Community-Level Social Support Responses in a Slow-Motion Technological Disaster: The Case of Libby, Montana

Abstract

Social support is an important resource for communities experiencing disasters. However, a disaster's nature (rapid- versus slow-onset, natural versus technological) may influence community-level responses. Disaster research on social support focuses primarily on rapid-onset natural disasters and, to a lesser extent, rapid-onset technological disasters. Little research has addressed slow-onset disasters. This study explores social support processes in Libby, MT, a community experiencing a “slow-motion technological disaster” due to widespread amphibole asbestos exposure. A comprehensive social support coding system was applied to focus-group and in-depth-interview transcripts. Results reveal that, although the community has a history of normative supportiveness during community and individual crises, that norm has been violated in the asbestos disaster context. Results are interpreted as a failure to achieve an “emergent altruistic community.” Specifically, community-level conflict appears to interfere with previously established social support patterns. The observed phenomenon can be understood as the deterioration of a previously supportive community.

Introduction

Social support is an important resource for communities experiencing disasters. However, the very nature of a disaster may influence community responses. Disaster research on social support focuses primarily on rapid-onset natural disasters. Although the literature has acknowledged the existence of both slow-onset natural (e.g., Sartore et al. 2008; Zamani et al. 2006) and technological disasters (e.g., Couch and Kroll-Smith 1985; Hernandez and Sedler 2003; Substance Abuse and Mental Health Services Administration [SAMHSA] 2004), a paucity of research has investigated social support processes in slow-motion technological disasters. This qualitative study explores social support processes in a community experiencing a “slow-motion technological disaster.” The disaster is ongoing in Libby, Montana, where a 70-year vermiculite mining operation resulted in widespread amphibole asbestos exposures among community members.¹ The Environmental Protection Agency (EPA) has called this situation “the most horrific environmental disaster in US history” (Sullivan 2007). Research in Libby may have implications for other Superfund sites (hazardous waste sites designated by the EPA as high priority for clean-up based on health impact). Today, nearly half of the country's population lives within 10 miles of designated or proposed Superfund sites (Sapien 2007; see Edelstein 2004).

The Nature of Disasters

The literature offers numerous definitions of disaster (e.g., Berren et al. 1989; Fritz 1961). Common elements of definitions include community-level effects and the overwhelming of resources. For example, the World Health Organization (1992) defined a disaster as “a severe disruption, ecological and psychological, which greatly exceeds the coping capacity of the affected community” (p. 2). Although traditionally, definitions of disasters (e.g., US Public Health Service 1976, p. 261) and the disaster literature focused on rapidly occurring natural disasters (e.g., hurricanes, earthquakes), contemporary scholars recognize that not all disasters are acts of nature with immediate consequences.

Two important dimensions distinguish types of disasters: they can be natural or human-induced and they can differ in their timing. Human-induced disasters typically have been called “technological disasters” (Elraz and Osmo 1995; Taylor 1991). Technological disasters are “human made” and involve accidents or other forms of human failure (Baum et al. 1983; Bolin 1993) and thus invoke human culpability (Kaniasty and Norris 2004). Although, following natural disasters, humans may be blamed for incompetent or failed responses, in technological disasters, humans often are culpable for the disaster's very onset.

A second dimension is the disaster's timing. Two related but important distinctions relate to timing. The first is speed of onset: (a) rapid-onset disasters, for which a clear initiating “event” marks the start of the disaster, versus (b) slowly emerging, disasters whose ongoing process is “discovered” (often by victims rather than authorities), often years or even decades into the disaster. The second distinction related to timing is duration: (a) disasters of relatively short duration (e.g., hours, days, weeks) imply a relatively clear end point, versus (b) slowly evolving disasters that persist (e.g., years, decades) with no clear end point in sight. Although these distinctions have been acknowledged in the literature for decades (e.g., Barton 1969), relatively little research has addressed slowly evolving “discovered” disasters (Couch 1996).

Unfortunately, the two distinctions regarding timing often are blurred and sometimes conflated in the literature, particularly with regard to technological disasters. For example, Erikson (1991) referred to the impact of the nuclear reactor accident at Three Mile Island (TMI) as indicative of “a new species of trouble” (p. 12) characterized by human culpability and toxic poisons. He included the Love Canal and Bhopal, India cases (thousands died from chemical exposure from a Union Carbide plant) in the “new species.” Erikson's TMI account captures both the long-term uncertainty experienced by residents (“deep and profound dread,” p. 12) and the “event” to which officials and the public responded immediately (by evacuating to reduce exposure risk). Clearly TMI had powerful and persistent psychosocial consequences (e.g., elevated distress more than 6 years later; Davidson and Baum 1992), but the onset of the disaster was marked by an undisputed “event.” Similarly, Edelstein (2004), in his work on “contaminated communities,” combined rapid-onset disasters (e.g., TMI, Bhopal, the Exxon Valdez oil spill) with slow-onset disasters (e.g., Centralia; Woburn, MA; dioxin contamination in Times Beach, MO). In his later work, Erikson (1994) acknowledged that “toxic disasters” differ, with some having “clearly defined beginnings” that “signal” the disaster (e.g., Bhopal), while others may be ongoing for years before they are noticed (p. 148). The latter are characterized by the absence of a clear beginning and the absence of a clear ending [except perhaps in cases of community relocation, which Edelstein (2004), notes is a rare outcome].

Slow-Motion Technological Disasters

We have chosen the term “slow-motion technological disaster” (Hernandez and Sedler 2003; SAMHSA 2004) to refer to slow-onset disasters of lengthy indeterminate duration. Relatively little research has been conducted on slow-motion technological disasters. Much of what is known comes from research on Love Canal and the Centralia mine fire. Levine (1982) referred to toxic contamination at Love Canal as “a new kind of disaster” (p. 1). Couch and Kroll-Smith (1985), who studied the Centralia mine fire, introduced the term “chronic technical disaster” to refer to disasters attributable to human action and persistent over time. However, both cases resulted in relocating residents. Other cases of slow-motion technological disasters include contaminated ground water resulting in a cluster of cancer cases in Woburn, MA (see Harr 1995; Zaillian 1998) and ground water contamination in Legler, NJ (Edelstein 2004).

Although limited in volume, extant research indicates that this combination of disaster characteristics (slow onset, lengthy indeterminate duration) has significant implications for disaster responses. The ambiguous beginnings and endings to such disasters create seemingly unending uncertainties and associated perceived loss of control, suggesting that their impact on psychosocial adjustment may be relatively more severe. Social support is an important resource for coping with any type of personal crisis or community-level disaster and has been shown to be associated with disaster-related psychosocial adjustment (Norris et al. 2002).

Social Support and Disasters

Social support has been conceptualized in varied ways, including as a communication phenomenon. Hobfoll (1988) defined social support as “social interactions or relationships that provide individuals with actual assistance or that embed individuals within a social system believed to provide love, caring, or sense of attachment to a valued social group or dyad” (p. 121). Albrecht and associates (Albrecht and Adelman 1984, 1987; Albrecht and Goldsmith 2003) defined social support as a communication phenomenon that functions to manage uncertainty and enhance perceived personal control. Social support has been conceptualized as a resource in both the stress (e.g., Hobfoll 1988) and disaster literatures (e.g., Kaniasty and Norris 1995a).

