ML20235D951

From kanterella
Jump to navigation Jump to search
Applicant Exhibit A-13,consisting of Article Entitled, Model of Evacuation-Decision Making in Nuclear Reactor Emergency, in the Geographical Review.Copyright Matl
ML20235D951
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-013, OL-3-A-13, NUDOCS 8709250418
Download: ML20235D951 (15)
v • d • e


Text

?

50

  • 322 l3 k

)

V&^ lA.[ [ }

7/g' /g 7 M

- ~ ~ -

g

/3

?DW s

A MODEL OF EVACUATION-DECISION MAh1NG IN

%P L

A NUCLEAR REACTOR EMERGENCY

  • 3 l

JAMES H. JOHNSON JR.

.g g g y gn AssinAci. A model for dectston maksng about tracuatwn durmg a radiological emer-gency ss tested by loglinear causal rnodehng and with data from a >urvey of households c m the victmty of the Shoreham nuc! car porcer station on Long Island. The model mcor.

tyrates mtention to evacuate as a surrogate for actual behartor and attributes of both locatson and local residents as predictors of cracuation. The reiutts strongly suggest that the modct can forecast magnitude and spatsal cxtent of spcntancous evacuatwrt for any commercsal nuclear mstallatwn, if a radiological emergency should occur LICENSEES of commercial nuclear power plants in the United States are required to devise, in conjunction with state and local governments, a

plans to ensure protective measures during a radiological emergency '

These plans must comply with prescribed federal guidelines, and final ap-p proval is the joint responsibility of the U.S. Nuclear Regulatory Commission I

(NRC)and the Federal Emergency Management Agency (FEMA).2 However, the criteria that are to be used in both the preparation and the evaluation of the plans are based almost solely on knowledge about human behavior in natural disasters and nonradiological technological emergencies.3 Studies conducted in the aftermath of the accident at the Three Mile Island nuclear-generating facility near Harrisburg, Pennsylvania, in March 1979 strongly implied that emergency plans, developed according to such criteria, are likely to be inadequate and ineffective for a major nuclear-reactor accident.

g During the Three Mile Island crisis, the advisory from the governor of d

t Pennsylvania instructed the evacuation of pregnant women and preschool-Q aged children within a five-mile radius of the malfunctioning reactor and everyone else within a ten-mile radius to stay indoors. This advisory should jy have resulted in the evacuation of approximately 3,500 persons. However, ym it was estimated that 144,000 persons, or 39 percent of the population within EB

%2 3

  • I wish to thank William A. V. Clark, J. Nicholas Entnkin, Bernard Logan, Meinn Oliver, Bruce W. Pigozzi, and Donald J Zeigler for their comments on dratts of this article.

' Nuclear Regulatory Commission Emergency Planning: Final Regulations, Federal Rerister, Vol. 45,

~i8 l

1080. p. 55402.

-M 2 Cntena for Preparation and Evaluation of Radiological Emergency Response Plans and Prepar.

O cr edness in support of Nuclear Power Plants (Washington, D.C.; U.S. Nuclear Re6ulatory Commio 3 3 sion and Federal Emergency Management Agency,1980).

w ::

' Robert T. Taske. Emergency Preparedness: Status and Outlook, Nuclear Jarcty, Vol. 24,1983,pp.

  • 2 l

1-11; James H. Johruon Jr., Planning for Spontaneous Evacuation during a Radiological Emergency, c: r*

Nuclear Saterv. Vol. 2s.1984, pp.186-194; Susan L. Cutter, Emergency Preparedness and Planning m$

for Nuclear Power Plants Accidents, Apphed Geograpny, Vol. 4.1984 pp. 235-245; and Richard T,

  • g Sylves, Nuclear Power Plants and Emergency Plannin6: An fhtergovernmental Nightmare, PuHic e3 o Admsmstrutwn Rmew, Vol. 44,1984, pp. 393-401.

Ot4 0 DR. johnson is an associate professor of geography at the University of California, Los Angeles, California 90024.

8709250418 870700 PDR ADOCK 05000322 g

PDR

i i

i 406 THE GEOGRAPHICAL REVIEW 2

I fifteen miles of the facility, departed.' The process whereby individuals and l

families left their homes without official directive was called spontaneous i

j evacuation.5 An analysis of responses to the accident from a geographical perspective used the term evacuation-shadow phenomenon to identify the tendency of people beyond or outside the designated dangerous zone to l

j i

l leave spontaneously or voluntarily.* This behavior had occurred in other 1

j crisis situations but never at the magnitude observed during the Three Mile Island incident. The gap between expected and actual evacuation was so i

i wide that it highlighted a need for improved methods to estimate the flow and extent of spontaneous evacuation movements.'

This article presents a causal model of evacuation-decision making dur-I l

ing a nuclear reactor emergency. The model is tested by a modified version of path analysis and with data obtained from interviews in a survey of a i

I representative sample of households in the vicinity of the Shoreham nuclear power station on Long Island, New York. The results of the test demonstrate i

that the model can be used to forecast for any commercial nuclear power l

plant both magnitude and geogra a radiological emergency occurs. phical extent of spontaneous evacuation THE MODEL The model is premised on the propositions that evacuation planning L

9

. be based on the behavioral intentions of the group that perceives itself i

tr

}

ik and that a social survey during the planning proce::s is the best way c

t itain the data needed to predict behavior during an actual emergency 8 j

i

, nce on intended behavior as a basis for evacuation planning is advo-I

c. T a mainly on theoretical grounds. The theory of reasoned action posits that intended behavior is the best indicator of actual behavior if the int tion does not change before an individual has the opportunity to act. Ap-plications of this theory show the proposition to be valid for a range of social be. miors, including voting on nuclear referenda, and it is especially/

applicab the specific research context here.' During the past three years, the nucle.

