the exxon valdez oil spill -...

Industrial Crisis Quanerly 6 ( 1 992 ) 235-257 Elsevier

Disruption and stress in an Alaskan fishing community: initial and continuing impacts of the Exxon Valdez oil spill

J. Steven Picou University of South Alabama, Mobile, A A, USA

Duane A. Gill Mississippi Stare University, Mississippi State, MS, USA

Christopher L. Dyer .University of South Alabama, Mobile, AL. U S

Evans W. Curry Texas Tech University, Lubbock, TX, USA

Picou, J.S., Gill, D.A., Dyer, C.L. and Curry, E., 1992. Disruption and stress in an Alaskan fishing community: initial and continuing impacts of the Exxon Valdez oil spill. Industrial Crisis Quarterly, 6: 235-257.

The study of technological disasters has been characterized by conceptual ambiguity and non-ecoIogica1 considerations of community vulnerability. This research employs an ecological-symbolic theoretical approach that identifies natural resource communities as particularly vulnerable to disasters that contaminate biophysical resources. A longitudinal panel study, including a control community, provides data on disruption and stress experienced by residents of a small fishing community in Prince William Sound, Alaska. The analysis reveals a continuing pattern of stress and disruptions some 18 months

Correspondence to: J.S. Picou, Department of Sociology & Anthropology, University of South Alahma, Mobile, AL 36688, USA.

092 1-8 106/92/$05.00 O 1992 Elsevier Science Publishers B.V. A11 rights resewed.

J. S. Picou ef nl. / Disrupf ion and srress

following the spill. The report concludes with a discussion of the specific long- term patterns observed.

Introduction

Disasters are situations which are socially defined in the context of human communities and their physical environment. An occasion is typically defined as a disaster if the social system's ability to reasonably ensure biological sur- vival, social order, social meanings and social interaction are disrupted (Fritz, 196 1 ; Barton, 1969; Dynes, 1970; Quaranteli and Dynes, 1970; Kreps, 1985, 1 989 ) . Disasters possess a variety of characteristics, i.e., source, speed of onset, scope of impact, duration of impact, etc., which identify unique stmctural consequences for the disruption of a wide variety of community activities (Dynes, 1 970; Mileti, et al., 1975; Drabek and Key, 1976; Erickson, 1976; Drabek, 1986; Gill and Picou, 199 1 ) .

Over the years, increased attention has been accorded to "man-made" or "technological disasters" in terms of their characteristics and consequences for human communities (Turner, 1978; Brown a d Harris, 1979; Baum, et al., 1 98 2, 1 98 3; Omohundro, 1 982; Couch and Kr011-Smith, 1 985; Shrivastava, I 98 7; Edelstein, 1 98 8; Bogard, 1989; Kroll-Smith and Couch, 1 989,1990; Gill and Picou, 1989, 1991 ). Recently a number of researchers have suggested that a variety of long-term mental hcaIth and community disruption problems characterize victims of technological disasters (Gill and Picou, 199 1 ; Freudenburg and Jones, 199 1 ; Kroll-Smith and Couch, 199 1 b ) . Erickson ( 199 1 ) has characterized this pattern of extended social impacts as "A New Species of Trou- ble." Toxic poisoning of t he environment "contaminates rather than merely damage," "pollute, befoul, taint, rather thau just create wreckage," and impact people and communities in "new and special ways" (Erikson, 199 1, pp. 14- 15).

Given these recent concerns in the disaster research literature, the primary objective of this study is an analysis of initial ( 5 months) and continuing ( 18 months) patterns of community disruption and stress resulting from the Val- dez oil spill. The research design includes data collected from both "impact" and "control" communities. A comparative analysis provides the basis for describing the social impacts of the spill in 1989 and 1990. Further, patterns of disruption and stress are examined for subpopulations within the impact community.

Industrial Crisis Quartdy, VoL 6, No. 3

J. S. Picou er a/. / Disruption and sfress

Technological disasters: conceptual approaches and empirical findings

The study of technological disasters has recently been characterized by debate and discussion in the sociological literature. The primary issues in this debate relate to how disasters are conceptualized ( Kroll-Smith and Couch, 199 1 a ) and the documentation of long-term negative impacts of technological disasters on human communities (Freudenburg and Jones, 1991 ). We will briefly review these issues and identify the conceptual framework for this investigation of the community impacts of the Valdez oil spill.

Disasters have traditionally been studied in the sociological literature from perspectives which emphasize the common, generic characteristics of all types of disasters (Quarantelli, 1985) or characteristics of the disaster event which are unique for understanding subsequent community consequences (Baum et ai., 1 98 3; Couch and Kroll-Smith, 1 985; Edelstein, 1 988,; Gill and Picou, 1 989, 199 1 ). Both of these approaches are limited in terms of a direct consideration of community vulnerability (Gillespie, 1 98 8 ) , the emergence of symbolic interpretations of the event by victims and an ecologically-informed view of human response and community change (Kroll-Smith and Couch, 199 la, p. 360; Dyer et al., 1992 ) . An. alternative conceptual model has bee4 identified as the " e c o ~ o ~ c d c s ; l ~ y ~ b o ~ ~ ~ approacht and rtsb on the faowing basic niorns. "( 1 ) people exist in exchange relationships with their built, modified and biophysical environments" and "(2) disruptions in t4e ordered relationship between individuals, groups and communities, and their built, modified and natural environments are labeled and responded to as hazards and disasters" (KroII-Smith and Couch, 1991a, p. 361 ). In sum, the ecological-symbolic a p proach to the study of disasters emphasizes the relationship between human communities and their physical environment and culturally-based, subjective responses to disasters.

The ecological-symbol ic perspective provides a conceptual approach to the Valdez oil spill, which takes into consideration the cultural consequences of the contamination of Prince William Sound and the subsequent devastating impacts 011 waterfowl, marine mammals and many o~ganisms in the \d ecosystem (Nichols, 1989). Furthermore, this perspective addresses the issue of the nature af community vulnerability to tec)lnolo~caJ disasters gi ven that human perceptions mediate interpretations of resource contamination. It is apparent that mbin types 01 communififf may be more vu\nefabIe to bisnpf1~on from resource contamination than others, given the fact that "sustenance mainte- nance" may be resouroe-based (Hawley, 1950; Gibbs and Martin, 1959; Fris- bie and Poston, 1975; Murdock, 1979).

