ML20235D969
| ML20235D969 | |
| Person / Time | |
|---|---|
| Site: | Shoreham File:Long Island Lighting Company icon.png |
| Issue date: | 07/08/1987 |
| From: | Jerrica Johnson CALIFORNIA, UNIV. OF, LOS ANGELES, CA |
| To: | |
| References | |
| OL-3-A-015, OL-3-A-15, NUDOCS 8709250428 | |
| Download: ML20235D969 (10) | |
Text
~
,'x y
.n.
.n..
gy
. p.
)J!f.Nhi$F
" '5W ~
u s
.M Ae4 % eshnew e t= E._
les...
/
de -322. M 3 J JL do k' Y~}w,..fkJ 7
- o. m vet ti w r m
,m s m sino t om sne w a.m.,
w~o
.a, tu
- /5 N0DCE Ms material may be prote$7d SEP 17 g130 copyright lew (Title 17 U.S. Code)
PREDICTING NUCLEAR REACTOR EMERGENCY f
EVACUATION BEllAVIOR
\\
J ms 11. Jonwn Ja j
Dopanme-* if Geog aphy, Unher i 3 of Cabrer.u at Lm.\\ngeles Los hg.... L \\ uX (*.SA
.{
shcul 12 Wm6cr IW5)
Abstract-!..;ime re:ression.uulpes er sursey d.:t.i from a sample d h; ec.qda withm the ucmit) ef ": Sherchm Nuclear Pouer Statien support nearly all of de h.po:-..sind Imkarcs in a theor: al model of nudear reactor emergene) c.act.ation decWr..- ab:.g. the results indicate the if a TMI type uccident were to oee r at the Shorcham plant. rerly one third of the i
t.ong Islar, popalaaen (-25%00 household" would esacuate wlunwr4 er pmtancously.
Implicatiora br future. oriented nucicar ernergen;> evacuation planmng are oeuned.
?
iNfhoiH CitON The accident at the Three Mile Island (TMii nuclear generating fadlity near liarrisburg, Pennsylvania 9. March,1979, has sparked a resurgence of research interest in human behavior in cris3 situations. Resuhing in part from human error and partly from mechanical failure, and im ching an off-site re! case of radioactis e materials, the T MI accident prompted the governor ci Pennsylvania to instruct pregnant women and prc>chool children within 5 miles of the rnalfunctioning reactor to esacuate, and everyone che within 10 miles to stay indoors. This protective action advisory should have precipita ed the esacuation of only about 3M people.' Post accident surseys rescal, honescr. that as many as 200,000 a
peopic within 25 mile radius of the crippled TM1 reactor actually esacuated.2 The process L
whereby individuals and families evacuated esen though they were not advised to take such drastic pre:ectisc measures has been termed spontancoits esacaation.3 Spontaneous evacuation has occurred in other crisis situations, but never of the magnitude obsersed during the TMI reactor crisis. In fact, the gap betucen actual and expected evacuation was so wide thu one group of experts concluded their review and esaluation of the TMI accident by rec 0mmending that further research be done to develop methods for estimating.
the flow and euent of spontaneous esacuation mosements from the time a crisis begins.)
Elsewhere, a theoretical model of nuclear reactor emergency deci3 ion. making and behavior is de. eloped.' in this paper, the results of an empirical te3t of this model are presented. The primary objectile here is to demonstrate that, for the population within the vicinity of any commercial nuclear installation, the model can be used to predict evacuation betnior should a major reactor accident occur in the future. To place the rescarch in its proper policy context, we begin by providing a general oserview of existing federal regulations governing commercial nuc! car power plant operations, focusing specifically on ahy esacuation plans developed according to these regulations are likely -
$ g to be inadeque: and ineffective in an actual radiological emergene).
n.
b k
Tilli R ESE ARC 11 CON f EXT CD O Implicit in the federal regulations governing commercial nuclear power plant operations hg in the U.S. is the notion that people's behasior in a radiological emergency will not be f
j Ei5 unlike public reactions to warnings of natural disasters (c.g. flood 3 and hurricanes) and :
c trHl:
non-nuclear. te.hnological emergencies (e.g. accidental explosions and transport accidents
[
imolving toxic materials).5 The guidelines assume that, if a major reactor accident were >
f$ to occur, emergney management ollicials' most significant challenge aeuld be getting thi CDG O potentially imps.:ted population to esacuate their homes until the immediate crisis is over.
Evidence ama> :d oser the last 3 decades in studies of non. radiological disasters indica'tes LU llM dl
i h
)
i -
l I-l-
1
,o s' 8
- A
$ \\'M t
e.
b
'9 N'N '
<,jd 1>\\x.
