ArticlePDF Available

The Protective Action Decision Model: Theoretical Modifications and Additional Evidence


Abstract and Figures

The Protective Action Decision Model (PADM) is a multistage model that is based on findings from research on people's responses to environmental hazards and disasters. The PADM integrates the processing of information derived from social and environmental cues with messages that social sources transmit through communication channels to those at risk. The PADM identifies three critical predecision processes (reception, attention, and comprehension of warnings or exposure, attention, and interpretation of environmental/social cues)--that precede all further processing. The revised model identifies three core perceptions--threat perceptions, protective action perceptions, and stakeholder perceptions--that form the basis for decisions about how to respond to an imminent or long-term threat. The outcome of the protective action decision-making process, together with situational facilitators and impediments, produces a behavioral response. In addition to describing the revised model and the research on which it is based, this article describes three applications (development of risk communication programs, evacuation modeling, and adoption of long-term hazard adjustments) and identifies some of the research needed to address unresolved issues.
Content may be subject to copyright.
Risk Analysis, Vol. 32, No. 4, 2012 DOI: 10.1111/j.1539-6924.2011.01647.x
The Protective Action Decision Model: Theoretical
Modifications and Additional Evidence
Michael K. Lindell1,and Ronald W. Perry2
The Protective Action Decision Model (PADM) is a multistage model that is based on find-
ings from research on people’s responses to environmental hazards and disasters. The PADM
integrates the processing of information derived from social and environmental cues with
messages that social sources transmit through communication channels to those at risk. The
PADM identifies three critical predecision processes (reception, attention, and comprehen-
sion of warnings or exposure, attention, and interpretation of environmental/social cues)—
that precede all further processing. The revised model identifies three core perceptions—
threat perceptions, protective action perceptions, and stakeholder perceptions—that form the
basis for decisions about how to respond to an imminent or long-term threat. The outcome
of the protective action decision-making process, together with situational facilitators and
impediments, produces a behavioral response. In addition to describing the revised model
and the research on which it is based, this article describes three applications (development
of risk communication programs, evacuation modeling, and adoption of long-term hazard
adjustments) and identifies some of the research needed to address unresolved issues.
KEY WORDS: Protective action decisions; protective action perception; risk perception
Researchers have long been interested in ex-
plaining the process by which people respond to
environmental cues or socially transmitted warn-
ings about environmental hazards and disasters.(110)
Findings from this research on environmental haz-
ards and disasters are compatible with those from
theories on social influence, persuasion, behav-
ioral decision making, attitude-behavior relation-
ships, protective action, and innovation processes in
identifying useful guidance on ways in which risk
1Texas A&M University, Hazard Reduction & Recovery Center,
College Station, TX, USA.
2Arizona State University, School of Public Affairs, Phoenix AZ,
Address correspondence to Michael K. Lindell, Texas A&M Uni-
versity, Hazard Reduction & Recovery Center, College Station,
TX 77843-3137, USA;
communication can influence immediate disaster re-
sponse and long-term hazard adjustments. The latter
are defined, following Burton, Kates, and White,(11)
as “those actions that intentionally or unintentionally
reduce risk from extreme events in the natural en-
vironment” (Ref. 12 p. 328). The relevant elements
of these complementary approaches have been in-
tegrated to produce a Protective Action Decision
Model (PADM) of the factors that influence indi-
viduals’ adoption of protective actions.(5,13) More re-
cently, additional evidence has been collected that
supports some of the PADM’s propositions and
has required reconsideration of others. Thus, the
purpose of this article is to describe an updated ver-
sion of the PADM, describe three applications (de-
velopment of risk communication programs, evacu-
ation modeling, and adoption of long-term hazard
adjustments), and identify some of the research
needed to address unresolved issues. This section will
616 0272-4332/12/0100-0616$22.00/1 C
2011 Society for Risk Analysis
The Protective Action Decision Model 617
Fig. 1. Information flow in the PADM.
Source: Adapted from Lindell & Perry (2004).
provide a brief overview of the model, which will be
followed by more detailed discussion of its compo-
nents in Sections 2–5. Section 6 will provide a theo-
retical assessment of the PADM and Section 7 will
describe its applications in three areas (development