The Complexity of Social Responses to Disasters

Ample evidence indicates that social support plays an ameliorative role in health and coping with stressors (Cobb 1976; Cohen and Syme 1985; Albrecht and Goldsmith 2003), including disasters. Norris et al. (2002) reviewed studies comprising 160 samples of disaster victims and found that social support often serves a protective function in psychosocial responses to disasters. Social embeddedness (size, activeness, and closeness of social network), perceived support, and received support were found to protect disaster victims. (However, beneficial effects of received support have been somewhat mixed and sometimes limited to particular outcomes or types of support.) Evidence finds social support inversely associated with negative psychosocial outcomes (e.g., distress, depression, anxiety) of victims of both rapid-onset natural (e.g., Cook and Bickman 1990; Cowan and Murphy 1985; Kaniasty and Norris 1993; Norris and Kaniasty 1996; Solomon et al. 1993) and technological disasters (e.g., Dalgleish et al. 1996; Dougall et al. 2001; Elklit 1997; Fleming et al. 1982). Although victims of both rapid-onset natural and technological disasters appear to benefit from social support, whether a community provides that social support to victims may vary based on the nature and resulting stressors of the disaster. Not only does social support affect the impact of stressors on disaster victims, a reciprocal relationship may exist: disaster-related stressors may influence subsequent social support processes (see Jerusalem et al. 1995).

Community-Level Social Support Processes in Response to Disaster

Research generally supports the hypothesis that, in the immediate aftermath of rapid-onset natural disasters, communities experience a surge in social support (e.g., Kaniasty and Norris 1995b, 2004; Norris et al. 2001; Rodríguez et al. 2006). Widespread social support also has been observed in communities experiencing slow-onset natural disasters (i.e., repetitive drought; Sartore et al. 2008; van der Meer et al. 2002). This surge in support is understood to be part of a larger community response phenomenon often termed an emergent altruistic community (Barton 1969; Dynes 1970), a post-disaster utopia (Fritz 1961), or a therapeutic community (Barton 1969; see Cuthbertson and Nigg 1987). That phenomenon is characterized by: a shared identity among victims, a sympathetic stance toward those most affected, and a shared understanding of the disaster's effects and victims' needs.

However, two lines of reasoning suggest that a surge in community-level support cannot be assumed during disasters, including slow-motion technological disasters. First, a more nuanced understanding of community responses to disasters suggests that social support processes not only are responses to disasters, but are themselves affected by disasters. Although rapid-onset natural disasters initially mobilize support, ongoing demand overcomes and depletes support resources, resulting in the deterioration of social support and social embeddedness (e.g., Kaniasty and Norris 1993, 2004; Kaniasty et al. 1990; Norris et al. 2005; Norris and Kaniasty 1996; see review Norris et al. 1995). The term “double jeopardy” describes the impact of disasters on availability of social support: “To cope with threats and losses, people need to marshal social support at the time when their social networks are most likely to be disrupted and potentially unable to carry out their supportive roles” (Kaniasty and Norris 1995b, p. 94). The very social structures and processes likely to facilitate coping with disaster are themselves “victims,” undermined or obliterated by the disaster.

Second, community responses to slow-motion technological disasters may differ from community responses to rapid-onset natural disasters. Reports from both Love Canal and Centralia identify a community in conflict (Levine 1982; Kroll-Smith and Couch 1990). Conflict in Centralia was so profound that researchers concluded that “the real disaster is above ground” (Kroll-Smith and Couch 1990). More specifically, Cuthbertson and Nigg (1987) used the term “nontherapeutic community” to refer to community conflict in response to slow-motion technological disasters, reflecting their conclusion that victims did not receive desired social support. These and other findings regarding community conflict in slow- and rapid-onset disasters, led Kaniasty and Norris (2004) to conclude that technological disasters are likely to be characterized by the deterioration of social support.

However, several arguments challenge the assumption that slow-motion technological disasters necessarily result in community conflict that precludes social support. First, the concepts of conflict and social support should not be conflated. People can disagree on important issues and still provide social support to each other in crises. Further, the absence of conflict does not necessarily ensure social support responses. Consistent with this argument, numerous researchers have suggested that a community's social history and cultural values (e.g., Aronoff and Gunter 1992; Kroll-Smith and Couch 1990) may predict the nature of community disaster responses. Second, the literature typically concludes that community conflict characterizes technological disasters. However, some evidence indicates that community consensus and cohesiveness can emerge in slow-motion technological disasters (e.g., Aronoff and Gunter 1992; Edelstein 2004; Freudenberg 1984). Third, although case studies have established a common pattern of community conflict in slow-motion technological disasters, with accompanying but largely passing reports of lack of civility and absence of social support, systematic studies focusing specifically on social support processes need to be conducted to confirm these observations and to clarify the nature and dynamics of communities' social support responses. Thus, the present study addressed the following general research question: What social support processes characterize community-level responses to a slow-motion technological disaster? Our analyses considered both disaster and non-disaster support contexts.

Libby, MT: Epicenter of a Slow-Motion Technological Disaster

Libby, a small town in rural northwest Montana, has a population of about 10,000 within a four mile radius (US Census Bureau 2000). Until 1990, a vermiculite mine and two processing plants were operated locally; from 1963 until its closing in 1990, the mine was operated by W.R. Grace Company. Vermiculite from this mine was contaminated with Libby amphibole asbestos. In addition to the miners and processing plant workers, who were heavily exposed, residents of Libby also were exposed to asbestos by dust brought home on workers' clothes, ambient air, contaminated insulation and building materials, and the common practice of adding vermiculite to garden soil. Additional occupational exposures included providing contracted services to the mine (e.g., vending machine, automotive maintenance) and harvesting and milling timber in the Libby area (Hart et al. 2007; Ward et al. 2006). Due to the latency period for diagnosing asbestos-related disease (ARD) (from 10 to as much as 40 years or longer; Whitehouse 2004; Wright et al. 2002), this disaster continues to unfold, and its magnitude continues to grow.

Amphibole asbestos exposure is associated with health consequences often collectively referred to as “asbestos-related disease” (ARD). Specifically, exposure is associated with increased risk for mesothelioma, lung cancer, asbestosis, and pleural disease, and may be associated with gastrointestinal cancers (Frumkin and Berlin 1988) and autoimmune disease (Noonan et al. 2006; Pfau et al. 2005). The rate of asbestosis mortality in Libby is 40 to 80 times higher, and the rate of lung cancer mortality is 30% higher, than expected when compared to rates for Montana and the US (Agency for Toxic Substances and Disease Registry (ATSDR) 2000, 2002, 2003). The National Center for Health Statistics (2008, June) reported that Lincoln County, MT had the highest age-adjusted mortality rate for asbestosis (262.5 per million population) among all US counties for 1995–2004 (the overall US rate was 6.1).