)ower industry has been plagued by problems of design con l!

f struction, and operation of commercial reactors. Surveys conducted near if l

'Cynthia B l'lynn. Three Mile Island Telephone survey: Preliminary Report of Procedures and lames H. Johnson Ir.. and Donald I. Zagier. Final Report un a so 8

Rendents. Michigan 5 tate Untverutv, Department of Geography. East Lansing.1979.

' William W. Chenault. Cary D. Hilbert, and Seth D. Rett.hlin. Evacuation Planning in the TMI Accident (W.uhington. D C.. Federal Emergency Management Agency,1979).

  • Don.sid J. Zeigter. Stantev D. Brunn. and lames H. Johnson Jr.. Evacuation from a Siuclear Tech notogical Daaster. Gruparhtart Re:ww. Vol.
  • Chenault. Hilbert. and Reichtin, footnote 5 above. p. 4.71.1981. pp.1-16. reference to p. 7
  • James 11. Johnson Jr. and Donald I. Zeigter. Distinguishing Human Responses to Rad Emergencies. honnanc Cropagnv. Vol. 59.1983. pp 38tr402. reference to p. 387.

l

' leek Aizen and Martin Fishbetn. Understanding AWtudes and Predicting social Behavior (E

]

wood Cliffs. N.J.. Prentice-Hall.1980) and Carole H.' Bowman and Ma rin Fishbein Understa i

i Public Reaction to Energy Proposals. An Application of the Fishbein Modellournalof Applied

)

Psgholory. Vol. 6.1974. pp. 319-340.

l i

f~

. e,.

. 7

. m.~ 3

..z

(..

9.

cv4Cuar:ON-DECISION MAKINC 407 commercial nuclear power plants in New York and California during that l

period indicated that behavioral intentions in the event of a nuclear reactor j

accident were relatively stable. Analyses of data from one survey, more-ovcr, showed that evacuation intentions closely paralleled behavior during the Three Mile Island crisis five years previously." On the bases of these and related considerations, intended behavior is judged to be a reliable in-dicator of actual behavior in a radiological emergency.

=

The model comprises six components: three exogenous categorical vari-ables (location, stage in life cycle, and weial status) and three endogenous 4

variables (attitudes toward nuclear power, perception of risk, and evacuation intention) (Fig.1). I now discuss the first set of variables as well as two endogenous ones that are hypothesized to have direct or indirect influence on evacuation during a nuclear reactor emergency.

There is no question that the location of a household's residence relative z

to a reactor will be a crucial, if not the most important, factor in the decision-making process. In general, persons residing near a plant will be more apt to evacuate than individuals farther away. Specifically, the propensity to evacuate will most likely decline with increased distance from the facility.

Evidence from the Three Mile Island crisis lends strong empirical support to this hypothesis. A survey there revealed that a high level of evacuation occurred within ten miles of the disabled reactor and that the propensity to

{

evacuate declined noticeably only with increased distance beyond that limit."

Multivariate analyses of data gathered on behalf of the NRC approximately three months after the Three Mile Island incident reveal evacuation rates of 66 percent within five miles, of 49 percent between six and ten miles, of 34 percent between eleven and fifteen miles, and of 13 percent between sixteen and twenty-five miles." These data strongly indicate that evacuaticn from a nuclear reactor accident will probably conform to the classic distance-decay phenomenon that has proved to characterize behavior in response to other kinds of environmental stress."

Locational considerations other than distance, including physical fea.

  • Stephen Cole. Testimony Regarding Contention 23 (evacuanon shadow phenomenon) on behalf of suffolk County in the Matter of Long Island Lighting Company, shoreham Nucicar Power Plant, Unit 1. Docket No 30-3*2-OL-3. Januarv 1984 (available at NRC Document Roomt and PLOG Research, Inc. Emergency Preparedness study: A Report of Research. Prepared for the san Luis l

Obispo County Otfice of Emergenev services (Les Angeles: PLOC Research,1984).

l

' Attituses toward Evacuation: Reactions of Long 1* land Residents to a Possible Accident at the shorenam Nuc' ear Power Plant (Setauket, N.Y.. social Data Analysts,1o820 and lames H. Iohnson

?

Jr. and Donald J. Zeigler, A spatial Analysis of Evacuanon Intentions at the shoreham Nuclear Power station, m Nuclear Power: Assessina and Managing Hazardous Technology (edited by Mar-un 1. Pasqualetti and K. David Pilawka: Boulder Colo.: Westview Press.19A4h p. 286.

l

" Brunn. Johnson, and Zeigler, footnote 4 above.

"Teh.Wei Hu and Kenneth s. slaysman. Health Related Economic Costs of the Three Mile Island I

Accident. Pennsylvania state University, Center for Research on Human Resources. Institute for Pouev Pesearch and Evaluation. University Park,1981.

i

" Milton E. Harvey, John W Franer. and Mincaugas Matulionis. Cognition of a Hazardous Envi-ronment: Reactions to Butta!o Airport Noise. Economic Gregrarhv. Vol. 55.1979. pp. 263-256: and I

Michael K. Lindell and Timothy C Earle. How Close is Close Enough: Public Perceptions of the Risks of Industrial Facilities. RisA Analysts. Vol. 3.19t 3. pp. 245-253.

t I

3

.i

t

' I 408 THE CEOCRAPHICAL REVIEW i

NuctzAn l !

M WER l

Artrnas 8

l LoCATON

\\ l MAEAR ACCCENT i

nm a

N l

STAGE l

}

g LFE j l CYCLE i

EVAGJATON

,I tanrtn SOCIAL I

$TATUS i !