Along these lines, Dyer et al. ( 1992) have identified cultural patterns di- reedy linked to renewable natural resources as the defining cbaracteristic of

Industrid Crisis Qwvter{y, Yo(. 6, No. 3

J.S. Picou et al. / Disruprian and stress

"natural resource communities." The majority of Alaskan communities have cultural traditions of subsistence and commercia1 harvests of fish, berries, and mammals (e.g., see: Stanek, 1985; Smythe. 1988; Statton, 1989; Fall, 1990). Natural resource communities are characterized by traditions of resource utilization which define community values and structure social participation in seasonal subsistence harvest activities. Furthermore, a cash-economy struc- tures commercial harvests and a wide variety of occupations related to resource utilization emerge to support this primary economic activity (Saiaman, 197 1 ). Several studies indicate that fear and uncertainty regarding contamination from the VaIdez spill substantially reduced subsistence harvests of Alas- kan Natives in Prince William Sound (Impact Assessment, Inc., 1 990; Rest* ration Planning Work Group, 1990; Dyer et d., 1992). We suggest that the vulnerability of natural resource communities to resource contamination can- not be understood apart from the complex patterns of interpretations that people "construct" to make sense of the extent and nature of biophysical contamination.

The contamination of a community's water supply, soil or air results in uncertainty, fear, outrage, and disruption among residents ( Ahearn and Cohen, 1 98 1; Picou, 1984; Smith et al., 1986; Kroll-Smith and Couch, 1989; Eriksun, 1 99 1 ; Gill and Picuu, 199 1 ) . Research on the psychosocial impacts of technological disasters has consistently revealed both initial and long-term negative consequences, For example, studies oFThree Mile Island by Houts et al. 1 1988 ) , Dew e t al. ( 1 987 ) and Davidson and Baum ( 1 986 ) all report the existence of various types of long-term ( 17 months to 6 years) psychologica1 stress for a variety of community members. Furthermore, Gill and Picou ( 199 1 ) docu- ment the existence of long-term (20 months) uncertainty, stress and fear among residents of a rural community which experienced a train derailment and toxic spill. In general, these patterns of long-term disruption and stress rarely character-

ize natural disasters (Quarantelli, 1 98 5 ; Freudenburg and Jones, 1 99 1 ) .' As Kroll-Smith and Couch state:

From an ecological perspective, natural disasters rarely result in long-term disruptions of the reIationship between a community and its biophysical environments, while a community's built and modified environments frequently sustain loss. On the other hand, technological disasters routinely disrupt the exchange between human settlements and their natural or biophysical environments, while in many cases resulting in relatively little or no damage to built and modified environments. [boll-Smith and Couch, 199 la, p. 3623

'For a notable exception to this trend, see Steinglass and Gemty ( 1990). The debate of the existence or non-existence of long-term psychosocial impacts of natural disasters may be related more to issues of research design, e.g, rhc luck of and the need for longitudinal research designs in disaster reserach (Solomon, 1989; Maser and Solomon, 1990),

Industrial Crisis Quarterly, Vol. 6, No. 3

J.S. Picou el a/. /Disruption and stress

This pattern of community-environment disruption results in three distinct problems for communities experiencing technological accidents: " ( 1 ) the ambiguity of biophysical damage; (2) the likelihood of a corrosive, rather ihan therapeutic, 'pus!-disaster ' phase and ( 3 ) the overall tendencies toward socio- cultural disruption" (Freudenburg and Jones, 1 99 1, p. 1 1 56; emphasis origi- nal ). These problems facilitate the development of long-term collective stress within the context of continuing uncertainty and disruption (Gill and Picou, 1991),

In summary, technological disasters, like all disasters, initially generate patterns of disruption and stress in communities. However, given that the basic relationship between the impacted community and its biophysical context is disrupted, perceptions of uncertainty, ambiguity and continuing disruption generate patterns of long-term community stress (Freudenburg and Jones, 199 1 ). This pattern has been empirically documented by longitudinal studies and appears to have identifiable characteristics ( Gleser et al., 1 98 1 ; Houts et al., 1 988; Freudenburg and Jones, 1 99 1 ; Gill and Picou, 199 1 ) . Following an initial impact of intense stress, over time, some reduction in overall collective stress typically occurs. However, certain subpopulations of affected communities experience elevated levels of stress over time (Gill and Picou, 1 99 1 ; Dyer et al., 1992). These subpopulations within communities may be defined in terms of relationships to the disaster, e.g., proximity to the contamination, or a unique relationship to the resources that have been contaminated, i.e., sub sistence and/or commercial harvests, property ownership, etc. ( Omohundro, 1982; Davidson and Baum, 1986; Houts et al., 1988; Gill and Picou, 199 1; Dyer et al., 1 992 ) . The existence of such a general pattern of initial and continuing impacts will be evaluated for a natural resource community which was directly impacted by the Valdez spill.

The Exmn Vd& oil spill

On 24 March 1989, at 12:04 AM the supertanker Exxon Valdez ran aground on Bl@ Reef, resulting in the largest oil spill in United States history. Within five hours of the accident, over ten million gallons of Prudoe Bay (North SIope) crude escaped into the pristine waters of Prince William Sound. The response to the spill was inadequate, revealing little preparedness for an oil spill of such magnitude in such a remote location. Furthermore, the initiation of dean-up activities was delayed by confusion in the administrative directives of the various contingency plans (National Response Team, 1 989 ) .' *For more detailed descriptions of specific events which characterized the Valdez oil spill, see: Davidson, 1 990 and Keeble, 1 99 1 . Industrial Crisis Quarrerly, Vol. 6, Nu, 3

J.S. Picou el al. /Disruption and stress

The immediate impact of the spill on the local ecosystem was devastating. Fig. 1 identifies the location of the Valdez spill and the community of Cordova. The environmental conditions characterizing this spill actually increased the severity of the environmental impact (National Response Team, 1 989). For example, tbe type of oil spilled and the low water temperatures resulted in a much slower rate of biodegradation, physical weathering and evaporation of the 02. The spill resulted in an oil slick which measured over 3,000 square miles and impacted over 1,200 miles of shoreline (National Response Team, 1989; Stewart et al., 199 1 ). Furthermore, the spill occurred approximately 60 days prior to the most biologically active time-period in the immediate area. Migration and reproduction cycles of birds, fish, mammals and marine inver- tebrates occurred within two to three months following the spill (Stewart et al., 1991). Six months following the spill the death toll for birds and marine mammals

in the Prince William Sound area was substantial. Conservative estimates had over 33,000 birds, 980 sea otters, 30 harbor seals, 1 7 gray whales and 14 sea lions documented in the death toll (NichoIs, 1989). Recent observations suggest that these initial evaluations may have severely underestimated the biolog-

Fig. 1. Map of the study site.