Even relatively low levels of exposure to Libby amphibole asbestos can lead to lung disease (Horton et al. 2008; Rohs et al. 2008; Wright et al. 2002; Whitehouse et al. 2008). More than 30 cases of mesothelioma due to non-occupational exposure have been diagnosed in the Libby population (Whitehouse et al. 2008). Mesothelioma is an extremely rare (about 10 cases per 1 million individuals in the US general population; US Cancer Statistics Working Group 2007) aggressive and fatal form of cancer nearly always attributable to asbestos exposure (American Thoracic Society 1990; Bourdès et al. 2000; Srebro and Roggli 1994), and in particular to exposure to amphibole asbestos (ATSDR 2001; Boffetta and Trichopoulos 2002).

Libby represents only the epicenter of a disaster that is spreading over time as well as geographically. Until the Libby mine closed, it was the source of approximately 80% of the world's vermiculite (Egilman et al. 1998). Almost 10 billion pounds of vermiculite were mined between 1960 and 1990 and shipped to hundreds of other locations for processing (Schneider and McCumber 2004). Although the public health threat to residents of Libby is documented, the potential threat to communities across the US who hosted the 243 Grace-owned facilities that processed vermiculite mined in Libby is less well known. In addition, vermiculite was shipped to hundreds of additional addresses in North America (see Schneider and McCumber 2004, p. 405, regarding EPA and US Department of Justice (2005) reviews of company records). Further, the EPA estimates that vermiculite insulation is found in up to 35 million US homes and businesses (Schneider 2002).

Methods

We conducted focus groups and in-depth individual interviews in the Libby community as part of a larger study that explores the psychosocial consequences of a slow-motion technological disaster. The present analysis focuses on community-level social support.² The question driving the present analysis is: What social support processes characterize community-level responses to a slow-motion technological disaster?

Participants

The community research partner (Center for Asbestos Related Disease (CARD)) conducted all recruitment and screening activities. Potential participants, who were invited to respond by telephone, were recruited via flyers placed in local venues (e.g., town hall, clinics, churches), community media ads (newspaper, radio, CARD newsletter), and presentations at community organization meetings. Responses to screening questions were used to assign participants to relatively homogeneous groups.

A total of 71 Libby-area residents who had lived in the area for at least the previous 5 years participated in nine focus groups and five in-depth interviews. Groups were constituted based on participants' experience with ARD: (a) people with ARD (four focus groups and two interviewees; two groups and two interviewees reported a mine association—worked, or had a family member who worked, at the mine or processing plants or worked in and around the mine—and two groups who reported no mine association) (n = 31); (b) undiagnosed family members of people with ARD (two groups and three interviewees) (n = 17); and (c) people not diagnosed with ARD and with no ARD diagnoses in family members (three groups) (n = 23).³ Our goal was to capture diverse subjective experiences and viewpoints on the disaster rather than to recruit a representative sample.

The sample consisted of 34 men (47.9%) and 37 women (52.1%) who self-identified as Caucasian (n = 66, 93%), Hispanic (n = 1, 1.4%), Native American (n = 1, 1.4%), or Other (mixed Caucasian/Native America; n = 3, 4.2%). The majority of participants were 40–65 years of age (n = 42, 59.2%); 35.2% (n = 25) were ages 66–85 years; and 5.6% (n = 4) were ages 21–39 years. The majority reported being currently married (n = 45, 64.3%) and having children (n = 66, 93.0%). Eight participants (11.3%) had not completed high school; 23 (32.4%) completed high school; 23 (32.4%) attended some college; and 17 (23.9%) completed a college degree. Of 67 participants who responded to the residence question, more than half had lived in the Libby area more than 30 years (n = 35, 52.2%); 21 (31.4%) had lived in Libby more than 10 and less than 30 years; and 11 (16.4%) had lived in Libby 10 or fewer years.

Procedures

Because of sensitive asbestos-related legal (i.e., lawsuits, pending criminal trial) and political issues (e.g., Superfund status, controversies regarding clean-up activities), substantial efforts were made to protect participants' identities and confidentiality of their responses. Informed consent processes (including administering an oath of confidentiality to focus group participants) were conducted by the community research partner; at no point did the lead investigators have access to participants' last names or contact information. Focus groups were video- and audio-recorded and interviews were audio-recorded. Upon completion of focus groups and interviews, participants were debriefed and received an honorarium ($40 in “Libby Bucks,” currency that could be used at local businesses). Debriefing included providing contact information for the researchers and for CARD's licensed clinical social worker who was available to help with psychosocial issues that may have resulted from the discussion/interview. Upon completion of transcription, transcript validation, and de-identification of transcripts, all video- and audio-recordings were destroyed.

Moderator and Discussion Guide

The moderator for the nine focus groups was a licensed clinical social worker who was not from Montana and was unfamiliar with asbestos-related issues in Libby. Two investigators jointly conducted the in-depth individual interviews. Both focus groups and interviews were conducted by following a discussion guide covering these topics: (a) Introductions and Opening Question (why agreed to participate), (b) Libby, Montana: Hometown Image (personal perceptions, what you would like others to know about Libby), (c) Experiencing the Asbestos Problem (costs and consequences), (d) Social Support and the Asbestos Problem (social support in general and associated with asbestos issues), (e) Diffusion of the Problem and Process of Understanding It (how perceptions changed); (f) Beyond Libby (what others exposed to asbestos-contaminated vermiculite need to know; (g) Summary; and (h) Libby and the Future (hopes and dreams). Focus groups ranged in length from approximately 2.5–3 h in length; individual interviews ranged in length from 1 to 2 h.

Unitization and Selection

De-identified transcripts were unitized into thought units following rules developed by Auld and White (1956). Two research team members unitized a sample of 30 transcript pages and established 97.5% agreement; in turn they completed, or reviewed and corrected, all remaining unitizing of transcripts. The final data consisted of 18,959 thought units. Focus groups accounted for 15,782 units (83.2%), ranging from 1,300 to 2,421 per focus group (M = 1,753.6, median = 1,675); individual interviews accounted for 3,177 units (16.8%), ranging from 316 to 961 per interview (M = 635.4, median = 543).

Due to the large sample, three researchers reviewed transcripts and used a set of selection rules to identify thought units related to social support processes. Two researchers reviewed each transcript; the third participated in resolving disagreements. The resulting sample consisted of 6,504 thought units.

Social Support Coding System

We developed a comprehensive social support coding system that accounted for disaster and non-disaster contexts, varied social support functions (e.g., emotional, instrumental, institutional),⁴ sources of support (self versus others), barriers to support, and support failures. Based on concepts found in the literature and reviews of transcripts, a preliminary coding system and codebook were developed. The codebook provided definitions for social support concepts (e.g., barriers, functions), coding procedures, coding rules, and examples. In the process of training coders, the coding system was fine-tuned and additional rules and examples were added to the codebook. (The selection rules and final codebook are available from the first author.)

