeless persons self-report, recorded with a pencil and paper instrument by the interviewer. We acknowledge the potential bias in self-reporting behavior, error potential in recording responses and the influence of confidentiality or the lack thereof in affecting individual responses. Finally, we acknowledge an important theoretical limitation regarding the health psychological orientation of the study. We are interested in the psychological suffering of homeless respondents that is assessed using suicide ideation. That orientation shapes the questions we ask. We are interested primarily in how social capital affects ideation rather than behavior, as such we cannot answer a very important and related public health questionHow does social capital affect suicidal behavior (attempted and completed suicides)? In addition, such a psychological perspective quite naturally focuses on the role of social capital in providing psychosocial support. We do not explore the role that various forms of social capital supplying material resources such as housing and health services might play in alleviating suicide ideation. Homelessness is a structural problem as well as a personal one, focusing only on the psychological aspects of the circumstance provides only a limited focus on the problem and policies to address it as a health issue.

References
Berkman, L. F., Glass, T., Brissette, I., & Seeman, T. E. (2000). From social integration to health: Durkheim in the new millennium. Social Science & Medicine, 51, 843857.

Boggs, C. (2001). Social capital and political fantasy: Robert Putnams Bowling alone. Theory and Society, 30, 281297. Bourdieu, P. (1986). The forms of capital. In J. Richardson (Ed.), Handbook of theory and research for the sociology of education (pp. 241257). Westport, CT: Greenwood Press. Campbell, C. (1999). Social capital and health. London: Health Education Authority. Cattell, V. (2001). Poor people, poor places, and poor health: The mediating role of social networks and social capital. Social Science & Medicine, 52, 15011516. Caughy, M. O., OCampo, P. J., & Muntaner, C. (2003). When being alone might be better: Neighborhood poverty, social capital, and child mental health. Social Science & Medicine, 57, 227237. Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal of Sociology, 94(Supplement), S95S120. Durkheim, E. (1957). Suicide: A study in sociology. Trans. J. A. Spaulding & G. Simpson. London: Routledge & Kegan Paul. (Original work published 1888.) Ensel, W. M., & Lin, N. (1991). The life stress paradigm and psychological distress. Journal of Health and Social Behavior, 32, 321341. Fitzpatrick, K. M., & LaGory, M. E. (2000). Unhealthy places: The ecology of risk in the urban landscape. New York: Routledge. Fitzpatrick, K. M., Piko, B. F., Wright, D. R., & LaGory, M. E. (2005). Depressive symptomatology, exposure to violence, and the role of social capital among African American adolescents. American Journal of Orthopsychiatry, 75, 262274. Gliatto, M. F., & Rai, A. K. (1999). Evaluation and treatment of patients with suicidal ideation. American Family Physician, 59, 15001513. Harpham, T., Grant, E., & Rodrigues, C. (2004). Mental health and social capital in Cali, Columbia. Social Science & Medicine, 58, 22672277. Hovey, J. D. (1999). Moderating influence of social support on suicide ideation in a sample of Mexican immigrants. Psychological Reports, 85, 7879. Kawachi, I., & Berkman, L. F. (2001). Social ties and mental health. Journal of Urban Health, 78, 458467. Kawachi, I., Kennedy, B. P., Lochner, K., & Prothrow-Stith, D. (1997). Social capital, income inequality, and mortality. American Journal of Public Health, 87, 14911498. Kessler, R. C., Berglund, P., Gorges, G., Nock, M., & Wang, P. S. (2005). Trends in suicide ideation, plans, gestures, and attempts in the United States, 19901992 to 20012003. Journal of the American Medical Association, 293, 24872495. Kushner, H. I., & Sterk, C. E. (2005). The limits of social capital: Durkheim, suicide, and social cohesion. American Journal of Public Health, 95, 11391143. LaGory, M. E., Ritchey, F. J., & Fitzpatrick, K. M. (1991). Homelessness and affiliation. Sociological Quarterly, 32, 201218. 759