/
- p s_
$gh y
- l, s' r, '
N, s.,jspe mm.
i
I e
m J.11. June h that there is a strong tendency for people who are threatened by impending di> aster to ignore the official esacuation advisory, rationalizing that the threat will no: materialize l
(denial reaction) or that they will he safe even if it does (avoidance react 7nd Cross community studies hase shown. however, that the level of compliance with evacuation advisories in natural disasters is directly related to. among other factors. the t;.;e content, and frequency of the the warnings.' More specifically, these studies indicate the a majority, if not all, of the potentiall> impacted population adhered to the oRicial evacuathn adsisory in crises where cicar, concise. and consistent emergency information was issued at frequent intervals.
Ilased in part on the fmdings and implications of the cross community natral di> aster i
studics, and in part on the fact that conflicting information was the second moc frequently cited reason for esacuating duri.ng the TMI crisis, the federal regulations regt.:re that the population within the vicinity of commercial nuclear power plants be exposed to 2 types i
of emergency information. The first is a pre-emergency education and informati.: a brochure which licensecs,in conjunction with state and local governments, are required te distribute (at least annually) to the population within a 10 mile radius of their facilities.The brochure is supposed to contain mformation pertaining to the risks olt radiation and the range of protective action recommendations which might be issued in case of a nue: ar reactor cmcrgency. The second is emergency notification messages, that is, informati:n which is -
to be issued at the onset and during a reactor crisis. The regulations require licensees, along with designated state and local organizadons, to establish the contents of the initial and follow-up messages to be sent from the plant. and to devise a system for dineminating this information to the public in an unambiguous and timely manner 8 It is argued here. honeser, that in the event of a nuc! car reactor acciden; the.most important emergency management problem, at least from a behavioral standpoint. will not be people's unwillingness to evacuate, as in non-nuclear emergencies. Rather, it most likely will be people voluntarily taking protective and more drastic actions,i. particular v
evacuating. soor.er than the situation may warrant, as happened in the TMI rea:ter crisis?
g It is further argued that, while efficient and effective communication with the public is 4
unquestionably crucial for the successful management of any crisis situation, the 2 sources of emergency information described above will not noticeably alter the pr:pensity of individuals and families to spontaneously evacuate in a nuclear reactor emerger:y because, in several critical respects crises involving the risks of radiation are significant: different from other kinds of emergencies.
First, as risk perception studies have repeatedly shown, nuclear power is U:wed with more dread than any other activity except terror sm and warfare,in large par % au.e of the potentially letnai. care:nogenic, and mutagenie ettects ofionizing radiation / Peaceful uses of nuclear power. such as the generation of electricity, are often associat:d with the atomic weapons program of World War 11 which culminated in the bombing o! Hiroshima f
and Nagasaki. The deaths and human sulTering caused by these catastrophic c.ents loom large in the collective social memory of the public." For this reason alone. a nuclear reactor accident can be expected to generate higher levels of stress and,in turn, c. asiderably more hypervigilant behasior than emergency nonfications of other types of er ces.
Second, unlike natural and other technological disaster agents. such as flood.uters and noxious gases from accidental explosions, radiation is an ambiguous hazard f: :he most s
part it is imperceptible to the senses, except in doses sullicient to 'nduce radiatt:. dekness.
2 For evacuation decision making this is an especiall> important distinctio: Previous i
rescarch has show n the ability to confirm official warning.s through direct sense cadence (e.g. the winds asveiated with a hurricanc) to be a critical determinant of evac.a: ion from s
natural disasters.' In a nuclear reactor emergeney, however. it is highly u..' Aely that l
[
environmental cues woulJ exist. The public therefore would be forced to reb u!cly cn
}
secondary (technical esperts and government olliciah) and tertiary (news meau. Inends.
and relatives) analy.es and interpretations of the events surrounding a mz reactor l
accident which ma3. as in the TMl crisis, be seen as unreliable. biased. and.or centradic.
l
,.W tory." Morcoscr. esen ifit is assumed that the emergency information issued k i he clear, i
/
T-
]
Preactir.g nudear reaemr enmency emor hr w
1 conei c,.onsigent. and timely, social >urveys conducted "ithin the sicinity of >eseral UX nuclear cactor. indicate that a major portion of the popuLdion would hate little or no faith " ollicial warnings and protectise action recommendanons, inespective of the sourcec This high lesel of distrust i> in large part a rc&ction of the fact tht the federal
- 5. ' _
govern ent, a h.: premeting its des elopment has failed to demonstrate that it can protect j
the pub: from se hazards of nuelcar power. In short. w hereas in non-nuclear emergencies the prennec of e:nironmental cues enhances compliance with omeial evacuation adiisories J
l and thus minimizes extreme behasior,in a nuclear reactor accident the absence of direct i2-sensory nidence may result in the voluntary or spontaneous esacuation oflarge numbers of peopb who technically may not be at risk."