The final coding system consisted of six major categories, comprised of 70 mutually exclusive and exhaustive social support topics. Barriers to Asbestos/ARD-Related Support indicated challenges, difficulties, or factors impeding acquiring or providing social support. Asbestos/ARD-Related Support Failures indicated failures to provide, ineffectiveness of, or withdrawal of social support. Non-Asbestos/ARD-Related Support (i.e., non-disaster context) addressed social support needed or provided. Asbestos/ARD-Related Support (i.e., disaster context) addressed social support needed or provided in response to the asbestos disaster. Costs and Benefits of Support indicated the costs or difficulties of providing support or the benefits of providing or receiving support—in both the non-disaster and disaster contexts. Definitions of Social Support defined or clarified the nature of social support or provided definitional examples. (Thought units that could not be coded into one of the six major categories were coded as Other.) See the “Appendix” for a complete list of topics organized within categories.

Coding

Six coders were trained to implement the social support coding system. Practice coding continued until coders achieved an average inter-coder reliability exceeding 86% agreement and average intra-coder reliabilities exceeding 96% agreement on three consecutive samples that comprised approximately 5% of the social support units. Due to the complexity of the coding system and the contextual nuances of the Libby situation, final coding was conducted by groups of three coders. Differences were resolved by consensus. Consensus decisions comprised the final data for analysis.

Analysis

Numerical data regarding participants (screening and demographic information) and coding results were entered into SPSS 15 files. Textual data (thought units) were entered into Microsoft Access files to facilitate identification of themes. Chi-square and Fisher's Exact Tests were computed to assess potential differences in discussion patterns due to ARD Experience.

Results and Discussion

We began our analysis by identifying the extent to which social support was addressed by participants and by identifying the most dominant social support processes by examining the distribution of thought units by categories and topics. The most commonly addressed categories and topics were examined further, both quantitatively and qualitatively, to identify the nature of themes and potential differences in them based on experience with ARD. Results are interpreted in terms of understanding the nature of the Libby community's social support responses to the asbestos-related slow-motion technological disaster.

Proportion of Discussion Devoted to Social Support

Across all focus groups and interviews, 34.3% (N = 18,959) of all thought units addressed social support processes (including asbestos- and non-asbestos-related; and needs for support, observed support, support failures, and barriers to support). Differences in proportions of thought units addressing social support versus other topics occurred based on ARD experience. ARD participants' discussions included proportionately more thought units addressing social support (36.8% of 8,864 units) than did Family participants' discussions (34.1% of 5,729 units), which included proportionately more thought units addressing social support than did Non-ARD participants' discussions (29.4% of 4,366 units) (χ² = 71.4, df = 2, p < .001). Thus, social support was an important topic for participants regardless of their experience with ARD, but discussions of those more directly affected by ARD gave proportionately greater attention to social support processes.

Dominant Social Support Categories

The most common of the seven categories of social support topics discussed with regard to percentages of thought units were Barriers to Asbestos/ARD-Related Support (38.9%), followed by Asbestos/ARD-Related Support (32.8%), and Non-Asbestos/ARD-Related Support (10.7%). See Table 1.

Table 1.

Distributions of percentages and numbers of thought units by seven categories of social support topics and experience with ARD

Category	ARD	Family	Non-ARD	Total
Barriers	39.0 (1,274)	40.1 (785)	36.8 (472)	38.9 (2,531)
Failures	12.3 (400)	6.6 (129)	9.4 (121)	10.0 (650)
Non-asbestos/ARD support	7.4 (240)	14.8 (290)	12.8 (164)	10.7 (694)
Asbestos/ARD support	34.3 (1,121)	30.6 (598)	32.4 (416)	32.8 (2,135)
Costs and benefits	5.0 (163)	4.5 (88)	2.3 (29)	4.3 (280)
Definitions of support	1.6 (53)	2.9 (56)	5.1 (65)	2.7 (174)
Other	0.4 (13)	0.5 (10)	1.3 (17)	0.6 (40)
Totals	100.0 (3,264)	100.0 (1,956)	100.0 (1,284)	100.0 (6,504)

Within the category Asbestos/ARD-Related Support (Topics 39–59) more thought units addressed needs for disaster-related support than observed support. Needs for Asbestos/ARD-Related Social Support (Topics 39–44) accounted for 13.4% of all thought units addressing social support. In contrast, 9.9% of thought units addressed Observed (available or enacted) Asbestos/ARD-Related Social Support (Topics 52–57). The majority of Observed Support was Institutional rather than resulting from naturally occurring social support networks. When Institutional sources are eliminated from both the categories Needs for Support and Observed Support, more than twice as many thought units addressed Needs for Support as addressed Observed Support (12.7% vs. 5.3% of support thought units).

At the same time, Barriers to Asbestos/ARD-Related Support⁵ (Topics 1–11) and Asbestos/ARD-Related Support Failures (Topics 12–18) accounted for nearly half of all thought units (48.9%). Ten percent of thought units addressed Failures of Asbestos/ARD-Related Support (Topics 12–18; 4.9% if Institutional Failures are excluded). In summary, discussion of disaster-related social support needs, barriers, and failures (with Institutional Support included or excluded from consideration) far outweighed discussion of observed support.

Most Commonly Discussed Social Support Topics

To explore social support themes, we first identified the specific social support topics addressed by (a) the greatest numbers of thought units and (b) the greatest numbers of participants. Of 71 topics, the same 11 topics were addressed most in terms of both criteria. Excluding thought units that defined or provided abstract examples of social support (Topic 69), these topics each accounted for 2.5–22.8% of support-related units; each topic was addressed by at least half of the participants (52.1–97.2% of 71 participants). All 11 topics related to the asbestos-disaster context. See Table 2 for the topics and related categories.

Table 2.

Most common of 71 specific social support topics discussed by experience with ARD: percentages and numbers of thought units (N = 6,504)

Category/Topic	ARD	Family	Non-ARD	Total
Barriers
Others' ARD death/illness	8.1 (263)	5.4 (106)	4.2 (54)	6.5 (423)
Physical limits	5.8 (188)	2.4 (46)	0.9 (11)	3.8 (245)
Stigma	17.1 (557)	28.1 (550)	29.4 (377)	22.8 (1,484)
Failures
Informational support	5.0 (162)	2.5 (49)	3.7 (47)	4.0 (258)
Institutional support	6.0 (195)	3.5 (68)	5.3 (68)	5.1 (331)
Needs
General support	3.1 (118)	2.8 (55)	3.7 (48)	3.4 (221)
Emotional support	2.9 (94)	2.3 (45)	2.0 (26)	2.5 (165)
Informational support	4.0 (131)	3.8 (75)	7.9 (102)	4.7 (308)
Available/enacted
Informational support	2.6 (85)	2.7 (52)	2.1 (27)	2.5 (164)
Institutional support	4.1 (134)	4.2 (82)	6.8 (87)	4.7 (303)
Self-help	6.7 (22)	1.2 (23)	2.8 (36)	4.3 (279)

The most commonly discussed topics clustered within the following categories: Barriers to Asbestos/ARD-Related Support, Asbestos/ARD-Related Support Failures (Informational and Institutional), Needs for Asbestos/ARD-Related Support (General, Informational, and Emotional), and Observed (available or enacted) Asbestos/ARD-Related Support (Informational, Institutional, and Self-Help). Thus, discussions established significant disaster-related social support needs, pointed to failures to meet those needs, and identified barriers that may help to explain why social support needs, in general, and among particular groups of people, may not have been met. Analysis of these discussions also hints at the need for alternative support sources (i.e., Institutional, Self-Help) in the absence of adequate support from social support networks. Thus, our analysis suggests that the surge in community-level social support typically observed in the immediate aftermath of natural disasters may have failed to materialize in the Libby community's response to a slow-motion technological disaster. If so, this failure occurred despite evidence of the Libby community's history of normative supportiveness in response to crises.