Frc. l.-Model of evacuation-decision making fer a nuclear reactor emergency.

tures of the local environment, population distribution, and local landuse patterns, may also play a significant role in the decision-making process.L8 Evidence from the Three Mile Island incident revealed that the direction of a residence from the plant might be the " generic" attribute that captures the effects of "other" locational factors which vary significantly from one a

reactor site to the next. In terms of origin of evacuees, rates of departure were higher in communities to the north and west of the plant than from those o the south and east. The preferred destinations were in the moun-tains and the sparsely settled communities to the north and west of the

,[

reactor. That directional bias, at least with respect to selecting temporarv l

shelter, probably reflected a desire by the evacuees to have a physical barrier between themselves and the disabled reactor as well as to avoid the possi-i' l

bility of subsequent movement from more densely settled centers that were downwind from the plant. '

i From an analytical viewpoint, inclusion of both distance and direction as exogenous variables in tests of the model seems reasonable. These two variables adequately assess the etiect of the local geography of a reactor site on the probability that a household will depart in the event of a radiological emergency.

The stage in life cycle of a household may al>o directly influence the I

evacuation decision in a radiological emergency. A general finding in studies i

i I

lohnson and Ze:gler, footnote 11 above.

t

  • Flynn, footnote 4.ibove.

i

  1. Zeigler, Brunn and Johnson, footnote 6 above.

+

k i

i

j. g-v. :

,-.- s EVACUAr!ON-DECislON MAKING 409 of behavior during disasters is that older persons are less prone than youn-

?'

ger-aged individuals to move. A nuclear reactor accident is no different in f

thu regard. For example, a study of this factor at Three Mile Island found g,p that 45 percent of the sample between ages 20 and 29 evacuated, while only y^

14 percent of persons over age 50 did so.In another survey of that incident, 3 percent of the persons who did not evacuate gave "too old"as the primary reason for their immobility."

Other stage-in-life-cycle indicators may influence the evacuation deci-sion. During the Three Mile Island incident, households with a pregnant woman were more apt to relocate than ones with no pregnancy, and house-holds with a current or past marriage had higher probabilities for evacuation than did ones composed of a single person. However, the presence of chil-3 dren was the strongest determinant of evacuation among the variables as-sociated with stage in life cycle. Households with children were 23 percent more likely to evacuate than childless ones, a situation attributed to per-ceived intergenerationa! effects of exposure to radiation.28 On the basis of these findings it can be posited that young families in the childbearing years, especially if children are present, are more likely than older, childless house-holds to evacuate.

A third factor likely to influence directly the evacuation decision is the social status of the head of household. In almost all studies of behavior after I

the Three Mile Island incidentine indicator of social status, years of school-ing completed, consistently emerged as a statistically significant correlate or determinant of evacuation; other indicators of social status, for example, income and occupation, failed to achieve this importance.28 The more edu-cation that a head of household had, the more prone it was to evacuate. In spite of the division arr.ong the indicators,it seems reasonable to posit that households with high social status will be more likely to evacuate than ones 3

with low status. At the individual household level, the three traditional measures of social status will probably be strongly correlated, and income and occupation as well as educational level have been crucial determinants in voting behavior on nuclear referenda.22 The attitude of the head of household is likely to have a direct influence on the evacuation decision. This attitude is formed on the basis of beliefs about the consequences of generating electricity from nuclear power.23 A study based on a heterogeneous sample of persons from various parts of Austria revealed that beliefs about nuclear power fell into four categories:

{

1 Susan L Cutter and Kent Barnes, Evacuation Benavior and Three Mile Island. Disasters. Vol. 6 1982, pp.116-124.

]

Brunn. Johnson, and Zeigler, footnote 4 above.

3

  • Hu and slavsman, footnote 13 above.

I 8' Hu and Slavsman, footnote 13 above, p.15. and Cutter and Barnes, footnote 18 above, p.121.

d David J. Webber,la Nuclear Power Just Another Environmental Issue? An Analysis of California l

\\oters Intiroarment and Schavim. Vol. 14.1992. pp. 72-83.

8' Harry J. Otway, Risk Assessment and the Socal Response ta Nuckar Power, lournal of the British j

Nucles-Energy Society, Vol.16,19~7. pp. 327-333 j

0

i t

410 i

THE GEOGRAPHICAL REvlEW l

f, psychological risks, economic and technological benefits, sociopolitical risks, I

1 !

and physical-environmental risks? Opposition to or negative attitudes to-j' ward nuclear power were strongly related to the psychological risk category, j

while support or positive attitudes emphasized the economic and techno-logical benefits. Research on this attitude in the United Kingdom, The Neth-i criands, and the United States supports the findings of the Austrian study? -

l One study in the United States stated that concern about safety of a nuclear power plant was the strongest discriminator between opponents and pro-ponents of the use of nuclear power? On the basis of these findings,it can i

l be posited that opponents would be more inclined than proponents to evac-I uate in a radiological emergency.

j In actual application of the model, the attitude of the head of household

l toward the local nuclear power plant should be included as an endogenous i

variable along with the general attitude toward the use of this form of i l energy to generate electricity. On the surface, these two variables appear to.

j be highly correlated and therefore redundant in most instances. However, f'

an individual might support the use of nuclear power in general, yet might i

be opposed to the existence of a local plant.r For example, a person might consider a plant to be too close to his residence or to be inadequately man-j aged and hence a source of potentidly catastrophic accidents. A.CBS-New York Times nationwide poll, taken approximately one month after the inci-I dent at Three Mile Island, supported this viewpoint: 46 percent of the sam-pie generally favored the construction of additional nuclear power plants, i

but only 38 percent approved of such facilities in their own communities?