240 Industrial Crisis Quarterly, Vol. 6, No. 3

J.S. P i m et a!. A Disruption attd stress

ical impacts of the spin. For example, a summary of environmental impacts recently lodged in United States District Court (Alaska) reports that 3,500 to 5,500 sea otters were killed directly by the spill and over 350,QM) birds per- isbed (Stewad et al., 199 1 ; also see: Shew onit, 1 99 1 ).

The consequences of the spill for microbiotic life in intertidai mnes and for the various fisheries in Prince WilIiam Sound remaids uncertain at this time. The 1989 herring season was closed following the spiU Lung-term impacts on the commercial Amon fishery in Prince William Sound will not be kxfaw~ until fry release in f 989 begin to return (National Response Team, 1 989 ). Therefore, local salmon fishermen may not experience the long-term environmental consequences of the spill until the 1 992, 1 993, and 1 994 fishing seasons.

The research setting

Cordova is a small, picturesque fuhing mmmunity located near Prince Wil- liam Sound in south central Alaska. Curdova is gsogtaphitally isolated from other communities by mountains, glaciers, rivers and the sea. No roads have connected Cgrdova to otha communities since the 1964 Alaskan earthquake. A maritime climate of heavy precipitation and moderate temperatures char- acterizes this region.

Histoically, the mmmunity of Cordova can be traced to four Eyak Villages and the territories ofthe Chugach Eskimos. Early documents identify the 1898 Alaska gold rush as a reason for growth in Cordova's population. The city was officially incurparated in 1909 and quickly W e the export center of mppr being mined in the Wmggell mountains north of Cordovel (Stratton, 1989). Following the closing of these mining operations and the railroad in 1939, Cur- dova residents became involved in an expanding commercial salmon fishery. Cordova's population during the next three decades remained mund 1,000 residents until the 1970s when the population almost doubled. This dmdt of growth stemmed from the diversifmtion of the commercial fishing industry in the area and increased in-migration trmds for Alaska in general.

The economy of Cordova is dombted by commercial fishiag. Cordova fishermen hdd 44% of all Prince William Sound herring fishery permits and 55% of aU Pdnce William Sound salmon fishery permits (Stratton, 1989). Subsist- ence activities characterize most of the rmidentq of Cordova. Harvesting, re- ceiving and giving away fish, moose, deer, berries, etc. are common to the majority of Cordova residents. Resource sharing provides the basis for maintaining social relationships. Over 90%~ of a recent sample of Cordova households reported t4at they had rgceived subsistence resources and over 400 pounds of resouroes were bested per household (Stratton, 1 989 1. These and other data

Indutrial Cririr Quarter&, Void 6, No. 3

J.S. Picou er a/. /Disruption and srress

provide evidence for classifying Cordova as a natural resource community (Dyer et al., 1992).

At the time of the Valdez oil spill Cordova could be described as an isolated community, highly dependent on commercial fishing for an economic base and having a cultural history of subsistence practices stemming directly from an Alaskan Native heritage. Approximately 18% of the residents are Alaskan Na- tives. Due to the community's location (see Fig. 1 ), no oil from the E m n Valdez reached the shores or immediate vicinity of Cordova. However, the spill directly impacted primary fishing grounds of Prince William Sound used by local fishermen.

Research design

Cordova was selected as the impact community because of its economic de- pendency on commercial fishing and its cultural heritage of subsistence activities. The overall research design is presented in Fig. 2. Data were collected from: ( 1 ) a stratified, random sample of households in Cordova; (2) a random telephone survey of residents in Petersburg (see below); and ( 3 ) a random telephone survey of Cordova residents.

This research design was developed from the social impact assessment literature, as well as recent applications of impact assessment methods to techno- iogical disasters (Lounsbury et a]., t 983; Livessay et al., 1984; Picou, 1984; GiH, 1 98 6; GilI and Picou, 1 99 I ) . Minimally such assessments require: ( I ) proper sampIing procedures; (2 ) appropriate control community comparisons; and ( 3 ) standardized indicators used in previous disaster research (Pi- cou, 1 984; Gill, 1 986; Solomon, 1 989 ).

Household sample select ion A research team of three social scientists arrived in Cordova on 1 9 August 1 9 8 9. Throughout the first day, the researchers identified seven residential areas in the community of approximately 2, I50 residents. Once these seven residential areas were identified, households were assigned numbers for random selection. Tables of random numbers were utilized for the selection of households. A geo- graphically stratified, random household sample was obtained. Data were collected in approximately 70 households, resulting in a final sample of 86 residents. HousehoId members present in selected locations were interviewed by members of the research team. The majority of the respondents provided information to the researchers which would allow recontacts and future interviews.

Indusfrial Crisis Quarrerly, Vul. 6, No. 3

spat ,

rl FI - -.-

J.S. Picw PI a/. / Dlsruprion and rprress

u Fig. 2. Research design for disaster impact assessment of Valdez oil spill.

IndusrriaI Crisis QuarterIy, Vul. 6, No. 3

J.S. Pimu ei al. / Disrupliun and stress

Cunlrol comrnunitjb selection After an evaluation of demographic and economic characteristics of Alaskan communities from census and cultural information, the city of Petersburg was selected as a control community. Like Cordova, Petersburg is isolated by not having roads l i ~ked directly to it and is highly dependent economical~y on commercial fishing. Petersburg is Iocated in southeast Alaska and has a population of 3,13 7 people. Petersburg also has an Alaskan Native population which comprises approximately 20% of the community (Smythe, 1988). In 1985, a $29.5 million salmon fishery existed in Petersburg, compared to a $36 million salmon fishery in Cordova ( Smythe, 1 9 88; Stratton, 1 989). Petersburg residents share subsistence harvests in a manner similar to Cordova residents. For example, in terms of salmon subsistence activities, 72% of Cordova residents harvest salmon, while 63% of Petersburg residents engaged in salmon harvesting. In Cardova, 58%~ of the residents receive salmon compared to 6 1 % in Pe- tersburg. Percentages of residents who give away salmon are also similar ( 6 1 % Cordova; 5 5% Petersburg ) . These data suggest that Petessburg, like Cordova, can be classified as a natural resource community. In addition to the demographic and cultural information discussed above,

research suggested that Pctetsburg received minimal direct impacts from the Valdez spill. The fishing seasons were not directly affected for most fishermen in Petersburg, although some limited leasing of boats by v ~ c o for the clean-up

' did occur, It should be noted that it is highly probable that some Petersburg residents were impacted negatively by the spill. All Alaskan communities seem to be sensitive to environmental stress and related political issues involving the oil industry and state government. Because of this fact, the use of Petersburg provides a conservative estimate of the magnitude of observed impacts.