Final:. a we lth of e(idenc'e exists which indicates that special programs designed to educate cople about simultaneously low probability and high risk events have not been 18 particule.rly successful in changing perceptions ind behasior?' With respect to nuc! car If power..orcoser the c.idence suggests that information which contradicts or conflicts p
with enting perceptions, beliefs, and attitudes most likely will be characterized as g-s inaccurre, erroneous or unrepresentative and thus will be rejected or dismissed. On the other ha.d. information that either supports or reinforces pre existing views and opinions fi most like!y will be accepted."It therefore shou!d not be autmed : hat a public education T
pmgr.m.vili necessarily result in a high lesci of compliance with onicial protectise action Y
recommendations in the event of a radiological emergency, f
For these rea ons, in short. csacuation decisions most likely will be made at the r
indisidea household level and be strongly innucnced by such factors as proximity to the 4;
malfunct.cning reactor. social background, and family status. In the section which follows.
R a model :f radiological emergency decision-making is described that incorporates these j-and other *.ariab!:s as predictors of who is mo3t likely and who is least likely to esacuate
+
voluntar-:y or spontaneously in the esent of a major nuclear reactor a:cident.
!_g h
A NtODEL OF E V ACtJ ATIO N DECISION.NI A K ING dk The th:oreticci underpinnings of the proposed model of esacuation decision making T,
hate bee. discuned in detail elsewhere and therefore are adJressed in only abbreviated
[
form her: # The model is premised on the propositions that, in the absence of sufticient e
empiricai evidence pertaining to actual behavior," evacuation plans must be based on the
[
behasioral inten:!ons of [the] population which perceives itself to be at risk, and that a c
social suney administered during the planning process is the best way to ascertain how
[
residents are likely to behave in the event of an actual radiological emergency "" Reliance y
on intended behavior (i.e. cvacuation intentions) as a ba3is for esacuatiun planning is f
advocated mainly on theoretical grounds. Fishbein and Ajren's theory of reasoned action 4-posits that intended behavior is the best predictor of actual behavior as long as the II intention does not change before the individual has a chance to act out the behavior.F 1
Surveys canducted within the vicinity of commercial nuclear power plants in New York V
and Califcrnia suggest strongly that this proposition is especially applicable in the present k
research context. Conducted over a 3 r period when the nuclear industry was plagued 3
e witn scr) urious probicms related to the design, construction. and operation of commercial j
reactors. dese suneys indicate that peoples' behavioralintentions in the event of a nuelcar i
reactor ueident r.ot only are relatively stable but also closely parallel actual behavior during th: TMI cnsis.'3 On the basis of these consideration > intended behavior is judged to be a re:.able ir.Jicator of actual behavior in the event of radiological emergency.
As Fig. ! shou. the model is comprised of 3 categorical esogenous variables (location.
stage in-L:: cycle. and social status) and 4 endogenous variables inuclear power attitudes.
nuclear ace: dent r:3k perception. behavioral intention and behasior). The structure of the model re. lses primaril.s around 5 research hypotheses which can' be summarized as follow s.
1 Hypoth@ 1. The location of a household's residence in relation to the plant will directly influence the esactation decision. Two variables. distance and direction of the house from l-
1 W,4 J.H.Jons>os h l
the plan't.are postulated to capture-m a statistical sense-the cifect of the local getjraphy of the reactor site on the hou>chold evacuation decision in a radiological emerger. 3 Hypothess 2. The hou chold's stage in-life cycle will directly influence the esa:uation decision. Families in their childbearing years, especially those who already hate young children in the home. will be rnore likely to esacuate than older families without children in the home.
~ Hypothesis 3. The head of household's social status will directly influence the eva:uation decision. More specifically. high status households will be more likely to evacua:e than lower status fam'ilies.
i Hypothesis 4. The head of household's pre accident attitude toward nuclear power will directly influence the es acuation decision. Those who oppose the use of nuclear power to -
generate electricity will be more inclined, than those who favor its'use, to evacuate in a-
' radiological emergency.
Hypothesis 5. The head of household's perception of the risk associated with a nuclear reactor accident will directly influence the evacuation decision. Irrespective of the distance of their residence from the plant and of the emergency information they receive from crisis management officials, those heads of. household who perceive' themselves and other j
members of their family to be at risk of exposure to a potential or actual release of radiation
}
will be more likely to evacuate than those who perceive that they are not in danger.
In addition to the foregoing hypothesized direct effects, several indirect influer:es on the evacuation decision are also specified in the model. As Fig.'I shows, the household's i
.{
l
%JeLEAR l
/ ' at E
/
g
/:
\\
LOCATION
}
N i'
-- oRECT EFFECTS y'5 g
-- ~ INoiRECT EFFECTS N.N N.