The converse is theoretically possible, but not likely in reality. Support for.

a local nuclear power facility may result from its contribution to the econ-l omy or a long term record of accident free operation, while opposition may stem from problems associated with siting, design, operation, and manage-ment of recently built plants. Therefore, it is plausible to include both atti-tudes in actual tests of the model of evacuation-decision making.

Finally the decision about evacuation may be directly intluenced by the f

l l

a Harry J. Otway. Damar' Maurer, and Kerry Thomas. Nuclear Power The Question of Public_

Acceptance. Futurr5. Vol.10.197& pp 100-118. reference en p.113.

" Enc sunderstrom and others Community.\\ttitudes toward a Proposed Nuclear Generanng Fa-l j cility as a Function of Expected Outcomes.'Iournal or Commumtu P39.hology. Vol. 3.1977. pp 1"-

20% Baroara D. Melber and others. Nuclear Power and the Public: Analysis of Collected survey l

Research isvattle:

Battelle Human Arf.tirs Researen CNnters.1977); J. Richard Einer and loop Vd" l

Der Phgt, Beliets and Valun in the Nuclear Debate, tournal et SThed sonal Psychclogy. Vol. 9.197E pp. 324-D6. Tae O. Woo and Carl H. Castore. Dpectancy-Value and Selective bposure as Deter-minants of Attttudes toward a Nuclear Power Plant. lournal of Syht) Social Psychcloyy. Vol.10.

1940.p 224-234; Energyp.

loop Van Der Pligt. Joop Van Der Linden. and Peter Ester. Atntudes to Nuclear i

Beliefs. Valun and False Consensus. fearnal ot Enwronmental Psychology, Vol. 2,1982. pP-

}

221 231; and Joop Van Der I'ligt, J. Richard Eher, and Russell spears. Public Attitudes to Nuclear Energy. Entryy Pohcv. Vol 12.1984, pp 303-305.

  • Melber and others, footnote 2s above, p. 2St.

" Frank J. Popper. LP/HC and LULUS; The Political Uses of Risk Analysis in Land Use Planning, hk Analvsis. Vol. 3.1983, pp. 255-263; and Lindell and Earle footnote 14 above.

" Robert C. Mitchell. The P

- esponse to Three Mile Island: Compilation of Public Opinion l

Data about Nuclear Energy,7 on Paper 0-3d, Resources for the Future, Washington, D.C.,1979-O r

e: ~ ::: y ;

ym EVACL'ATION-DECislON MhKING 411 4

I head of household's perception of risk from a nuclear reactor accident. Risk-j perception studies repeatedly have shown that the public views nuclear power with more dread than any other activity except terrorism and war-fare, in large part because of the potentially lethal, carcinogenic, and mu-tapnic effects of ionizing radiation," Unlike other agents of natural and technological disasters, radiation is often imperceptible to human senses, l

except in doses sufficient to ca tse radiation sickness. These distinctions, i

especially the ambiguous nature of the disaster agent, are probably very important in decision making about evacuation.' Research has shown that the ability to confirm official warnings through direct sensory evidence like winds associated with a hurricane is a crucial determinant in evacuation from natural disasters." However, in a nuclear reactor emergency there would be no similar environmental cues, Instead the public would have to rely solely on secondary and tertiary-the former from technical experts and governmental officials, the latter from news media, friends, and rela-tives-analyses and interpretations of events. Such sources may be seen as unreliable, biased, and contradictory, as was the case during the Three Mile Island crisis.22 Moreover, social surveys conducted in the vicinity of several American nuclear reactors indicated that a large portion of the population j

would have little or no confidence in official warnings or recommendations

{

for protective action, regardless of the source." This high level of distrust j

reflects, in great measure, the failure of the federal government to demon-strate that it can protect the public from the hazards of nuclear power."

The decision to evacuate during a nuclear reactor emergency, for the most part, will be made on the basis of an individual's perception of risk to

Paul slovic. Baruch fnchhoff, and Sarah Liechtenstein, Ratin5 the Risks, Enorronment, ' ol 21, 1979, pp.14-39; Paul slovic. Baruch Fischhoff, and sarah Lientenstein, Facts and Fears: Undentand.

ing Perceived Risk, m societal Risk Assessment: How safe is safe Enough?(edited by R. C. knwing and W. A. Albers, New York: Plenum Press,1930); Erte J. Johmon and Amos Tversky, Represen-tations of Perceptions of Risks. Tournal et E.wcriment.rl Psychology: Central. Vol 113,1984, pp. 35-70; and Ronald W. Perry, Population Evacuation in Volcanic Eruptions, Floods, and Nuclear Power Plant Accidents'some Elementary Comparisons, fournal of Commumty Psychology, Vol. 11,1983,pp.

47

' Donald J Zeigler James H. Johnson Jr., and stanley D. Brunn, Technological Hazards, Resource Pt.hotans in Gevgrapny (Was.tington, D C.: Association of 'Amencan Geographent 1983), pp, 91-83.

H Ronald W. Perry, Michael. Lindell, and Majorie R. Green, Evacuation Planning in Emergency Management (Lexington..Wssa Lexington Books,1981).

' Daniel M Rubin and eth ers. staff Report to the President's Commission on the Accident at Three Mile Island. Report of trie ' ublic's Right to information Task Force (Washington, D.C.: Covernment Printing Office,1979); Dan Nimmo and James E. Combs, Fantastes and Melodramas in Television Network News: The Case of Three Mile Island, Western fournal of speech Commumcatren, Vol 40, 1962, pp. 45-33: and The Publi(s Right to Know: Communicator's Response to the Kameny Com-mission Report (Washington, D.C.; Media institute,1980).

u status Report c,n Public Response to Emergency Planning Efforts (prepared for the Long Island Lighting Company by Yankelovich, skelley, and White. New York,1983), and Attitudes toward Evacuation, footnote 11 above.