Libby's History as a Supportive Community

Both qualitative and quantitative evidence support the conclusion that the Libby community had a history of responding with social support in both personal and community crises.

In Their Own Words

When discussing Libby in general (i.e., outside of the asbestos disaster context), participants characterized their community as “caring.” Topic 24, Community Supportive in General, was addressed by about one-third of participants (32.3%). Characterizations of Libby as a supportive community often came in response to one of the first questions: “What would you want others to know about Libby or to have as an image of this town and of its people?” Responses clarified that participants love and are strongly attached to their community. Examples include:

• Everyone helps one another (ARD);

• It's people caring about each other and being hardworking and honest (Family);

• We take care of each other (Family); and

• There's generous, there's good people here (Non-ARD).

Participants also provided numerous examples of the Libby community's supportiveness, evidence that “taking care of each other” was a social norm and that supportive responses were almost routinized in non-disaster contexts. However, the nature of that support often was instrumental and in response to events affecting a relatively small number of people and where the need was visibly evident and short-lived (e.g., in response to floods, fires). For example:

• If you have a flat tire, there'll be three or four people from Libby to stop to see if you're alright, you know, if you need assistance (Non-ARD);

• If there's a house fire, you know there's just people that are Johnny-on-the-spot to take care of it. If there's, you know, somebody dies, there's people that bring food in. I mean it's just uh, I have just seen it. But there are just giving people (Non-ARD);

• There was a woman that had to have a heart transplant. And I'll never forget standin' on the side of the road… and we're watching these logging trucks go down the road with big hearts on the front of them and all these people, all were stopping what they were doing and the whole community was standing there and watchin' these logs that they were donated for this woman (Family);

• And where I noticed that, what you're talking about, was when people would go out and help others, um, protect their property, um, sandbagging and… pumping out basements and doing whatever they could to help each other (Non-ARD); and

• Every time there's a death, or say like, for an example, 3 years ago my [family role]⁶ was killed on a [accident type]… at a [event] and our refrigerator was totally full of baked foods, baked, I mean everything you would ever want. I mean people brought, they just brought, every envelope had money in it for our [role] (Family).

The history of community supportiveness includes spiritual support. Two friends (ARD) described support in their church: “Well, in our church we have what we call the Relief Society and we have home teachers. And the idea of the home teacher is just to be friends with whoever you're to visit” (Participant 21). “You don't condemn. You don't say anything bad. You just visit them and see that everything is all right. That's all you do” (Participant 23). “And be friends with them. That's it” (Participant 21). Another ARD participant described church-based support for a neighbor following a hernia operation: “Well, his church group came in and, and took, basically took care of the family, ya know? They brought them meals and got his wood in and took care of him for the whole winter until he, until he recovered.” A Family participant described church-based support:

Our church also makes prayer shawls…for anybody in the community who would need one. And they are immensely helpful to people, you know. They, they just feel so comforted when they're wearing it. And so when the people put them on and wear them, they know that people are saying, thinking, of them, and have said prayers for them. Yeah, especially if they're terminal ya know.

Quantitative Evidence

To assess potential differences in likelihood of participants addressing specific topics based on their experience with ARD, we used Fisher's Exact Tests to conduct pair-wise comparisons (2 × 2, Experience by Yes/No) due to relatively small samples sizes (i.e., ns of participants; ARD = 31, Family = 17, Non-ARD = 23); exact p values are reported. Although Libby as a Supportive Community in General (Topic 24) (i.e., in non-disaster contexts) was addressed by 23 of 71 participants (32.4%), who spoke to Libby's caring nature, the proportion of participants who discussed the topic varied based on ARD experience. ARD participants were significantly less likely to address Libby as a Supportive Community in General than Family (p_FE = .004) and Non-ARD (p_FE = .011) participants. Only 4 of 31 ARD participants (12.9%) addressed this topic while more than half of Family (n = 9, 52.9%) and 10 of 23 (43.5%) Non-ARD participants addressed this topic. Thus, the more affected participants were by the health consequences of the disaster, the less likely they were to discuss Libby as a caring community. ARD participants' expectations for community support may have been violated by their experience in the asbestos context. As a result, they may no longer tend to view the community as supportive and caring. (Note: Underlining in participant quotations signifies participant's emphasis.) Their comments included:

• It's very isolating. I have very few people I can speak about… this disease with;

• But like I said in the beginning, it's all about the victim, and it ain't seem to be nobody's worried about the victim. They're getting overlooked. And they're the ones having to struggle; the bills and the medicine… [emotional, voice cracks];

• The only thing I can say that… CARD came in… to take care, and to do testing and stuff… That's the only thing; and

• You feel like there's nobody on your side.

We further explored the possibility that ARD experience negatively influenced the perception that Libby is a caring community by reviewing Topic 45, Community in General Provides Asbestos/ARD-Related Support. Only about one-third as many participants discussed Libby's caring nature in the disaster context compared to in general (i.e., non-disaster contexts) (7 versus 23, 9.9% versus 32.4%). And, in fact, only two ARD participants (6.5%), two Family participants (11.8%), and three Non-ARD participants (13.0%) discussed this topic. Thus, few participants discussed Libby as a “caring” community in the disaster context; the likelihood of addressing this topic did not differ based on ARD experience (p_FE's > .26).

Differences in Most Commonly Discussed Topics Based on ARD Experience

To assess potential differences in likelihood of participants addressing the 11 most commonly discussed topics based on their experience with ARD, we computed Fisher's Exact Tests. Results showed no differences, based on ARD experience, in the probability of participants addressing the following topics: Stigma, Institutional Failure, Informational Needs, and Observed Informational Support. However, the proportions of participants who discussed Barriers to Asbestos/ARD-Related Support, and Needs for Emotional and General Support varied significantly as a function of ARD experience.