. g s
\\
"cEN'"r
)
A
'* U cNE (#
stAct m
)
w
\\
N
.s
\\..
s
'i I
,/
/
/ /
egy i,o t
eg vion l
sooAL status #'/
1 Fig 1. A genera l:M radd of nuclear reutor ernergency evaeustion decision-mning, phpical location in relation to the plant, stage-in-life cycle, and social stat.s are hypothesized to influence the esacuation decision indirectly through both pre a;:ident attitude toward nuclear power and nuclear reactor accident risk pcreeption. Youn;. hi;;h status families who lite clo3e to the plant will he more likely to oppose nuclear power and to perceive that they are et risk of an accident than older. lower status househous who live farther from the reactor site. Also, pre-accident attitude toward nuclear p0 e.- is hypothesized to intluence the esacuation decision indirectly through reactor accident risk perception. Those w ho oppose the use of nuclear power to generate electricity will N more likely to perceive that they are at risk of an accident than those who favor its use.
In sum, the proposed model of nuclear reactor emergency evacuation decision-mking incorporates intention to e'.aeuate as a surrogate for actual behavior, ano attrib :e5 of 1
'i,
A ' J
v as Prediaing nWe r mw emergency esacuaan # at both :he location and the local residents as predictors of cueuation. In the section which follows. :he results of an crnpirical :st of the model are preser.uJ.
A N F.M NRical. TrsT OF THr. MODM.
]
Data.: :.1 methvJ Data used in this test of the model of esacuation decision-making were taken from a J
social arvey conducted by Yankelovich Skelley, and Nhi:e WSW) for the Long Island Lightir.g Company, the owner of the Shoreham Nuclear Pm'.cr Station (SNPSI. This facility a located on the north shore of Long Island in Su:Telk County approximately 60
]
miles e.ut of New York City. Designed to awess emergency planning needs associated with 11 start up of the SNPS. the sur ey was administered :o a tratified random sampic of 2631 households in Suliolk and adjoining Nassau coundes during a 2 week period in
)
May 1983. 8 Elesen variables were selected from the YSW survey for the purpose of this analysis (Table 1). Consistent with the structure of the model these sanables can be disided into Tn5 e 1. Concepis. seriables und ceilic measures, compiled bs the 4 :t : Irmn i sw sursey Concep'., a,6 s ansrtas y
6 LocauerC Percen ! diatanca X )
Aht how many miles would.s sf e.
- h..e is from the shoreham Nue: car 3
W 4er Piani? Categones rangcd fro? ii. ' m:.e5 er len.. to 14116 to 20 miles t,
i M
ifi 21 to 10 miles.. (%) more than F m: er M
Direct "t (Y;)
CeJed into 2 categoner t1) Ca t 60' We M
Siaye.in. : c3cle Age o auschold nand (X,)
CvJed iniu 5 categonet tu IM 34 2 2f.. ib 35 M 14i 50-64 e5: 65 and oser.
Marn. siatus &
Ceded into 4 casei;ones t n ung'c C % -.:2 di w idow ed (di dhursed. separated.
Your.; :hildren it. *.he home tXa CoJed inin 2 categones: di yes ah e Any * ! pregnant in the home (V )
Cod.d inin 2 categones. tu ye, du r I
Socioem. mee stat.,
1 Educc.an (X 4 Coded intn 5 categones:(0 Lew the hi whool Q H s. grad th some codese
- 14. evliege grad til post grad.
i PsS:h6.e p.eal Gener attitude n nuclear power iN.i See peopte beliese that we houis x
.t.e tJ huild nuCICar Power pIanis in the l, S as a source of clectnc p
.c., %:5 others heliese that the) crease prablems for wmmunities and shm. 4 re,e built. Which one of these pouuons wmes closest to how you feel? Cod:3 m" ) evegones: t t) continue to buiid (2) too many prob ems /stop buildmg 13 # : know Attih..! to SNPs iX.)
LILCO i. currently completmg the WPi. It is not 3et operating. Suppose that it begms to operale, would thr. ma'.c you feet. (1) scr) pica,ed to hee a new sourec of power p) modereely fea :d i e moderately concerned about the l
nn awoci.iied uith it (4l sery cance r.eJ
%eiec reactor accident Gnen uherc 3ou live, do 30u thei ;+..os.:d he afferied if a serious problem fisk pereeption t\\,,,)
. doeloped at the sNPS after it viamJ per.cmg? Coded:(fl ye: 21 no th not
- u. e.
Beha m st trtera:d behavice IX,i)
Sgro.c the you and your farm;y.ere at home and there was an accident at the shPS ihat created a powibhty of c:rea ing ihe lesel of radiation around the plant. All pregnant women ane ;* estocd children hving withm 5 miles of i% plant were advised to evaet. ate DJ eser;.one else lhing wnnin 10 miles of the plant was advised to remain An Ghen where you the. do you thmk l
you aiid other members of your FAT.:i aoCd be hkcl> to: di Do what you uvuld normally do(2) or, stay inud: p c heme, or (3) lease your horne and go l
wrnewhere clwe.
l
- c. ogenous and endogenous variables. The former consi>t of the survey respondents' l
\\
i locational and sociodemographic attributes. while the latter ir.clude their attitudes toward l
nuclear power in general and the SN PS in particular, their reactor accident risk perception.