"Christoph Hohenesmer, Roger E. Kasperson, and Robert Kates, The Distrust of Nuclear Power, Science, Vol 196.1 April 1977, pp. 25-34; Roger E. Kasperson and others Pub!!c Opposition to Nuclear Energy Retrospect and Prospect, scrcnce. Technology and Human Values, Vol s,1960, pp.

1143, and Othce of Tecnnology Assessment, Nuclear Power in an Age of Uncertainty (Washington, D.C.: Covernment Pnnting Omce,1984)

l t

412 THE GEOGRAPHICAL RcVIEW

l self and family? In view of the high degree of fear and distrust of nuclear
l power, of the low level of public confidence in individuals and organizations that have been assigned the responsibility to issue warnings and recom-3 I

i mendations for protective action, and of the absence of environmental cues,

'j the postulate for the model is that persons who perceive themselves to be in the path of a passing radioactive plume will evacuate, while individuals 3

who perceive no risk will not depart.

i :j In addition to the foregoing direct effects, physical location of a house-hold in relation to a plant, social status, and stage in life cycle are posited i

to influence decision making about evacuation indirectly through both at-l titudes toward nuclear power and parception of risk from a nuclear accident.

3 Young families of high social status with or without children and residing near a plant are more apt to oppose nuclear power and to perceive them-N selves at risk than older households of low social status that live dis j

from a reactor site. Another postulation is that the attitude of the head of' household has indirect influence through perception of risk. Individuals q

who oppose nuclear power are more likely to perceive risk than persons who favor its use.

i 1

In sum, the model of evacuation-decision making in a nuclear reactor emergency incorporates intention to depart as a surrogate for actual behav-ior and attributes of both location and local residents as predictors of evac-g 7

L uation. The model embodies thirty research hypotheses (Table I). The initial nine specify variables that are posited to have direct influence on the deci-sion. The rest pertain to the variables postulated to have indirect influence through perception of risk (hypotheses 10 to 17), through attitudes toward

(

)

a local nuclear power plant (hypotheses 18 to 24), and through attitudes 4,

toward use of nuclear energy to generate electricity (hypotheses 25 to 30).

TESTING THE MODEL l

Data used to test the model came from a social survey conducted by a private firm for the Long Island Lighting Company, the owner of the Shore-l, ham nuclear power station. This facility is located on the northern shore of j

i Long Island in Suffolk County approximately sixty miles east of New York i

j City. Designed to assess emergency-planning needs associated with the opening of this facility, the survey was administered to a stratified random sample of 2,630 households in Suffolk and adjoining Nassau counties during a two-week period in May 1983?

L i

For the purpmes of this study I selected ten variables from the survey l

(Table II). Consis ent with the struaare of the proposed model, the data are i

t 1

3

" E. Jonathan soderstrom and otners. Risk Perception in an Interest Group Context:An Exarnmation of the TMI Restart l$4ue. Ra An.stysa. Vol. 4. l984. pp. 231-244; and susan L Cutter. Residen Nuclear Power, footnote 11 vibove. pp 247-2s7. Proximity and Cognition of i

j

" Status Report. footnote 33 above, Appendia A. pp. 74-8s.

i

., - R.

e er= -

Qgg '

e -

EVACUATION-DECislON MAKING 413 l

divided into exogenous and endogenous variables. The former are the lo-cational and sociodemographic attributes of survey respondents, while the j

latter include attitudes towr.rd nuclear power generally and the Shoreham l ;j,,

facility specifically, perception -of risk during a reactor accident, and in-tended behavior during an emergency at the Shoreham plant. The final I g endogenous variable, evacuation intention, was derived from responses to j

an accident scenario, in which no one was advised to depart, because spon-taneous evacuation is the primary concern of this study. Responses to this scenario were redefined so that, under conditions approximating those in it, individuals who would continue normal activities or stay indoors were labeled potential nonevacuees, while persons who would leave their homes

)

were called potential spontaneous evacuees.

The model was tested by loglinear causal modeling, a modification of path analysis. In the loglinear variation, logit formulation solves the struc-tural equations corresponding to the hypothesized linkages in the model.2' Logit is a statistical technique, developed specifically to analyze relationships between dichotomous dependent variables and either categorical or interval

)

scaled independent variables. The proposed model reduces to four structural equations, one for each of the endogenous variables: general attitude toward nuclear power, attitude toward the Shoreham facility, perception of risk in j

a nuclear reactor accident, and intention about evacuation. In standard no-1 tational form, the equations are:

(x,xi,...,4)(w)

(1) i Ixi.xi,...,x H 4)

(2) fx, x2,..., 4 H x,)

(3) i lx, x:,..., x.} (xio)

(4) i I

where terms are defined in Table 11 and Figure 2.

l These four equations embody all of the hypothesized direct and indirect causal linkages in the model. The analytical objective here is to solve them in succession by logit to derive estimates of the direct and indirect effects of variables xi through 5, as measured by beta coefficients, on variable xio and to determine how well the overall model fits the observed data. Because the loglinear variation does not assign single values to causal paths where po-lytomous variables are involved, it was necessary to dichotomize the vari-ables in Table II with more than two categories of response: xi, x3, %, x,,

and w.)*

j The final path model for the Shoreham facility appears in Figure 2. My discussion is organized around the thirty research hypotheses that corre-I spond to the direct and indirect linkages embodied in the model. Eight of l

P David Knoke and Peter J. Burke. Loglinear Models. Sage Umt'ersitv Papers 20. sage Publications.