The data for the control community were collected by telephone interviews conducted by the Survey Research Unit of the Social Science Research Center, Mississippi State University, In addition, telephone interviews were conducted in Cordova during the same timea3 Households were randomly selected from a list of all possible telephone numbers in both communities (people interviewed in the Cordova household survey were excluded). Once a household was reached by the telephone interviewer, they asked to talk to: ( 1 ) the oldest male; ( 2 ) youngest male; (3) the oldest female; or (4 ) youngest female (over 17 years of age). This procedure reflects a modified version of the Throldahl- Carter approach for random household sampling for telephone interviews (for more information, see: Dillman, 1978 and Frey, 1983). The interviews were conducted in early to mid-December 1989.

3This procedure resulted in data which allowed for an evaluation of the impacts of "data collection techniques'" for social irnpacr assessments. For example, the data collection technique was found nor to predict srress levels (for more information, see Picou et a]., 1990).

Industrial Crisis Quarlerly, Poi, 6, No. 3

J.S. Picotd et a/. I Disrupfion and srr~ss

Table 1

Demographic characteristics for control (Petersburg ) and impact (Cordova ) commun hy samples

Demographic characteristics Cordova Petersburg ( N = 118) (N=73)

Percent married 69% 64% Percent male 48% 44% Percent white 94% 88% Percent in comm~rcial fishing 34% 28%

i Education level 13.42 13.44 ii No. of people in household 2.84 2.76 X Age of respondents 40.23 42.03 R Years lived in Alaska 21.10 24.60 R Years lived in community 18.10 2 I .80

A >

R-mean scares.

The data collection procedures in 1989 resulted in a final sample size of 1 1 8 respondents in the community of Cordova (impact community) and 73 respondents in the Petersburg community (control community ). Table 1 provides comparative demographic data for the samples drawn in both communities. In general, respondents in both samples have similar characteristics, with Petersburg respondents reporting slightly longer periods of community and state residency.

Follow-up surveys in 1990 were conducted by mail sumeys and telephone interviews originating from the Survey Research Unit of the Social Science Re- search Center, Mississippi State University. Data were collected on perceptions of disruption, stress and general demographic information. A total of 69 respondents were successfully recontacted and interviewed in the fail (Octo- ber-November ) of 1 990 in Cordova, reflecting a 62% response rate. Approxi- mately 74% of the 1989 Petersburg sample was reinterviewed, resulting in a 1990 sample of 53 respondents.

Indicators and measures Questionnaires containing similar indicators were used for data collection in both communities in 1989 and 1 990. Questionnaire items were selected because of their use in previous disaster research and their relevance ta t he actual disaster. Five indicators of perceived social disruption were utilized. The specific items were operationalized in terms of changes in the way famiIy members get along, changes in future plans, having relatives make changes in their future plans, changes in the work setting and changes in the community which re-

Ittdustrial Crisis Quurterly, Vol. 6, No. 3

J.S. Picou et a/. / Disruption and stress

Table 2

Indicators of ,social disruption for control and impact communities, 1989 and 1990

As a result of the spill., . (Control ) I ~ w c t 1 %Yes %No %Yes %No

I . Have you wt id any changes in the way your family gets along together?

2. Have you made any changes in your glans for the future?

3. Have other family members changed their future plans?

4. Have things changed for you at work?

5. Have things changed in your community?

*Chi-squart analysis (see Yaemans, 1 963).

sulted from the spill. These items provide a general, straightforward indicator of social dismption for community groups, i.e., families and work groups.4

Stress resulting from the spill was measured by the "Impact of Events Scale?' ( IES ), This instrument consisted of 1 5 items which elicited reports of potential responses to the spiIl. The contents of the guestions which make up the ES were originally derived from statements made by people who had experienced stressful life changes (Horowitz, 1 974,197 6; Horowitz et al,, 1 9 79 ) . Past studies have identified two general response patterns from the TES-inlrusion and avaidance (Harowitz et al., 1979)+ Intrusive stress is characterized by recur- rent thoughts, feelins and dreams which remind one of the stressful events. Avoidance behavior includes purposeful restriction of ideas and behayion

The disruption irrdicamrs also solici~d information on the nature orthe disruption or change, i,e., positive or negative. All rwpdnses were s a n d and only perceiwd negative changs were coded as a disruption.

Industrial Crbis Quarterly, Vol, 6, No. 3

J.S. Picou et a!. / Disruprion and strms

which remind one of the stressful event. The IES appears to be a valid and reliable indicator of community patterns of post-traumatic stress disorders which result from disasters (Davidson et al., 1989; Shore et al., 1989). However, this indicator may actually underestimate the wide variety of stress rates which characterize communities experiencing disasters (Shore et al., I 989 ) . None- theless, the IES provides a valid and reliable indicator of stress which has been used in previous disaster studies (Share et al., 1989; Solomon, 1 989).

Data analysis

Given our objective of describing the initial and continuing social impacts of the Valdez oil spill, the data analysis will proceed in the following manner. First, patterns of socia1 disruption resulting from the spill will be compared for the impact and control communities for 1 989 and 1 990, Secondly, factor ana- lytic scales are generated from the IES indicators and comparisons of levels of intrusive stress and avoidance behavior are made between the impact and control communities. Finally, the relationships between perceptions of source of social disruption and stress are described for the impact community in both 1989 and 1990.

Table 2 provides responses to the five indicators of social disruption for both control and impact communities. In 1989 relatively substantid proportions of respondents in the impact community rereported that the spill had caused social disruption. A11 five chi-square tests revealed that statistically significant differences existed between control and impact community responses. Almost seven out of ten respondents in the impact community in I989 reported they experienced disruption at work (68%) and more than half indicated that they changed personal plans ( 5 1%) because of the spill (Table 2). Almost all of the respondents (96%) in the impact community reported that changes had occurred in the community five months following the spill. The patterns of social disruption which characterized the control community were similar to those observed for the impact community, but at a substantially reduced rate. These findings indicate that five months following the spill, residents of Cordova experienced significant social disruption in personal, family and work-settings, while almost everyone felt that the spill resulted in changes in the community.