Relatively more participants with personal experience with ARD discussed Barriers and Needs than those without experience with the disease. ARD (p_FE = .021) and Family (p_FE = .013) participants were more likely to discuss Others' ARD-Related Death/Illness than were Non-ARD participants. In addition, ARD were more likely than Family (p_FE = .015) and Non-ARD (p_FE < .001) participants, and Family were more likely than Non-ARD participants (p_FE = .007) to discuss Physical Limitations on Providing Social Support. ARD were more likely than Non-ARD participants to discuss Informational Failures (p_FE = .049) and Needs for Emotional Support (p_FE = .046), and less likely to discuss Needs for General Support (p_FE = .021).

With regard to Observed Support from others, ARD (p_FE = .01) and Family (p_FE = .026) participants were more likely than Non-ARD participants to discuss Observed Institutional Support. ARD participants also were more likely than Non-ARD participants to discuss Self-Help (p_FE = .03). Together, these two findings further suggest that people with ARD may rely more on institutional support and self-help than on their social networks when compared to people who do not have ARD. One possible explanation for this trend is that that they have support needs that are unmet by their social networks.

Our analyses suggest that people less or not at all directly affected by ARD continue to hold the view of the Libby community as supportive and may have assumed that previously routinized supportiveness has been extended to people with ARD. In contrast, ARD participants likely once held that view but, in light of their current needs and experiences, now are less likely to view Libby as a supportive community. These current experiences include encountering support failures and barriers and having unmet support needs.

Understanding Failed Community Support Responses in Libby

When investigating the “emergent altruistic community,” researchers generally operationalize the construct by observing social support (e.g., Kaniasty and Norris 1995a, b, 2004). Although research has repeatedly found a surge in social support in response to rapid-onset natural disasters, and evidence of the same phenomenon in slow-onset natural disasters, this phenomenon does not appear to characterize the Libby community's disaster response. Our analyses indicate that, although the Libby community's culture is one that includes responding to others' crises with social support, that norm has been violated in the asbestos-disaster context.

Our literature review indicates that researchers have implied that community-level conflict may inhibit a community's social support responses (see, e.g., Cuthbertson and Nigg 1987; Kaniasty and Norris 2004). Conflict typically is understood conceptually as “incompatibility” that occurs when interdependent parties perceive that they have different goals that place them in competition for scarce resources (see Roloff and Soule 2002). Although our analysis focused on social support processes and thus did not code the nuances of conflict as a concept, a sub-topic within one of our Barrier codes (Stigma, Code 6) identified discussion that explicitly acknowledged “us-them” relationships and associated conflict (interpersonal separation, distancing, and overt references to conflict) surrounding asbestos issues in the community.⁷ These data provide evidence that participants perceive Libby to be a community in conflict.

References to Community Conflict

Participants commonly acknowledged community conflict surrounding asbestos-related issues regardless of their experience with ARD. Community-level conflict was addressed explicitly by a total of 194 thought and was discussed in every focus group and interview. More than half (53.5%, n = 38) of all participants addressed community conflict. Conflict was addressed by 45.2% of ARD participants, 64.7% of Family participants, and 56.5% of Non-ARD participants. The probability of discussing community conflict did not differ based on ARD experience (p_FE's ranged from .10 to .22). In addition, ARD, Family, and Non-ARD participants used similar language to refer to or describe community conflict, suggesting shared perceptions. That language points to a conflict characterized by “either-or” and “us-them” rather than a continuum of points of view. See Table 3 for representative examples.

Table 3.

Examples of language reflecting perceived community conflict: ARD, family and non-ARD participants

ARD	Family	Non-ARD
“us and them”	“a real separation between factions”	“us and them”
“sides” of conflict	“sides” of conflict	“two sides”
“polarization”	“conflict between the people”	“divisiveness to the community”
“tension within the community”	“it really did divide the community”	“political tensions”
“community is too split over it”	“split in the community”	“very polarized in the community”
“the town kinda split”	“big break between the people”	“people are split down the middle”
“two sides of the town”	“break here in the town”	“a lot of controversy”
“controversy”	“brought a lot of divisiveness”	“it's contentious”
“conflict”	“conflict”	“conflict”
“there's no middle gray matter”	“it depends what side of the fence”	“a lot of contention”
“sides” of spectrum	“people…have a divided issue on it”	“big controversy”
“two divisions”	“real separation of community”	“very divided”
“community was two-sided”	“different factions”	“town very divided”

The Potential Role of Community Conflict in Inhibiting Social Support

Evidence indicates not only that community-level conflict was consensually perceived as a dominant response to the disaster, but that it has striking implications for social support responses. Two stories capture the reality that the conflict does not merely exist at a cognitive or intellectual level, but also is manifested in interpersonal transactions. In the first, the participant, whose husband had worked at the mine, describes an encounter with a woman she believed to be a friend. The woman's husband was in management at the mine and the participant was attempting to console her over the recent death of her husband but was rebuffed due to assumed conflict:

And so I've had, and I guess I've had some people in town get nasty with me because they said “It's because of you that we're being sued… This one lady, her name was Mrs. [name], and she just really got in my face one day. I was working… And she came in and I just go and put my arm around her, “Well, how are you doin'?” and everything, because her husband has passed away. And she said, “It's you that is suing me! It's you that's wrecking my life!” And I'm going, “What!?” And she said, “You guys have all these lawsuits against us and you're trying to take everything away from us.” You know, and I just, and I don't like that, and I'm just going, “I feel really bad,” but I said, “I don't even know what you're talking about.” I said, you know, “We would never sue you or we would never take anything away from you, Mrs. [name].” And, I still think she feels that way.

A second example of overt interpersonal conflict illustrates that conflicting perceptions regarding the existence of the disaster likely have consequences for motivations to provide social support to those whose health has been affected by the disaster. A woman with ARD describes a confrontation between her son and a stranger, a flagman near the entrance to the mine during clean-up work:

Well, I wanta tell you something that happened to me. I went up 37. That's up toward W.R. Grace. We was going somewhere, my son and my daughter-in-law and I. And that's when they had the flagmen out. And the flagman came up to the car and said it would be a few minutes. And he says, “I don't know what I'm here for. Oh, everybody knows there's nothing wrong with this town.” And my son, I had to grab him. He was going to get out and beat the heck out of this old man. And he says, “Well thanks, mister. [sarcasm] My mother's sittin' here. She's dying, and so is everybody she went to school with. Do you have anything else you want to say?” [sarcasm]

In addition to community-level conflict, ARD and Family participants discussed conflict within families and between friends. Examples include:

• A lot of families are kinda torn a little bit (ARD);

• It caused a big rift… in my own family (Family);

• That's the way my brother and I have been about it [in conflict] (Family);

• It [conflict] even set in the family… and amongst close friends (ARD);

• But I know better than to mention it… because she's [best girlfriend] been very open about her feelings about it (ARD); and

• People that used to be friends have a divided issue on it (Family).

Traditionally, family and friends represent commonly expected sources of social support. Thus, not only has a community with a history of supportiveness not responded supportively in the context of the asbestos disaster in Libby, conflict among family and friends may result in losses of the most commonly expected and accepted social support sources.