and their behasioralintention in the event of a radiological craergency at Shoreham. The l
fmal endogenous variabic. esacuation intemion, was derised from responses clicited to a nuclear reactor accident scenario modeled after the protectis e action advisory issued during the TMI reactor crisis. For the purpose of this analysis. the responses to this scenario i
(sumrnarized at the bottom of Table 1) were redefined in the following manner: Individuals
)
who indicated that, under conditions outlineJ in tbc secrark, they would either continue
n
- 1. II. 2nnwss b normal activities or stay in>ide of their homes were labeled potential nonevacum. while those who >tated that they would leave their homes and go some pla' c cl>e ur. icr such c
conditions were classified as potential evaeuces.
Loglinear causal modeling. a modi 6cd version of path analysis, was emplogi in this empirical' test of the mod:1. Whereas in path analysis ordinary least squard tOLSI regression proecdures are used to solve the structural equations corresponding; to the hypothesized linkage > in a causal model, in the loglinear equivalent the logit fur.alation is u. sed to accomplish this' task. Logit is a statistical technique wh' has been dr. eloped specifically for analyzing relationships between dichotomous dependent variaY.es and either categorical or intersal >cated independent variables.
The model reduces to 4 >tructural equations, one for each of the endogenous varabics-general attitude toward nu.. ear power, attitude to SNPS, nuclear reactor accider:t risk perception, and evacuation intention. In standard notational form, the 4 equatiora are as follows:
{ X, X,.., X,}
lX,},
(1) i 2 lX,X,., X, }
lX,},
(2) i 2
lX,X,..,X,)
lXo},
(3).
i 2 i
lX,X,...X o) lX }.
(4) i 2
i ii where the terms defined in Table I and in Fig. 2.
4 All of the hypothesiecd direct and indirect causal linkages in the model are embodied in the>e 4 equations. t; sing logit analysis to solve in succession these 4 equati:ns the analytic objectives were Arst, to derive estimates of the direct and indirect effects of triables X through Xn,. as measured by the beta coefHcients. on evacuation intention (Xal; and i
1 second, to determine how well the overall model fits the observed data. It should be noted here that. because the loglinear version of path analysis is unable to assign single '.alues to causal paths when polytomous variables arc invohed,it was necessary to dichc:omize those variables in Table I with more than 2 response categories.
Restdts Figure 2 depicts the fin d path model of evacuation decision making for the Shc reham plant. For the sake of bresit). the discussion here is limited to the findings pertair/ng to the $ major rc. search hypotheses around which the modelis structured.
The logistic regreshn.ma! pes of th: YSW <urve) data indicate that in the evn of a radiological emergency on Long Island the location of household's residenec in re'ation to the Shoreham facility IX; and X:)is likely to in0uence the evacuation decision directly, and this lends empirical support to hypothesi > 1. As Fig. 2 shows, people who lise : loser to (within 5 miles) and west of the plant are more likely to evacuate than those wb lisc further from (beyond 5 miles) and ecst of the facility.
The YSW survey data ako support the second hypothesis, w hich states that in a r.;elear reactor emergency the es acuation decision will be directly induenced by the life-eyeb stage of the household. Three of the 4 stage-in-life cycle indicators, marital statu> <.V
- d. pt sence of children in the home IXg and the presence of a woman who is pregnant iNp. are 'akely to in0uence the esueuation decision dircelly in a radiological emergency. On Long Fand married couple.4 with children Or a pregnanc) are more likel) to evacuate than unt rried individuals without children or a pregnancy in the home. As Fig. 2 shows. the fourn hfe e>cle indicator, age of hou>ehold head iXf). i< likely to in0uence the esacuation duision indircelly through nuclear attitude iX,d.
The logistic regression ar.alpes do not support hypothesis 3: that the evacuation duision in a nuelcar reactor emergene) will be directly influenced by the social >tatus. the household head. Howeser, the re>ults do sugge>t that. if a reactor accident were te Neur ou Long Island, the social tau, of the household head.as mea >ured b) years of sch:
0!ing i
- 4
'*s" Piedh; ting %4tcof TCat' Lor t'Intrpenc) c\\atw., tsT M a'. Or W
f~'~~>---
N fN
/
s.
'f ij ' #{ 208
~ ~
i x
.a
,N g
+,
M 7-
/
'N
/
u,
,f\\
lL " -%
'x'
% "'</
As 3
'//\\
\\
%_^
~
- 174 g~~y
~ N.s / 5,,, ' di iCf' "6
I N
{
,2c 0
,,.31( N / 's
/
j
't e N
5'?
E-
/
F0t l
s' f.
i @~ 4 / '
/
's.
N Ess't
/<
'.165 - d ' '
/
l m:.; m
/
/
/
p.3 71 y
^
l
-- -F 4-_
- 197 407 i
/:
//
.n J
l EDW ?'