The continuing impact of the Valdez oil, spill is apparent from the comparisons of 1990 responses for the five indicators of social disruption (Table 2 ). In general, a redudion in the prevalence of disruption was found for all sources. However, all but one chi-square test applied for contrasts between the control and impact communities were found to be statistically significant indicating the continuation of social disruption in Cordova 18 months following the spill. Once again, work-related changes, making changes in future plans and community changes were the major sources of disruption. The fact that i n 1 990,

Industrial Crisjs Quarter/y, Vol. 6, No. 3

J.S. Picou et al. I Drsrlrption and srress

Table 3

Related factor matrix for impact of Events Scale and conuol-impact cornmunily response T-tes~ analysis, 1989 and l9gQ

- . . -.

Intrusive+ Stress Avoidance 1989 1990 (Factor I )* Behavior T-test T-tes~

(Factor 2 )* with control wilh wniral cornmunlty communi~y

I th6ugh1 aboyl it when I didn'l mean to. (The thought of !he spill just popped into my head.)

4.9 1 PrS.0001

2.92 P r f ,004

3.'38 P r S .OD1

3.73 Prl.OGQ1

Piclures about it popped into my mind.

Other things kcpt making me have thr>u@ts aboui it (even when I dldn't want to).

I had to s b p myself from getting upset when I though abut ~t or was reminded of it.

1 had trouble falling asleep or staying asleep because pictures or thoughts a k u l i t came into my mind.

I had waves of strong feeling about i~. (Feelings about it jusr seemed lo wash over me. )

1 had a lot of feelings a b u l il thar 1 didn'l know how to deal with or handle.

I had dreams a b u t it.

Reminders of i l broughl back fee l ing 1 firs1 felt about i t .

I suddenly felt like i t was happening all over again.

I ~ried 10 remove it from my mcmory. (To acl as though it had never happened. )

I stayed away from retninders of 11.

r I\\ as ii ii ha6 nu ~W.lkj liappned (or as if it wasn'~ real).

I tried not lo talk about it.

I ~ried dot 10 think abour it. ( I rrkd to force my atternion away from it -perhaps lo other things. )

1989 'Cronbach's Alpha fur Intrusive Stress Scale=.895 Tronbac h's Alpha for Avoldance Behavior Scate= 3 1 5

1990 Tranbach's Alpha for Intrusive Stress SCale=.9 12 "Cronbach's A l p h for Avoldance Behavior Scalr-,885

Ind;~~!rinl C r i ~ i ~ Ounrf~rlv. Vnl. h Nn. 3

J.S. Picou er a/. / Disruprion and stress

84% of the respondents in Cordova perceived negative changes characterized the community, reveals a continuing pattern of community disruption. A substantial reduction of disruption was observed for the control community in 1990. Nonetheless, 13% of the respondents in the control community were found to have work-related disruptions in the fall of 1 990.

The second phase of the data analysis included an analysis of the rEs items and a comparison of stress levels between impact and control communities. The TES was subjected to a principle components factor analysis (DeVellis, 199 1 ). The rotated factor matrix clearly identified two factors which parallel the "intrusive stress" and "avoidance behavior" patterns observed in previous research (Horowitz et al., 1 979 ) . The indicator of intrusive stress consisted of 1.0 items (Table 3 ) . The Cronbach's Alpha for this scale was 3 9 5 in 1 989 and .9 t 2 in 1 990, indicating high reliability. The avoidance behavior scale was comprised of five items, which loaded on the second factor. The reIiability coefficient for this scale (Cronbach's Alpha) was .8 1 5 in 1 989 and ,885 in 1 990 (Table 3 ) . These results suggest that intrusive stress and avoidance behavior scales were identifiable outcomes of the Valdez oil spill five months following the spill and continuing over an 1 8-month period.

Table 3 also provides an item-by-item comparative analysis of levels of stress between the control and impact communities. T-tests were calculated for each item, revealing significantly higher rates on 14 of 15 comparisons in 1989 and all 15 item comparisons in 1990. Higher stress levels were found in the impact community for all item comparisons.

A more concise indicator of stress was created by summating scale scores for the leitern intrusive stress scale and the 5-itern avoidance behavior scale (DeVellis, 199 1 ). Table 4 presents the mean scale scores for both indicators of stress for 1 989 and 1990.

Statistically significant differences in levels of intrusive stress and avoidance behavior were found to exist between the impact and control communities in both 1989 and 1990. A comparison of mean intrusive stress scale scores in the impact community in 1989 ( ji = 24.47 ) and 1 990 ( % = 19.32 ) suggest an overall reduction in the level of intrusive stress occurred over the 12-month period following the initial data coUection. Nonetheless, the corresponding reduction observed in the control community (t = 19.5 1 in 1989; x= 13.42 in 1990) resulted in continuing statistically significant differences in stress levels for the impact community. It appears from the data that the relative impact of the Valdez oil spill in the two communities was greater for Cordova in both 1 989 and 1990. However, the intmsi ve stress level observed in Cordova in 1 990 is actually comparable (in terms of mean values) to the stress generated by the spill in a "minimally impacted community", i.e., Petersburg, for the fall of t 989 (Table 4 ) .

Industrial Crisis Quarterly, Fol. 6, No. 3

Table 4

T-tesl for Intrusive Stress and Avoidance Behavior Scales, 1989 and 1990

Scale ii S.D. T-test

fnlrusive Stress impaci community 24.47 7.72 4.46 (n= 118) Prs.0001

1989 Control community 19.51 6.81 (n=73

Impact community 19.32 7.47 1990 (n=69)

Control community 13.42 4.94 (n=53)

Avoidance Behavior Impactcommunity 8.78 3.89 (n= 118)

1989 Control community 6.3 I 2.56 (n= 73)

Impactcommunity 8.93 4.17 6.38 (nm69) PrS.000i

1990 Control oornmunity 5.47 1.48 (n=53)

The comparative analysis of the avoidance behavior scale reveaIed a differ- ent pattern (Table 4). Significantly higher levels of avoidance behavior were found for the impact community in both 1989 and 1990. Correspondingly, no change in the level of avoidance behavior was observed in the impact community from 1 989 to 1 990. A slight decrease in the mean for avoidance behavior scare was observed over time for the control community (x= 6.3 1 in 1989; %= 5.47 in 1990). These findings further clarify the patterns of stress observed in Table 3. Over time, the levtI of intrusive stress has declined in Cordova while the level of avoidance behavior has remained stable. This reduction in intrusive stress levels was relative, given the finding that significantly higher intrusive stress levels characterize the impact community when contrasted with the control community in both 1989 and 1990.