Many people may also expect that their religious organizations and fellow church members will be supportive. Notably, the only type of Observed Asbestos/ARD-Related Support discussed by proportionately more Non-ARD than ARD (p_FE = .023) or Family participants (p_FE = .030) was Observed Spiritual Support (Topic 56). More than one-third (34.8%) of Non-ARD participants discussed this topic compared to 9.7% of ARD and 5.9% of Family participants. Instead of church-related support, ARD participants provided poignant examples of conflict among and stigmatization by church members that appeared to preclude adequate social support. Four participants in a Non-Mine ARD group engaged in an extensive discussion indicating that church members “do not visit the people [with ARD]… about it”: “We have a large congregation in [town/city]. And you know what? Um… it is a real issue, even amongst them. You do not go and visit the people… about it. This is an absolute no” (Participant 19). Two other participants explicitly agreed. Participant 19 continued:

That's one of those gray- one of those areas that you don't… Oh. Well, it's a real sensitive subject because there's a lot of the congregation that,… that are not able to come to church any more because they have it so bad. Some of them are dyin' from it. Some of them have it, I mean, they're lucky. There's a couple of them that really doubt they'll make it a year. Uh, there's one individual that's just a couple years older than me that's extremely sick. Um, but, I'm the only one that they can really visit with about it. Because nobody else wants to even… visit about it.

Two group members responded, “They don't want to hear it.” Participant 19 continued, “And you're either their champion, or… they don't wanna talk about it, period.”

Participant 20 followed with his own experience of failure to receive spiritual support from a fellow church member that was rooted in conflict regarding the validity of asbestos-related disease as a serious health problem:

I had, uh, a prominent member of my church… helped me get a load of wood… and, uh, he just started talking about it out of the blue and said what a bogus thing it was and on and on and on. And I'd-I'd, I'd been through enough of it by then… that this was a one-on-one issue. And I stood my ground and said, “No, I'm sorry, man. You're wrong. It is not bogus. I have the stuff and I have it pretty bad. And I know what it's doin' to me. And, you know, I'm a fighter. I exercise thoroughly and, uh, rigorously three or four times a week.” And, anyhow, you know, one of his opening statements was, “Well, have you ever smoked?” And I said, “When I was about eighteen I thought it was cool.” And he said, “Well, there you have it. Emphysema.” And I said, “Wait a minute, [name]. Emphysema does not look anything like asbestosis in a lung x-ray. You know, you need to go sit down and talk with Dr. Black or Dr. White-house, or read some of this stuff that's being printed because they're printing what is the truth,” you know.

Finally, an ARD interview participant recounted an experience that reflected conflicting views among church members and implied a resulting absence of social support for people with ARD: “There's a lady in our church that has a bumper sticker that says, ‘I love Libby and I don't have asbestosis.’ And I thought, ‘How nice, but what about people that do have it?’”

Conclusions

Previous research indicates that communities often respond to rapid-onset natural disasters with an outpouring of social support (e.g., see review by Kaniasty and Norris 2004). The question driving our research has been: What social support processes characterize community-level responses to a slow-motion technological disaster? We conducted qualitative research to identify the most dominant social support patterns and themes characterizing those responses.

Scholars have suggested that a community's history and cultural values may affect its responses to a disaster (e.g., Aronoff and Gunter 1992; Kroll-Smith and Couch 1990). The Libby community's history and cultural values suggest that in the past (and outside of the asbestos context), an outpouring of social support typically occurred in response to crises, particularly when they could be addressed with instrumental support. This norm of supportiveness was evidenced by community members' general perceptions of Libby as a caring and supportive community. Logically then, community members held the expectation that an outpouring of social support should have occurred in response to the asbestos-related disaster and, in particular, in response to people most seriously affected (i.e., those with ARD).

However, despite the Libby community's history and cultural values of supporting each other during personal and community crises, our analyses suggest that the asbestos disaster has had deleterious effects on naturally occurring social support, particularly for people with ARD. ARD participants indicated disappointment in and insufficiency of social support, including non-disaster-related support. Instead of an “emergent altruistic community,” which would be consistent with the community's history and values, they have experienced the deterioration of a previously supportive community. Community-level conflict regarding the asbestos-related disaster and its consequences appears play a prominent role in that deterioration.

Although the Libby disaster clearly has had significant health and economic consequences, our results suggest additional profound negative consequences due to the social environment that evolved in response to the disaster. Consistent with Hobfoll's Conservation of Resource's (COR) Theory (e.g., Hobfoll 1988), Edelstein (2004) argued that the meaning of a disaster “is most clearly indicated by what has been involuntarily lost or changed” (p. 35). Those losses can include familiar and comfortable social processes (e.g., friendliness versus conflict) and social resources that otherwise might facilitate coping. Social support is a particularly powerful resource as it helps to preserve other valued resources (Hobfoll 1988). In Hobfoll's terms (Hobfoll and Lilly 1993), inadequate support resources render victims more vulnerable to additional resource losses. In Libby, one of the disaster's meanings is the loss of previous social support patterns. Worse, we observed that among those whose health has been compromised by the disaster, expectations for social support appear to have been violated. This is consistent with Erikson's (1994) argument: “The mortar bonding human communities together is made up… in moments of crisis, of charity and concern. It is profoundly… disturbing when these expectations are not met… and now they must face the future without those layers of emotional insulation that only a trusted communal surround can provide” (p. 239).

Increasingly, the disaster literature emphasizes that understanding the stressors of disasters is found not just in issues of toxic exposure, but in disasters' symbolic (e.g., meanings, interpretations, attributions) and social processes, whose impact on responses to disasters have been underestimated (e.g., Blocker and Sherkat 1992; Kroll-Smith et al. 2002; Edelstein 2004; Freudenburg 1997; Picou et al. 2004). This may be particularly true for slow-motion technological disasters that have the potential to damage the social fabric of a community. In fact, Edelstein estimates that half of the stress of living in a “contaminated community” is attributable to the disaster's negative social consequences (p. 31) or what Erikson (1994) refers to as “a loss of confidence in the scaffolding of family and community” (p. 242).

Most of what is known about social responses to slow-motion technological disasters comes from Love Canal and Centralia, where abatement from fear came only after residents were relocated (e.g., Couch and Kroll-Smith 1992). In essence, the outcome was that these communities disappeared geographically. In Libby, and in other ongoing and future slow-motion technological disasters, the community likely must continue to live together despite its torn social fabric. Thus, understanding social processes that evolve in slow-motion technological disasters may be critical to effective community and agency responses.

Implications

The very nature of slow-motion technological disasters, combined with community conflict, has implications for community-level support responses. First, slowly evolving technological disasters necessarily mean that the entire community is not affected in the same ways at the same time. Disaster-related social support needs may both evolve slowly (and thus be more difficult to recognize) and affect different community members on different dimensions (e.g., health, economic) and at different points of time. Because some people are seriously affected by disaster-related health issues, while others appear to be unaffected, conflict may emerge about the very existence of a disaster. Community members whose health currently is unaffected may question the cause of health problems and be unaware of the needs of those whose health has been jeopardized. As a result, those in greatest need of social support may not receive it, further jeopardizing their psychosocial health.