1.e
, /
x, -o......
o x,-e...
-<in ts
//
x,-e.,ui...c...o n, -.m.,.,,, s. < c.a, J
g
,,..~.....m.,
q
=,. u.,,... o. o x g.4...,:....a n.o 3;
,,-cu.. o.o n,,-i.u........
4-x,.,,....
o.o
.w CIAECT ErrsCTS k
WCI:tCT EFFECTS L._.
rig. 2 Final p.uh mdi for the Shoreham nuclear pew st., tion.
r1 j
F comple:ed (X-L would most likely influence the evacuation decision indirectly through g
nuclear reactor accident risk perception (X o).
i i
That the evacuation decision will be directly innuenced by the househvid head's attitude 3.-
toward nuclear power (Bypothesis 4)is supported by the YSW survey data. Figure 2 shows i
i that those Long Islanders who oppose nuclear power in general (X,) and the SNPS k
partice:ar (Xyl are more likely to esacuate than those who favor both.
7 Fina!!y, the YSW suncy data lend empirical support to hy pothesis 5: that the evacuation y
deci>io. will be innuenced directly b the household head's reactor accident risk perception 3
IXat Nassau and Suffolk counties residents who beliese that they and other family
{
membert would most likely be alTected if a serious problem descloped at the Shorcham
(
plant et more likely to evacuate than those w ho feel that they would not be at risk if an en crger.cy were to occur once the plant begins to operate.
y in >t.m. the results of the loE stic regression analyses of the YSW suncy data lend strong i
i t.mpirini support to 4 of the 5 h)pothesized direct linkages in the model of nuclear reactor
)
ernerger.cy evacuation decision making. In addition, the YSW suncy data also support I
I nearly sii of the indirect linkages in the model. although these results are not discussed in detail bre. Exactly how well does the model fit the obsened data? In the loglinear sersion of path analysis the goodness-of-fit statistic is p ; for the proposed model, the p salue is 2
0.21. wh.:h is indicatise of a relatisely good tit. In fact. p salues between 0.2 and 0.4 are consider:d to he catremely good tits " Also indicative of goodne3s-of fit, when the overall
gg.
J. H. JoHwss Jn i -
model was applied to the individual cases to see how wellit could predict known responses, 71% were correctly reclassified mto the group to which they were a< signed initiany on the basis of their stated behasioral intentions in the event of an accident at SNPS.
)
l Dl5CUSSION AND islPl.lCATIONS One would expect a T.\\ll. type accident at the SNPS to precipitate the evac.;ation of only about 2700 families tauuming that only the pregnant women and the pre school children instructed to lease were to take such protective action). Iloweser, the bregoing empirical test of the model of evacuation decision making suggests strongly that, under
~
circumstances approximating those at TMI. nearly one third of the Long Island pcpulation
( - 289.000 households) would esacuate. As shown elsewhere, a substar,tial portien of the households who would evacuate reside in excess of 25 miles from the Shorehart plant."
The findings of this research and of previous empirical tests suggest first, that socialsurveys, similar to the one administered on Long Island should be conducted within a $0 mile.
radius of all commercial nuclear polver plants: and second, that application of the poposed model of evacuation decision. making to the resuhing data would hssist emergency planners in estimating the magnitude of spontaneous evacuation and assc> sing its likely i:. pact on the time it would take to esaeuate the population from the plume exposure pathway
(
zone.22
-4 t
l REFFRENCES 5
- 1. M. K. Goldhaber and J. E. Lehman. Crisis evacuation during the Three Mile Island nuclear :::ndent. A.
Meer. Am. Publ. HIth.lu. Wntreal. Quebec (1952t
- 2. C. B Flynn.Three Mile (Cand telephone survey preliminary report of procedures and findings. U 3. Nuclear
_?
Regulatory Commission. Waning.in. D C. (1979t K. Barnes. J. Drosius. S. Cutter and J. K. Mitchell, Responses of impacted popula:.an o the Three Mile Island nuel ar reactor accident: an initial asessment.
Discusion Paper No.13. Dep4:ttrent of Gengeaph). Rutger> Unisenity, New Brunswick. New Jney e1979;.
s 3
- 3. W. W. Chenault. G. D. Hilbt:' r.J S. D. ReKhlin. Evacuatien Planning in the TMI Accidte;. Federal 7
Emergency Management A m.). Washmgton. D C. il979L i
1 4 J. H Johnson Jr. Genert Ret ~5. aoS t 19%.