Indmtrid Criss Quurteriy, Vd 6, No. 3

J-S Picou et a/. / Disruption and sfrm

Table 5

Pearson mrrehtion coeficients for stress, disruption and demographic characteristics for impact community, 1989 (N= 1 18) and 1990 (N= 6 9 )

Characteristics Scale

Intrusive Stress Avoidance Behavior

Family disrupt ion Changed persoad plans Family members changed plans Work disruption Community disruption Educational achievemen1 Sex Years lived in communily Fishermen

Table 5 presents Pearson correlation coefficients between the stress scales, social disruption indicators and selected demographic characteristics for impact community respondents. In 1989, five months following the spill, increased levels of intrusive stress were significantly correlated with family disruption, changing of future personal plans and having family members who changed their personal plans (Table 5 ). Additionally, when compared to community members in other occupations, commercial fishermen were characterized by higher levels of intrusive stress ( r= .27 ). In 1989, higher avoidance behavior scale scores were found to be significantly correlated with the three disruption items found to be associated with intrusive stress. The finding that most demographic variables were not associated with increased levels of stress suggests that initial stress generated by the spill in the impact community was associated primarily with patterns of social disruption rather than demographic characteristics.

A pattern of association between several indicators of social disruption and psychological stress persisted 1 8 months after the spill. The Pearson correlation coericients between family, work and personal disruption and the two stress scales increased in magnitude in 1990. The most dramatic increases occurred between reports of disruption at work and both stress indicators. A coef-


J.S. Picou et a/. / Disrupriorl nnd srrexs

ficient of . 39 was observed between work disruption and intrusive stress, while a coefficient of .4Q was found between changing personal plans and intrusive stress (Table 5 ), In addition, higher rates of avoidance behavior were found to be specifically correlated with family disruption and changing personal plans in 1990. These findings suggest that, over time, the disruption of work-related activities is associated with higher levels of stress. Furthermore, continuing persona1 and family disruption is also associated with higher stress-levels, although the initial consequences of work disruption may not involve elevated stress- level^.^ In 1 990, the correlation between being a fisherman and work disruption was significant ( r r . 2 8 1 , suggesting that the long-term impacts of the Valdez oil spill include a continuing pattern of work-disruption far people involved in commercial fishing some 18 months following the spill.

Summary and conclusions

The ecological-sym bolic approach to technological disasters facilitates an understanding of natural resource communities, the cu ftural traditions associated with their harvests of renewable resources and their vulnerability, both em- nomic and cultural, to resource contamination. Given the fact that most Alas- kan communities in the Prince William Sound area can be classified as natural resource communities, the primary objective of this research was an evaluation of the initial and continuing social impacts of the Valdez oil spill. Initial impacts were measured five months following the spill and continuing impacts were measured approximately eighteen months after the spill. The specific social impacts considered were social disruption (personal, family, work and community settings ) and psychological stress (intrusive stress and avoidance behavior). The research design included a matched control community for comparing relative levels of disruption and stress in both Cordova, Alaska (impact community) and Petersburg, Alaska (control community ) .

The statistical analysis of the data clearly and consistently documents significantly higher levels of social disruption and psychological stress in the impact community in 1989 and 1990. This general finding collaborates the ever-in- creasing research literature which documents a variety of negative, long-term social impacts for technological disasters ( e.g., see among others, Davidson and Baum, 1986; Smith et al., 1986; Hauts et d., 1988; Freudenburg and Jones,

'The statistical significance of the calculated Pearson correlation coeficients are sensitive to sample size. The smaller sample size of the 1990 data-sets are responsible for the lack of statistical significance, despite coefficients of similar magnitudes. Given this consideration, the coefficients in Table 5 for 1990 should be evaluated in terms of "substantive" significance, as well as "stat istical" significance.

Industrip1 Crisis Quarterly, Voi. 6, No. 3

J.S. Picou el al. / Disrupt ion and sfress

199 1 ; Gill and Picou, 199 1 ). More specifically, the findings reveal a longitudinal pattern indicating some reduction in the scope of disruption and Ievel of stress from 1989 to 1990, This decline was not observed for one stress scale (avoidance behavior), which revealed consistent levels for both initial and long- term assessments. Furthermore, overtime, stress and disruption appear to be increasingly correlated, suggesting a focusing of negative social impacts among community residents who experience continuing disruption in family, work and community settings and higher-levels of personal stress.

Given that demographic characteristics considered in the correlation analysis were not found to be significantly associated with our stress scales, it appears possible that cultural values unique to naturd resource community mem- bership may influence the longitudinal pattern observed for disruption and stress. Although participation in commercial fishing was initially found to k modestly correlated with intmsive stress in 1989 ( r = .27 ), no such relationship was observed in 1990. However, work disruption was correlated with intrusive stress in 1990 and fishermen experienced more work disruption than respondents in other occupations. It may be possible that other natural resource community activities such as participation in subsistence harvests, etc. may identify subpopulations more vulnerable to long-term negative social impacts. Future research on technological disasters in general, and especially on the Valdez oil spill should attempt to further identify characteristics of s u b p pulatioas in impacted communities that experience long-term disruption and stress. At this time, empirical research indicates that both Native and non-Na- tive residents of Cordova have experienced long-term negative social impacts 18 months following the Valdez oil spill in the form of continued disruption of work-roles and increased levels of personal stress (also see Dyer et al., 1992). In summary, although the research design employed in this research focused

on a detailed analysis of one fishing community in Prince William Sound, i.e., Cordova, we suggest that the natural resource community model provides a conceptual link for generalizing the observed patterns of social impacts to other resourcedependent communities in Prince W iiliam Sound. Furthermore, the ecological-symbolic approach to studying technoIogica1 disasters provides an appropriate framework for understanding that cultural and personal consequences result from the disruption of the relationship between human communities and their biophysical environment through technological contam i- nation. The findings of this research provide a description of initial and continuing patterns of disruption and stress resulting from the Valdez oil spill. Future studies should attempt to develop explanatory models of the long-term social impacts of technological disasters for natural resource communities. Such research would enhance the development of mitigation strategies based on an ecologically-informed model of the human community.

Industrial Crisis Quarterly, VoI- 6, No. 3

J-S, Picou er al. / Disruption and stress

Acknowledgements

Funds for this research were provided by grants from the National Science Foundation ( DPP-9 10 1093 ) and the Natural Hazards Research and Applica- tions Information Center, University of Colorado (Boulder). This research was also supported by resources provided by the Coastal Research and Develop- ment Institute, University of South Alabama, the Social Science Research Cen- ter and the Mississippi Agricultural and Forestry Experiment Station (Project [ MIS-43 1 S ) , Mississippi State University and the Fisheries Art Collective, Santa Cruz, California. Special thanks are extended to Bob Shipp, Brad Davis, Art Cosby , Pat Picou, Nancy Bird, Chris Christensen, J, Stephen KrolI-Smith, Wolfgang Frese, G. David Johnson, Jon Carr, Penny Oswdt, Rhonda Bates, Linda Burcham and Patience Anderson for their respective contributions to the completion of this study. The on-site field support provided by the respective staffs of the Copper River Delta Institute and the Prince William Sound Science Center is also acknowledged. The authors remain solely responsible for the contents of this article.