Beverly Hillo Cal.,1980.

l

" Knose and Burke. footnote 37 above, p. 42.

1 4

1 l

l

l i

I 1

l 414 THE CEOCRAPHICAL REVIEW i

j I

Taste I-RtsEARCH HYPOTHr.5f.S j

s a

l DEPENDENT VAalABLES dt DIRECTION OF PFLATIONSHIP j

Evacuation Attitude toward Attitude toward INDEPENDENT VAalA0LES intention Perception of local P ant MMClear POweT l

l dr HYPOTHI5L5*

(equation 4P nsk (equation 37 lequation 2P (equation lf

{

4 l

Distance (1) -

(10) -

(18) -

(25) -

i Direction (2) -

(II) -

(19) -

(26) -

i l

Age (3) -

(12) -

(20) -

(27) -

Pregnancy (4) +

(13) +

(21) +

(28) +

i Children (5) +

(14) +

(22) +

(29) +

Education (6) +

(15) +

(23) +

(30) +

General nuclear i

attitude (7) -

(16) -

(24) +

i Attitude toward

(

j local plant (B) -

(17) -

1 l ;

Perception of nsk (9) +

i source: Cotnpiled by author.

  • Hypotheses nutnbered in parentheses.

[

' Direct effect.

I

  • Indirect effect.

J the nine direct paths or links in the model are supported by survey data.

The exception is that the probability of evacuation does not appear to be

]

I y

directly linked with the presence of a pregnant woman in a household. Nor is this variable likely to have direct influence on the decision about evacu-ation through either perception of risk or general attitude toward nuclear l

power. However, this stage-in life-cycle indicator did indirect!y influence intentions to evacuate through very strong interaction with two other vari-1 a

ables, age of head of household and attitude toward the Shoreham plant.

6 The propensity of individual households on Long bland to evacuate in the event of a nuclear accident will most likely decrease with increased distance between home and the Shoreham plant. However, the beta coef6-cient of

,208 for this variable is an indication that the evacuation rate is likely to decline only modestly with increased distance from the site. These results are consistent with previous investigations showing that the pro-i pensity to evacuate would begin to decline at approximately ten miles, but i

not markedly until twenty-five miles from the Shoreham plant." Also the

! I perceived distance between home and plant is likely to affect the decision about evacuation indirectly through perceptions of risk and attitude toward the Shoreham plant, but the indirect effect of attitude is positive, not neg-s

{

i ative as hypothesized. This variation suggests that concerns about risk as-sociated with the plant remain high well beyond a five-mile radius from i

l the reactor site. The survey data do not support the postulated indirect effect j

of distance on intention to evacuate through attitude toward use of nuclear

)

power to generate electricity. This finding was not unexpected, because previous research had shown beliefs about nuclear power to be the principal 4

determinant of attitudes."

I j

" fohnson and Zeigler, footn~e 11 above.

  • Otway, Maurer, and Thor:
  • tootnoie 24 above.

9

' T. --

{

(

$WQ -),7 EVACllATION-DECISION MAKINc 415 a

TAntt Il-CONCF.*?5. VARIABLES. Axo MExst;REs CONCE!'ts & t. ARIA 6LES MLASL RES Location Perceived distance f )

distance from shoreham plant by categories ranging from 3 or A

fewer miles to more than 30 miles Dire: tion (n,)

east or west of Shoreham plant Stage in life cycle Age of household head (xp 18-24 yrs,25-34 yrs,35-49 yrs,50-64 yrs, or 65 or more years Pregnancy m household (s.) yes or no Young children (m.)

ves or no socioeconomic st.tus:

Iess than high school, high school graduate, some college, col-education (n.)

lege graduate, or postgraduate Psychological Geneat attitude toward continue building plants. stop building them, undecided nuclear power fu,)

Mou werd shoreham very pleased with prospect of new power source, moderately plant (a.)

pleased, me,derately concerned about risks, very concerned i

about risks Perception of risk (m.)

affected by risk from operation of shoreham plant: yes or no Behavior intention (x,,)

action if advi.sorv told people within 5 miles of plant to stay i

indoors: continue normal activity, stay indoors, or evacuate j

Source: Compiled by author from Status Report, text fooutote 33.

8

/

I Paralleling results of previous descriptive studies, the logistic regression analyses of the survey data indicate that the direction of a household's residerice from the plant is apt to have direct influence on the decision to evacuate. Persons who reside east of the Shoreham facility are less disposed to evacuate than ones who live west of it. This finding reflects what has been characterized elsewhere as the dilemma that an official advisory for l

evacuation would pose for anyone who resides east of the Shoreham plant."

On the one hand, movement eastward on the island from the plant might lead to entrapment if the situation worsened or persisted. On the other hand, g

evacustion westward would require passage through the hazardous zone, j

because. the island is only approximately ten miles wide and has only three routes of egress," Thus some individuals who live east of the p! ant do not l

perceive evacuation as a viable protective option, as the corresponding beta coefficient of.272 for the variable " direction" indicates. Nevertheless, un-i der conditions approximating the ones in the accident scenario, a substantial j

portion of the population residing east of the plant would attempt to evac-uate, in spite of the perceptual constraints.42 Survey data support the hypothesis that two stage-in-life-cycle indicators l

are likely to have direct influence on the decision to evacuate. On Long i

Island, individuals under age 35 and with children in the household are l

more likely to evacuate than older persons without children in the house-I hold. Additionally these two variablespave indirect influence on the deci-sion through the attitude toward use of nuclear power to generate electricity, i

I

Johnson, footnote 3 above.

I

" Attitudes toward Evacuation, footnote 11 above.