- 5. R. T. Ja>ke. Mu l. Lt. 24.14 N J L y
- 6. G. O. Rogers and J. Nehnewa, Behasior and.\\ttitudes Under Crisis Conditions: Selectee ;m.et and 4
Findings. Federal limeryere> Mar.2pement Agenes. W.nhington D.C. (1934L
-n
- 7. R. W. Perry, M. K. Lindell and M. R. Greene. Tracuation P.'annin;e in Emergency Macaeemer: l.cungton j
\\
Books. Lesington. Ma 3,n. hurt.*> < ! el t M. U.S Nuclear Regulatory C.immhan and Federal Emergen;> Management Aycncy.Cr:teria for Preparation
.md li.duation ei Ir @h y ' Er. un. 3 R.:9ar c"-
ad Preparo!neo in ';u pe : of St. m Pa n v
j/
Plants. Washington Dr. < !9p" g.
- 9. D J. Zeigler. S. D. Hrunn and J. H Jc hnwn Jr. Gcovel Rer. 71, I (1981).
-i
- 10. P. Slosic, B. Fi>ehho:T and 5..Ueh ens:ein. In Sovicial Rht. 4nenment: How Safe is S. lie En%th? # Edited by R. C. Schwing and W A A:Nr,t p.181. Plenum Pren. New York Il980L
- 11. E. Cook. Soc. Sd. Q. 63. 3 iNC
- 12. D. Rubin. Staff report to the prrud.nt'. commis> ion en the accident ut Three Mde bland. Rep. oF. e Public's -
Right to Information Task Fere L15. Gosernment Pnnting Ollice. Washin; ton. D.C. il979L
- 13. Social Data Analysts ine. Attr.u La Towards Esacuation. Reactions of Long l> land Re.idents 1:.1 Possible Accident at the Shoreham Nu.:e.c Power Plant. Setauket. New York tl982t 14 R. E. Kasperson and K. D Pi.Mt rahIic Admin. Rcr. 45. ' (1985L
.k 15 J. H. Sims and D. D. naumann 1.tironmcru & Itchar.15. If 5 t1983t F
16 J.11. Johnson Jr and D. J Ze:;ie-hen. Geof:r 59. 24 Il943t l
p-IT M. Fishbein und L A,iten Uc..,
.!:u.le. Iraention.md Heh trior: An Intro.hulon to Theary a c Roeorch.
-\\
1,'
Addison.Wesle). ReaJing. W..:xus (1975L
- 18. Wnkelosich. Sketley and u t:. Ir, Status Rep. On P@lic Responw to I mcrger.e> Plarn *.; Efrorts.
- ?,
Prepared for the 1.org lenJ W:..ng Conipany. I ong ( A. tad. New York il%3L i
19 D Knoke and P. J. Bark:. L. : l: at tiedeb. Nge. Bescr!> lhlis. Californ:.i (1980L J"
20 D. A. Henser and I. W. Jahr..
,;lini l>ix eue ( h"in WJcIline. Wde) New York il93li i
- 21. D. J. /eigler and J 11 J.~ran Jr 1%t Gcovr 36, 20' ilett "I
- 22. J. H. Johnson Jr and D J /ege. m m Econ Plami. Ai 20.165 (IWi+
. 4 q
m
d4fM&a$$$h@UCLEAF,I.R, EGULATOR,Y.COMMIS. SIONP WPM wWMUNITED STATESw pyg q
g
< pdWA"f' ~ yt gg k MNN MN#nM M
~M+h#N m.mdgy n a9%
q%
m c
c -
c
~
4 -
. n w
n k
$E$.w$ga w,w:ww >m@wux.E N,8 Nu y d qa M m, p n 4 % y m p p - m e m.; w w. m& p n g %u n w w +e g a n ;,x p A A ; g;m A
u4 g
wn n
99 yw m
ye m wdn
,a s w.mmyM 4mp A
QMC#@J4 E80 R A N D O M y F_0 R@MMJ o h nj M.Q11 g a_nsM%d@@QMy? WM@@@N N
%,,M F.a s @j- @FROMIM M+ 6,o-Ek M M M.9 9 q~ r@e F h fEmil W@
> %qlp%sm. va&%M+(mDo~ck.e?M~ Ju,lian a kWK T%hl6. yServ.1m.Branchp$B y@Q c'tingichfef Gw wn u
w n
/
w m c
eting~and we&m%m
- m w
e ce 4me
. ]l,$Jf%%a QU
'EXHIBITSM @~QQ Q-a M. R] m &pJ
&4 A
....W %T.yP 4
M &$iSUBJECT,Hu SHo A E HA" p 4 ap r
a m
m
@@g"f v3My > A ; N, Wm' ' %^ C V @ ' y %%%gNM 7J, qu@~h%gM M pf56W2dyW T
%y p%4 x g ; m.
aM~
4W fugf i M n%L 2%g%pg dtheV@p Q h W% U JNe e d i n d d n d Mid eta $a ftl$ fbt k e n tr e c o yg
- p% & %
4 MWH%
m'MDW p24
- ew y mL.!X y 5 % p;1Any"xdoc;umentsefiledfon A w QW in that SWR ENAM fpr0&rMi x
W o
%@%d Me x h i b i tM s% 6n s i d ef e dh x e mp t%
eT6fficia'i
%$3%
opisions%6fathe M ited % n h
?