References

A h m , F.L. and Coheo, R.E., 1984. Disasters and Health: an Annotated Bibliography. National Institute of Mental Health, Rockville, MD.

Barton, A.H., 1969. Communities in Disaster: a Sociological Analysis of Collective Stress Situ- ations. Doubleday, Garden City, NY.

Baurn, A., Fleming R. and Singer, J.E., 1982. Stress at Three Mile Island: applying psychological impact analysis. In: L. Bickrnan, ed. Applied Social Psychology Annual, 3, Sage, Newbury Park, CA: 2 I 7-248.

Baum, A,, Fleming R. and Singer, J.E., 1 983. Coping with victimization by technological disaster. Journal of Social Issues, 39 (2 ): 1 1 7- 1 38.

Eogard, W., 1989. The Bhopal Tragedy. Westview, Boulder, CO. Brown, G.W. and Hams, T., 1979. Laying Waste: Lave Canal and the Poisoning of America.

Random House, New York. Couch. S.R. and Kroll-Smith, S., 1 9 8 5 . The chronic technical disaster: toward a social scientific

perspective. h i d Science Quanerl y, 66: 564-575. Davidson, A*, 1990. In the Wake of the E m n VaIdex: the Devastating Impact of the AIaska Oil

Spill. Sierra Club Books, San Francisco, CA. Davidson, J.. Smith R. and Kudler, H., 1989. Validity and reliability of the DSM-I11 criteria for

posttraumatic stress disorder. Journal of Nervous and Mental Disease, 177 ( 6 ) : 336341. Davidson, L.M. and h u m , A., 1986. Chronic stress and post-traumatic stress disorders. Journal

of Consulting and Qinical Psychology, 54: 303-308. DeveUis, R.G., 199 1 . Scale Development: Theory and Applications. Sage, Newbury Park, CA. Dew, M.A., Bromet, E.J. and Schulburg, H.C., 1987. Mental health effects of he Three Mile

Island nuclear reactor restan. Amenan Journal of Psychiatry, 1 44: 1 074- 1077. Dillman, D.A., 1978. Mail and Telephone Surveys. Wiley, New York.


J.S. Rcou et al. / Dismpfim andstress

Dahrenwend, B.S., 1 978. Social stress and community psychology. American Journal of Corn- munity Psychology, 6 ( 1 ): 1-14.

Drabek, T.E., 1986. Human Systems Responses to Disaster. Springer, New York Drabek, T.E. and Key, W.H., 1976. The impact of disaster on primary group linkages. Mass

Emergencies, 1 ( 2 ) : 89-105. Dyer, C.L,, Gill, D.A. and Picou, J.S., 1992. Social disruption and the Vak&z oil spill: Alaskan

Natives in a natural~urwcommunity. SociologicalSpec'trurn, 12 ( 2 ) : 105-126. Dynes, R.R., 1970. Organized Behavior in Disaster. Heath, b i n g o o , MA. Edelstein. M. R., 1 9 88. Contaminated Communities: the Social and Psychotogical Impacts of

Residential Toxic Exposure. Westview, Boulder, CO. Erikson, K.T., 1976. Everything in Its Path: Destruction of Community in the Buffalo Creek

Flood. Simon & Schuster, New York. Erikson, K.T., 1991. A new species of trouble. la: S.R. Couch and J.S. Kroll-Smith, eds. C o n

munities at Risk. Lang, New York. Fall, J.A., 1990. The Division of Subsistence of the Alaska Department of Fish and Game: an

overview of its research program and fmdings: 1980-1 990. Arctic Anthropology, 27 (2): 68- 92.

Freudenburg, W. R. and Jones, T.R., 1 99 1. Attitudes and stress in rhe presence of technological risk: a test of the Supreme Court hypothesis. Social Forces, 69 (4): 1 143-1 168.

Frey, J.H., 1983. Sumey Research by Telephone. Sage, Newbuly Park. CA. Frisbie, W.P, and Poston, D.L., 1975. CompocLents of sustenance organization and nonmetro-

poiitan population change: a human ecological investigation. American Sociological Review, 40: 773-784.

Fritz, C.E., I96 1. Disaster. In: R.K. Merton and R.A. Nisbet, eds. Contemporary Social Prob- lems. Harcourt, New York: 65 1-694.

Gibbs, J.P. and Martin, W.T., 1959. Towards a theoretical synem of human ecology. Pacific Sociological Review, 2: 29-36.

Gill, D.A., 1986. A disaster impact a W m e n t model: an empirical study of a technologicat disaster. Unpublished dactobl dissertation, Department of Sociology, Texas A&M Univer- sity, College Station, TX.

Gill, D.A. and Picou, J.S., 1989. Toxic waste disposal s i ts as tachnological disasters. In; Psycho- social Effects of Hazardous Toxic Waste Disposal on Communities. Thomas, Springfield, IL: 81-97.

Gill, D.A. and Picou, J.S., 1991. The social psychological impacts of a technological accident: collective stress and perceived health risks. JournaI of Hazardous Materials, 23 ( 1 ): 77-89.

Cdlespie, D.F., 198%. Barton's theory of coUective stress is a classic and worth testing. Interna- tional Journal of Mass Emergencies and Disasters, 6 ( 3 ) : 345-36 l.

Gleser, G.C., Green, B.L and Wing* C.N., 198 1. Prolonged Psychosocid Effects of Disasters: a Study of Buffalo Creek. Academic, New York.

Hawley, A, 1950. Hunaa Ecology: a Theory of Community Structure. Ronald, New York. Horowitz, M. J ., 1 974. Stress response syndromex character style and brief psychotherapy. Ar-

chives of General Psychiatry, 31: 758-781. Horowiu, MJ., 1976. Stress Respnse Syndromes. Aronson, New York. Horowitz, M., 1990. Post-uaumatic stress disorders: psychological aspects of the diagnosis. In-

ternational Journal of Mental Health, I 9 ( 1 ) : 2 1-36. Horawitz, M.J., Milner, N. and Alvarez, W., 1979. Impact of event scale: a measure of subjective

stress. Psychosomatic Medicine, 41 (3 ): 209-2 18. Houts, P.S., CLary, P.D. and Hu, T.W., 1988. The Thm Mile Island Crisis. PennsylvaoiA State

University Press, University Park, PA.