" Johnson and Zeigler, footnote 11 aoove; and Donald J. Zeigler and James H. Johnson Jr.. Evae.

untion Behavtor in Response to Nuclear Power Plant Accidents. Professional Geogreipher, vol. 36, 1954. pp. 20"-215.

E'

416 THE GEOGRAPHICAL REVIEW g

THE VAMAOLES AAE.

'X7' xi oaace "+)

I

,} f K2 Orecron (east)

(

\\

x3 Age ( < 35) f g; Xl N

,x4 arecene (vois

- l g

\\

ix5 cair'a tv's)

/

h

\\

!x6 Etcaten (

  • 17) l l

' s/

lX8

/

N 1x7==- *== tcoa)

Y

!x8 SNPS Altmne(con) e l

!i

)X21 f/

?

x*

SaPrcaoa(m) t x 10 Evacuanon eventen i es) d Nhxe t

g4

~

h' s [x 4/,

\\ a i,

N

'lX5l/j

q pD m

I j

f 1 i

' 6;!:::'-

l I Fic. 2.-Final path model for the $; oreham facuiry.

t Survey data also sustain the proposition that the decision to evacuate

~,

will be directly influenced by the social status of the head of household.

Logistic regression analyses of one indicator of social status, number of years t

of schooling completed, show that families in which the head of household i

has completed twelve or more years are more apt to evacuate than house-y[

holds where the head has fewer than twelve years of formal education. This indicator also indirectly influences intention to evacuate through perception of risk.

1 The postulated direct linkage between general attitude toward nuclear power and the probability of evacuating receives support from the survey data. Long I3 landers who feel that construction of commercial nuclear power plants should cease are more likely to evacuate than those who hold the l

opposite viewpoint. The attitude toward the Shoreham facility emerged as i

an even stronger determinant of intention to evacuate than the general nuclear attitude. Residents concerned about the risks associated with the Shoreham plant are more prone to depart than individuals who expressed satisfaction about having a new source of electric power. In addition to bein$

highly correlated, both variables indirectly influence intention to evacuate through perception of risk. Finally, the survey data support the posited i

linkage between perception of risk and intention to evacuate. Respondents who believe that they and other family members would be affected by a i

.. _.n

g. - -,
i%"O &

EVACUATION-oECislON MAKING 417 serious problem at the plant are more apt to evacuate than persons who believe that they would not be at risk after the opening of the plant.

In sum, the results of the logistic regression analyses of the survey data give strong empirical support for most of the hypothesized linkages in the model. But exactly how well does the model fit the data? In the loglinear version of path analysis, the goodness-of-fit statistic is p, which is analogous 2

to r in ordinary least squares regression; the p value for the model is.21, 2

2 which is indicative of a relatively good fit. There is another indicator: when the overall model was applied to individual cases to' determine how accu-rately it could predict known responses,78 percent were correctly reclassi-fied into the group to which they had been initially assigned on the basis of their stated behavioral intentions during an accident at the Shoreham facility.

CONC 1.UsION AND IMPLICATIONS This article offers a model of decision making and behavior that will enable emergency-management personnel to estimate both magnitude and extent of spontaneous evacuation in event of a major reactor accident. The model has six components and incorporates evacuation intent as a surrogate of actual behavior and attributes of location and local residents as determi-nants of evacuation. Analyses of survey data from Long Island residents revealed that the decision to evacuate would be strongly influenced by lo-cation of homes relative to the Shoreham plant, stage in life cycle and social status, attitudes toward nuclear power, and perception of risk during an accident at the plant. Typically the potential spontaneous evacuee resides west of the Shoreham plant, is young and well educated, has children in the household, opposes nuclear power generally as well as a local plant, and firmly believes that negative effects will result from operation of the plant.

Approximately 215,000 households, or 25 percent of the population, would depart spontaneously to create an evacuation shadow extending in excess of fifty miles beyond the site.

From a statistical standpoint, the reliability of the estimates of the mag-nitude and the geographical extent derived here remain unknown unless i

an accident of the type outlined in the scenario occurs at the Shoreham facility. However, the statistically proved good fit of the empirical data sug-1 I

gests strongly that thghouseholds responded seriously and hence that stated behavioral intentions would be acted out if such an emergency occurs.

Federal regulations require licensees to devise plans for evacuation within i

a ten-mile radius of a facility. The findings of this research indicate that l

social surveys, similar to the one administered on Long Island, should be conducted within a fifty mile radius of all commercial nuchar power plants I

and that the results should be evaluated in terms of the technical emergency-planning criteria established by the NRC and the FEMA. An outcome 'would be improved understanding of distance as a factor in perceiving risk and in t

i i

I-

418 THE GEOGRAPHICAL REv!EW

-l taking protective action. The data could also be used to assess the effects of i

l the evacuation-shadow phenomenon on the time required to move the pop-ulation in the ten mile zone where risk of exposure to radiation would be j

greatest. If the evacuation-shadow phenomenon is ignored, as in existent regulations, estimates of time needed for evacuation are probably grossly t

j understated and thus are an inadequate basis to argue that evacuation is a

! l feasible protective action during a major reactor incident. The approach outlined here, if adopted, simultaneously would deal with an important behavioral issue hitherto ignored by federal agencies and would institute realistic criteria for both development and evaluation of preparedness and response plans for radiological emergencies.

s h

8 i

l t

hI 3

i i

f I

Ii 1,

1 3

i i

h h

Y Yh

?