Downloaded from hpq.sagepub.com by RAVI BABU BUNGA on October 29, 2011

JOURNAL OF HEALTH PSYCHOLOGY 12(5) LaGory, M. E., Ritchey, F. J., & Mullis, J. (1990). Depression among the homeless. Journal of Health and Social Behavior, 31, 87101. Lin, N. (2001). Social capital: A theory of social structure and action. New York: Cambridge University Press. Mazza, J. J., & Reynolds, W. M. (1998). A longitudinal investigation of depression, hopelessness, social support, and major and minor life events and their relation to suicidal ideation in adolescents. Suicide and Life Threatening Behavior, 28, 358374. McKenzie, K., Whitley, R., & Weich, S. (2002). Social capital and mental health. British Journal of Psychiatry, 181, 280283. Mitchell, C. U., & LaGory, M. E. (2002). Social capital and mental distress in an impoverished community. City & Community, 1, 199222. Portes, A. (1998). Social capital: Its origins and applications in modern sociology. Annual Review of Sociology, 24, 124. Putnam, R. D. (2000). Bowling alone: The collapse and revival of American community. New York: Simon & Schuster. Putnam, R. D., Leonardi, R., & Nanetti, R. Y. (1993). Making democracy work: Civic traditions in modern Italy. Princeton, NJ: Princeton University Press. Radloff, L. S. (1977). The CES-D scale: A self-reported depression scale for research in the general population. Applied Psychological Measurement, 1, 385401. Reifman, A., & Windle, M. (1995). Adolescent suicidal behaviors as a function of depression, hopelessness, alcohol use and social support: A longitudinal investigation. American Journal of Community Psychology, 23, 329354. Ritchey, F. J., LaGory, M. E., Fitzpatrick, K. M., & Mullis, J. (1990). A comparison of homelessness, communitywide and selected distressed samples on the CESDepression scale. American Journal of Public Health, 80, 13841386. Rossi, P. H. (1989). Down and out in America: The origins of homelessness. Chicago, IL: University of Chicago Press. Rudd, M. D. (1990). An integrative model of suicide ideation. Suicide and Life Threatening Behavior, 20, 1630. Schutt, R. K., Meschede, T., & Rierdan, J. (1994). Distress, suicidal thoughts, and social support among homeless adults. Journal of Health and Social Behavior, 35, 134142. Thoits, P. A. (1995). Stress, coping, and social support processes: Where are we? What next? Journal of Health and Social Behavior, 35(Special Issue), 5379. Vilhjalmsson, R., Sveinbjarnardottir, E., & Kristjansdottir, G. (1998). Factors associated with suicide ideation in adults. Social Psychiatry and Psychiatric Epidemiology, 106, 97103. Weissman, M. M., Sholomskas, D., Pottenger, M., Prusoff, B., & Locke, B. (1977). Assessing depressive symptoms in five psychiatric populations: A validation study. American Journal of Epidemiology, 106, 203214. Whatley, S. L., & Clopton, J. R. (1993). Social support and suicide ideation in college students. Psychological Reports, 72, 201202. Ziersch, A. M., Baum, F. E., MacDougall, C., & Putland, C. (2005). Neighbourhood life and social capital: The implications for health. Social Science & Medicine, 60, 7186.

Author biographies
KEVIN FITZPATRICK, Professor in Sociology & Jones Chair in Community at the University of Arkansas. His research interests include disadvantaged populations and the mental health and well-being of at-risk adolescents. JESSICA IRWIN is a PhD student at the University of Alabama at Birmingham studying medical sociology. Her interests include mental health, disadvantaged populations, sexuality and gender. MARK LAGORY is Professor and Chair, Department of Sociology, University of Alabama-Birmingham. His current research deals with social capitals role in promoting the quality of life of disadvantaged populations. FERRIS J. RITCHEY is a Professor of Sociology at the University of Alabama at Birmingham. His research specialties include statistics/methodology, medical sociology and disadvantaged populations.

760

Downloaded from hpq.sagepub.com by RAVI BABU BUNGA on October 29, 2011