Industrial Crisis Quart~rly~ V d 6, No. 3

J.S. Picm ct al. / Disruption and srress

Keeble, J., 1991. Out of the Channel: the Exxon Vddtz Oil Spill in Prince William Sound. Har- per Collins, New York.

Krep. G.A., 1985. Disaster and the social order. Sociolagical Theory, 3: 49-65. b p , G.A., ed., 1989. Social Structure and Disasters. University of Delaware Press, Newark, DE.

Kroll-Smith, J.S. and Couch, S.R.. 1989. Some thoughts on natural disasters: technological hazards and social change. Environment, Technology, and Society, 56: 2-4.

-11-Smith, J.S. and Couch, S.R., 1990. The Real Disaster is Above Ground: a Mine Fire and Social Conflict. University of Kentucky Press, Lexington, KY.

Kroll-Smith, J.S. and Couch, S.R., I99 la. What is disastef? An ecological symbolic approach to resolving the definitional debate. International Journal of Mass Emergencies and Disasters, 9 ( 3 ) : 355-366.

Kmll-Smith, J.S. and Couch, S.R., 199 1 b. Tschnological hazards, adaptation and social change. In: S.R. Couch and J.S. Kroll-Smith, eds. Communities at Risk. Lank New York.

Livessay, J.M., Boyer, J.C. and Harding, J.R., 1984. Social impact assessment: conceplual frameworks and their implications for planner. In: W. Nellsap, ed. Applied Social Science for Environmental Planning. Westview Press, Boulder, CO: 23-35.

Lounsbury, J. W., Van Liere, KD. and Meissen, G.J., 1983. Psychosocial assessment. In: K. Fin- sterbush, LG. Lltwellyn and C.P. Snow, eds. Social Impact Assessment Methods. Sage, New- bury Park, CA.

Maser, J.D. and Solomon, S.D., 1990. Psychosocial and psychobiological consequences of exposure to extreme stresmrs. f aurnal of Applied Social Psychology, 20-2 1 : 1 725- 1 732.

Mileti, D.S, Dmbek, T.E. and Haas, J.E., 1975. Human Systems in Extreme Environments: a Sociological Perspective. Ins ti tute of Behavioral Sciences, University of Colorado, Boulder, CO.

Murdock, S.H., 1979. The potential role of the ecological framework in impact analysis. Rural Soc i~ log~ , 44 (3): 543-565.

National Response Team, 1989. The Exvon Valdez Oil Spill. Report to the President, US Pe- partment of Transportation, Washington, DC.

Nichols, B., 1989. Bud deathsalann Alaska. USA Today, 7 September: 3A. Omohundro, J.D., 1982. The impacts of an oil spill. Human Organization, 4 1 ( 1 ): 1 7-25. Pimu, J.S., 1984. Ecological, Physical, Economic, Sociological, and Psychological Assessment of

the Iliinois Central Gulf Train DErailment. Volume 5: Sociological Assessment. Gulf South Raarch Institute, Baton Rouge, LA.

Piwu, J.S., Gill, D.A, Dyer, CL. and Curry, E., 1990. Social Disruption and Psychological Stress in an Alaskan Fishing Community: the Impact of the Exxon Vddez Oil Spill. Quick Response Research Report QR35, Natural Hazards Research and Applications Information Center, University of Colorado, Boulder, CO.

Quarantelli, E.L., 1985. What is a disaster: the need for clarification in definition and conup tualization in restarch. International Jwmal of Mass Emergencies and Disasters, 5: 7-32.

Quarantelli, E.L. and Dynes, R.R., 1970. Editors' introduction. American Behavioral Scientist, 13 (3): 325-330.

Restoration Planning Work Group, 1990. Restomtion Planning Following the E m n Valdez Oil Spill: August, 1990. Progress Report, State of Alaska, Depanment of Interior, Anchorage, AK.

Salaman. G., 197 I. Some sociological determinants of occupational communities. Sociological Review, 19 ( 1 ): 53-76.

Shewonit, B., 199 1. Whales on the wane. Anchorage Times, 15 December H1 and H4.

Industrial Crisis Quarrerly, Voi. 6, No. 3

J.S. Picou et al. / Disruption and sfress

Shore, J.H.. Yollmer, W .H. and Tatum, E.L., 1989. Community patterns of posttraumatic stress disorders. Journal of Nervous and Mental Disease, 177 ( 1 1 ) : 68 1-685.

Shrivastava, 1987. Bhopal: Anatomy of a Crisis. Harper & Row, Cambridge, MA. Smith, E., Robins. L.N ., Przybeck, T.R., Goldring, E. and Solomon, S., t 986. Psychosocial con-

sequences of a disaster. In: J.H. Shore, ed. Disaster Stress Studies: New Methods and Find- ings, American Psychiatric Press, Washington, DC: 49-76.

Smythe, C.W., 1 988. Harvest and use of fish and wildlife resources by residents of Petersburg, Alaska. Technical Paper No. 164, Alaska Department of Fish and Game, Division of Subsist- ence, Anchorage, AK.

Solomon, S.D., 1989. Research issues in assessing disaster's effects. In: R. Gist and 3. Lubin, eds. Psychosocial Aspects of Masters. Wiley, New York: 307-340.

Stanek, R., 1885. Patterns of wild resource use in Tatitlek, Alaska. Technical Paper No. 181, Alaska Department of Fish and Game, Division of Subsistence, Anchorage, AK.

Steinglass, P. and Gerrity, E., 199 1. Natural disasters and post-traumatic stress disorders: short- term versus long-term recovery in IWO disaster-affected cummunities. JournaI of Applied SQ- cia1 Psychology, 20-2 1.: 1746- 1765.

Srewan, R.B., Gersorl, S.M. and Bottini, 1 .W., 1 99 1. Nolice of lodging of United States' summary of effects of E n o n Valdez oil spill an natural resources. Filed in US District Court, District of Alaska, 8 April, Civil Action No. A91-082.

Stratron, t., 1989. Resource use in Cordova: a coastal community of Sou~hcentral Alaska. Tech- nical Paper No. 1 53, Alaska Depanment of Fish and Game, Division of Subsistence, Anchor- age, AK.

Turner, B.A., 1978. Man Made Disasters. Crane, Russak, New York. Yeomans, K.A., 1968. Applied Statistics: Statistics for the Social Scientist, Vol. TI. Penguin,

Middlesex.

Indusirial Crisis Quarterly, Vol. 6, No. 3

the exxon valdez oil spill -...

Documents