MGMM$dsMM $$

g@@@d Amp $p$WWMdE@MUNITEDSTATESMEMy@gMM OW%QMMA$

Q

@@7QWMAMEM0 RAN DUM UO R @@@TJ o h nlMill,i g a

$$MW MSN M@W$Fp ge hnassocHag{s$ ggpQ j

@ $$yGWJ@@ ROM:dA '$dkGWNM;c,%$NM@lian.a

  1. WM R#dh Md@W$% @

$w W

WM MQM s q

SMMMEmileGWJu tin schieftpy d h Mp*M 0

i l

MMa MMMR h SUBJECT S Oi SHORENAMEXHIBITSMSMS

@w&y$$e%, g&s:agg g%q$g&mme %)WMb W MbM W}gggg g g g$

ggp9 ; aqqpap Mw MMM WnWWW Y

MRW W

qom xe mo m wgp Wra uwg MAM - s um R" prom %g g

W L 1 Anyvdocumentsaf.MedionithesopentrecorddnKthe SNoRem wNym m

g 2, ff dgeeelingyandfmiidela7plagoffthe 0fficia1 ahi'r@

ecordya$$hMh6 Wp.

M e x hi b,i t%i_stc. on s i d e re,d t e x,emp t d{r.o;me SW n

4 W%

" M W M, awithithescourt{eipresslyibec

.u n

c W % 05tatesiCopyright~kA.c tVu'n l e s s s i tWa s g o ri g i.n a l ly fil e'd su-WW n, a ng. myg. w%g.gy4Qx,EOM M b

3 p y;ym.,m;.7 g

ph agg

~a

-, w ym m-y fy q,

.gs

,g 4

4p W 4 7 M W Mq 4

s.

f sexempt?c1'assif,ication,M ' ? ~ l" '

R1 W M & P W d<g'.4 y Mi 3 % 3,( s n/. A t >w,W G wg mm a

w

.m,

~,

i,

l.t h h,Ffk j M

'N1 P ' O nq J ug #

  • O:

lJ

}--

x J.

ppm y,.a r

a

~ y mw$ y$ 1',-

~

f c.%c e

1~

e e

  • p'

@c..

t i

nar g ' [p' ; '

o Q-Q'!w;p i f (

s

.(

/

')

2 q

4 t

yy_

+

di;, :. g

'M w., y f;j

, ;c;p. pJ

+

t mi w-y ?.? e,9 kv jf

~;

I, C

15. Tj 5

.::;;r Q

y*.

s

@"43 y

q.ltv;&, -,

.t?

a 4

4 w.:, n m.,

4

j-

'4 Ip7 J :p t

e. 2 -

3 '.,.i S M i ;... g

r.,' i % :j

.-" a y N's. [- ! L ;s s

4 j

L l

m fy 7'

MK}* l ', ' l', &( ' RQ:) :% h s ', L ;

't

.{ {

l} t,

-f; Y (yf s k.[lf

\\

c i

\\

y

. !?

a.',,' ' m :-y

m,,,%,

ggL f:1 3 Q_

s.b

p m, c ry y,,

e, jg

.x Q,..

,s.

i 4,

p u 3 p. m..

3 o

e

,g 4

,1

)

f;

}(_

dI

'a W' ggi meyq p

6 5

t 4e.

i3 i

1'

+ i

'3

,i % !$

's

@ Q, M}

Ms>

) ;,

?

g-

' p

.- - t a b2 ',

..[

i (4

)

='

q-1

%, b ;.a.

.1 n

i-,

v

[,

/

t 4

t

,,,.. 3

/

=.q,s p,

q" j

1 5

r s

5t,,

.; h,

(

l

,. p.r p

, e b

I I

5 i

i,

]

  • ' T ~

~ p, pq t

_(

x e

{Q w) 3:

c

',y g &l '.

~;rlM&

$b p

i

';Ty g) )

1 1

u._

a

g:
,y, n

gc

, r,y.

9 l

t.

a f' f

t~

m',

..,y l?:',

KQ' ':l l y, L

Q[Q'

\\

.,p t, Q

N]f s

_c lO T Ylgi.

~.

h>,-

[

, ' ' ', h[,

q e,&

>a s.

v~^

L

i, I j.

';3,

e.

i n

/

y

,.4

....g-,s.

I 4

.p:'

I

(

t

)

,h~ 4 9

'.,.,ia[..-

a r

., s

< (:-

,.y, uh^k

)

,. ~

r

- (p j

6*M[

.s x

sy f

c.

$'i fJf

,' 3_g,,

,)

pj

',r

'e 6.I;'

j f

'?#k }, _,

A v

g 4 ->) r..d!%

C ' p? ' '

G 0.); p i

J-l', t I W.

,,[Q y l,:Q!a g ~)f r

g,y 3,

t wa e -. 4 -j c y~y

.. ~g d,

=

e

+

A 3

A' m.

~ - -

.1 m

3 s

5

'Of j y-y

_j jf f

,,,<4 d(

h

?

7 1

- /

0 I.'

L,

y A..

..t's S

h j

p f

i e

j s

s

~

u i

4

?

1 8

S',

>5 suvg w{

q 1

i_ Na..i

,l.

I

,f g

I k

.l.,.:,-

I I

,k.

j g

'i 1

")

.. h

, ) M J :,1, M

w s:W3 I 2 bb 2 j~j NU[x +

n (~.w,w.w u' <w%

hl

.i$

>Ap 3

m,{.mgiu.

4

>(

-d anem u, < y

..I gh m u,,,

m e i d.

I,.

~2 j,1s <

7 wgeg yn i

, w

,a; ww g,g#

e peteWG awT p

M M. -

%yt'-1

_T i,

i

..'-T

.s J

f 2

5 6

a

, i,

a.

... ij.

[~.( W's, YbN h* IN

'h.)

ihb

.k Ui,- ~ ~ [h hb kb M~f

)

r an

, ~

.

Retrieved from "https://kanterella.com/w/index.php?title=ML20235D951&oldid=3952843"