D M, ~St'ateMCopyri a,m. m e %w~ithithescou ghtTAct dunless9. 6.N Roh jina;hts 11F.ifindGViher,ns.eaMa,: W,w W J
~
w
.r tl e x p+re s s l ~y d b e c a u s e vo~f4 c o py r;4,kn
.n,+.m ; p %-.t.".m%. n~+m
.w
,i g Weg m
mu w m{,4. m1 7% A 1 1To fi t h efd o c u m'e n t s? s e n tft'oTTE, f o W, M,::O_:,1 mms a
- ,t-n,
(
,4 4:p-gs.:
n, u.w..
?
v' er 4: 4. :i. v
+ n%t. -
7 y
m n:v ~q n i t ',
rs -
x y $7,
c,,,
a e.
n..,
~
n
.,;a
- m..
v..
g9
. "ro ces. s ih. g E_f. a l_.1Ni.thW_ths..%.
P M+U, M "g
s c
7 w " g" -
'exe.m,pst1Cl8SSificati.on N.._ _ _
?
[o -
1 m.
dWhN
' Mx
,d 3
s e a o a,- d' R@,, f?....g 1u s:s
+
m e' 1 s.
vm
' u
' s ;.
+
t ~N:
- { J,l0Q:;.5 I ti?,
2?'
4' U
- %.'j.
M. W WW
' p,' o
' f^( f,
m<
^ ' "
a
~s
.Q, 7 j';
' L W;,Y
- l'b.Y:l.Q l(p ' r
', } l 1
Y R 3:Q' Hl s
..a y:
_g~
g, my4 v, q y
a.u y y
+
y
.,. g:...
f.t. :.z. w( r
- ',y
- :
s;>5i,#.
k '1.m p#
- Q y i '
a
~ ~,. -
8 cs c
4>
i ' < < >
o r es I
.f :, -
M r.
I i.,p.
e.d
%)
?v U:>Q D:'. -l
?
s s
s s
, A,:.
g n,
g-
- y l, s ;
4',
S
(
1 x
.s.
k l'{f ;
b l
- w. y4
<W i
.j i : s i %y
4 3
,+
,,d rI
( [
T iN f.
-Q s p.
I h.-
i
' hl
' 5'f
..;y
? 3 p?, :1 A
...M r s :m
,4 %y. g' V
,L;g t
+ '
n..
.w e9 L,
.- 'tw t. - s s-x
- 3
- -
ce p
e.
I)-
'-T (
d:p,,, <
1' i
- w,.
a 5
f' L.
t 4
3' t
rt c.~,
n GK; 3 ( }jP
Q t((
N]
4 3
g y
h, s, Sp '
Q v
s, f
<e p
o.
1 P
,I
- h'1lx I
v.
",1 y.s...
'g.,
p A
i
.'c.,.a.-.~,
'y, e,.
i' d (
I t '
4
-')',
.,',,2 Y
.t t
j g.
s,. /.,
a s..:.
f 3'
s;
.. r "F
c4 M,
y A
~e e
i m, ' i
^
1
'y
)
O..; D h
f' i
4.. '
m[.
Lh.
i f,l
(
.<..a-q Bf a
.s Ny Yh ' b
'L h
%.3, a:f:L e
l r
r a
, (; +
b$ MGt y
s
... 3, NP:.
g t
't s
)
T' yi s
4 a
,a y-6 n
- q;
,n m 7 :b
.c' s
Q$;6s
- {
7
- y
.g, r
t y
i y.("f $
VN
,g
- n. ;' A.m :gr., w,
j fi:q "d
t z, --
tW De ' lq, y
i a,
a ll '
- 3 w
s l'&.,,
y T[
%h,.
s t
n a
~ g
.E E
~'
- a. :..
N r.Q./
N
- .e
<3 ;
A
.hp m,
4,
" u] t
."q,, &:,.
a c.
y J
- h. ag%[IMh. 7 5
M, w, g%c,.
1
,s
.,,M[
v w
p
, ; (/;} 9 9
hb/'h y
f (q 7 p
.q-ww'. w $
.c
.{,
l /
,I s
t 5,'I.,-f,
, '?
N ;, '-
{
}
\\
4-r&
i 7
+, s i
ibi 1
h g
4
}
- ?h (y -
e gj.[
a y, r.
?
r S y i
lt.j/,
/
f W.-
i r
.s
- 7, 1
'h
,,3 i
g.7
.u m 4@< jy ' <
W L
uM 8
Qj Q.,
4 N
&vg nd p.syas, u
,h w%-,
w+
- a, -
n a
, n 49-4..
3 4
-se s
Q @v g ge.V M Q
[.
. M<,
d,
' @# :a a$.% gg4'h - - x <,
- Mti,
...ms t
1 N/
hk"
'