Second, to the extent that community conflict is a dominant response, community members may dispute existing health hazards and, as a result, may not engage in appropriate social and health/medical behavioral responses. Community members whose health currently is unaffected and who are skeptical about the health consequences for others, may not only be unaware of the social support needs of others, but may focus instead on their own needs (e.g., disaster-related economic needs). As a result, they may violate community social support norms.

Third, although any disaster may create long-term social support needs in a community, a slow-motion technological disaster has no clear endpoint in sight. When health consequences have a lengthy latency period (as is the case with asbestos-related disease), and when the possibility of continued exposure to a disaster's causes exists, community members' support needs will continue to evolve. Thus, the chronic and ambiguous nature of the disaster creates unique personal and community-level coping challenges.

Fourth, agencies responding to this type of disaster will be challenged to develop strategies to (a) quickly assess a community's social dynamics in order to facilitate and/or re-ignite community-level support as needed, and (b) respond to changing needs of segments of the community, even as other segments appear unaffected by disaster-related health issues. Standard approaches to disasters by federal agencies (e.g., by EPA, ATSDR), tend to focus on resolving technical issues directly associated with toxic exposure and may fail to sufficiently appreciate and address the role of the social environment in traumatic stress experienced by community members.

Limitations and Future Research

Limitations of our present study suggest directions for future research. Although our results suggest that community-level conflict is a prominent factor inhibiting community-level supportiveness in the Libby disaster, our analyses did not focus on conflict. We plan to conduct further analyses that will explore the root causes of that community conflict, specific issues in contention, and issues of community-level consensus so that interventions can be designed and tested to reduce community-level conflict and thereby improve both social support and health/medical responses to the disaster. Aronoff and Gunter (1992, p. 85) suggested that conflict is greatest in disasters when competing social constructions of the disaster stigmatize segments of the population. Stigma was one of the most commonly discussed topics among our participants. Thus, we also plan to analyze the nature, sources, and potential consequences of that stigma in future analyses. Research in the context of HIV/AIDS suggests that stigma often functions in tension with social support processes (e.g., see reviews, Cline and McKenzie 1996, 2000).

Second, we applied a comprehensive complex coding system to a very rich data set in order to capture numerous dimensions and nuances of social support processes in the Libby disaster. Unveiling all of the relevant and critical patterns in these data that may inform future responses to slow-motion technological disasters is beyond the scope of any single analysis. Thus, in future analyses, we will further explore this data set with regard to the barriers to and failures in providing social support and the specific types or functions of social support that may be particularly needed in this type of disaster (e.g., emotional, instrumental, informational).

Third, our results, based on qualitative research and volunteer participants, cannot be generalized to the community of Libby without research to quantify the disaster's impact. Using the results of our focus group research, we have begun population-based survey research in Libby that includes over-sampling of randomly selected patients with ARD from the CARD clinic in Libby. Finally, the Libby disaster is just one example of a slow-motion technological disaster. Numerous such disasters are ongoing in the country (e.g., uranium mining in the southwest, uranium ore production sites, toxic waste sites; see Pasternak 2006; Ohio EPA 2007; Sapien 2007) and are in need of parallel investigation in order to assess the generalizability of the social dynamics and related consequences we observed in Libby.

Appendix

See Table 4.

Table 4.

Social support coding system

Categories and codes

Barriers to asbestos/ARD-related support 1 Others' ARD-related death or illness. 2 Geography. 3 Physical limits on social activities. 4 Emotional limits on providing support. 5 Physical limits potentially limiting ability to provide social support. 6 Stigma. 7 Self-sufficiency, independence values. 8 Anxiety about availability of future support. 9 Lack of awareness of people with ARD or their needs for support. 10 Costs (negative impact) of receiving asbestos/ARD-related social support. 11 Other barriers to asbestos/ARD-related support. Asbestos/ARD-related support failures 12 General failure to provide support [multiple types, type not specified]. 13 Others' failure to provide emotional support. 14 Others' failure to provide informational support. 15 Others' failure to provide instrumental support. 16 Others' failure to provide spiritual support. 17 Institutional failure to provide support. 18 Other failure to provide ARD-related support. Non-asbestos/ARD-related support 19 Need for support: General. 20 Need for support: Emotional. 21 Need for support: Informational. 22 Need for support: Instrumental. 23 Need for support: Spiritual. 24 Community is supportive in general. 25 Speaker provided support: General. 26 Speaker provided support: Emotional. 27 Speaker provided support: Informational. 28 Speaker provided support: Instrumental. 29 Speaker provided support: Spiritual. 30 General willingness/desire to provide support. 31 Support: General [type not specified or multiple types]. 32 Support: Emotional. 33 Support: Informational. 34 Support: Instrumental. 35 Support: Spiritual. 36 Support: Institutional. 37 Self-help: Non-Asbestos/ARD-related, general. 38 Other Non-Asbestos/ARD-related support. Asbestos/ARD-Related Support 39 Need for asbestos/ARD-related support: General. 40 Need for asbestos/ARD-related support: Emotional. 41 Need for asbestos/ARD-related support: Informational. 42 Need for asbestos/ARD-related support: Instrumental. 43 Need for asbestos/ARD-related support: Spiritual. 44 Need for asbestos/ARD-related support: Institutional. 45 Community in general provides asbestos/ARD-related support. 46 Speaker provided asbestos/ARD-related support: General. 47 Speaker provided asbestos/ARD-related support: Emotional. 48 Speaker provided asbestos/ARD-related support: Informational. 49 Speaker provided asbestos/ARD-related support: Instrumental. 50 Speaker provided asbestos/ARD-related support: Spiritual. 51 Willingness/desire to provide asbestos/ARD-related support. 52 Observed asbestos/ARD-related support: General. 53 Observed asbestos/ARD-related support: Emotional. 54 Observed asbestos/ARD-related support: Informational. 55 Observed asbestos/ARD-related support: Instrumental. 56 Observed asbestos/ARD-related support: Spiritual. 57 Observed asbestos/ARD-related support: Institutional. 58 Asbestos/ARD-related self-help by person. 59 Other Asbestos/ARD-related social support. Costs and Benefits of Support 60 Costs/difficulties of providing support in general (hypothetical, general, or non-asbestos/ARD-related). 61 Costs/difficulties of providing asbestos/ARD-related support: Personal (family, friends, advocates). 62 Costs/difficulties of providing asbestos/ARD-related support: Professional (health care workers). 63 Benefits of providing support in general. 64 Benefits of providing asbestos/ARD-related support. 65 Benefits of receiving support in general. [abstract] 66 Benefits of receiving asbestos/ARD-related support. [abstract] 67 Lack of costs of providing or receiving support. 68 Other costs of providing support. Definitions of Support 69 Definitions of support or definitional examples (helpful and unhelpful support). Other 70 Other social support.