of risk communication programs, evacuation model-
ing, and adoption of long-term hazard adjustments).
The findings of studies on individual response
to environmental hazards and disasters can be dia-
grammed in a flow chart that provides a graphic rep-
resentation of the model (see Fig. 1). The process of
protective action decision making begins with envi-
ronmental cues, social cues, and warnings. Environ-
mental cues are sights, smells, or sounds that sig-
nal the onset of a threat whereas social cues arise
from observations of others’ behavior. Warnings are
messages that are transmitted from a source via a
channel to a receiver, resulting in effects that de-
pend on receivers’ characteristics. The relevant ef-
fects are changes in receivers’ beliefs and behaviors,
whereas receivers’ characteristics include their phys-
ical (e.g., strength), psychomotor (e.g., vision and
hearing), and cognitive (e.g., primary and secondary
languages as well as their mental models/schemas)
abilities as well as their economic (money and ve-
hicles) and social (friends, relatives, neighbors, and
co-workers) resources.
Environmental cues, social cues, and socially
transmitted warnings initiate a series of predecisional
processes that, in turn, elicit core perceptions of the
environmental threat, alternative protective actions,
and relevant stakeholders. These perceptions pro-
vide the basis for protective action decision making,
the outcome of which combines with situational facil-
itators and impediments to produce a behavioral re-
sponse. In general, the response can be characterized
as information search, protective response (problem-
focused coping), or emotion focused coping. In many
cases, there is a feedback loop as additional environ-
mental or social cues are observed or warnings are
received. The dominant tendency is for such informa-
tion to prompt protective action decision making, but
information seeking occurs when there is uncertainty
at a given stage in the protective action decision-
making process. Once the uncertainty is resolved,
processing proceeds to the next stage in the process.
The stages in the PADM characterize the way
people “typically” make decisions about adopting
actions to protect against environmental hazards.
These stages are sequential, as are those within the
information-seeking process. However, few people
are likely to follow every step in the model in the
exact sequence listed in Fig. 1. For example, an ex-
tremely credible (or powerful) source might obtain
immediate and unquestioning compliance with a di-
rective to evacuate an area at risk—even if there
were no explanation why evacuation was necessary
or what alternative protective actions were feasi-
ble.(14) The important lesson is that—unless warn-
ing sources have an extreme amount of credibility or
they have substantial power to compel compliance—
the more stages in the PADM they neglect, the more
618 Lindell and Perry
ambiguity there is likely to be for message recipients.
In turn, greater ambiguity is likely to cause warning
recipients to spend more time in seeking and pro-
cessing information rather than preparing for and im-
plementing protective action.(9,15) Indeed, ambiguity
can initiate a repetitive cycle of information process-
ing and information seeking that persists until it is
too late to complete a protective action before haz-
ard onset.
The physical environmental component com-
prises the geophysical, meteorological, hydrological,
or technological processes that generate a hazard and
transport it to the locations where people are ex-
posed (see Ref. 16, chapter 5, for further discussion).
These processes generate hazards that vary in the
speed of onset, and magnitude, scope, and duration
of their impacts. One important hazard character-
istic that is often quite relevant for people’s emer-
gency response is the availability of environmental
cues such as sights and sounds that indicate hazard
onset. In tornadoes, for example, the sight of a funnel
cloud and a roar “like a freight train” (a common de-
scription by tornado victims) provide unmistakable
indications of imminent threat. In other cases, such
as the recession of coastal waters before a tsunami,
the cues are much more ambiguous and likely to be
misinterpreted. Still other hazards, such as ionizing
radiation and some toxic chemicals, provide no envi-
ronmental cues whatsoever.
The transmission of social information is based
upon the classic six-component communication
model of source-channel-message-receiver-effect-
feedback.(1721) Sources in the social context consist
of other people who may transmit information about
hazards and protective actions, as well as providing
assistance to reduce the hazard or providing material
resources that assist protective response. For exam-
ple, authorities, news media, and peers (i.e., friends,
relatives, neighbors, and co-workers) can provide in-
formation about environmental threats and alterna-
tive protective actions. Moreover, emergency man-
agers can dispatch city and school buses to evacuate
those who lack their own means of transportation.
They can also provide public shelters to those who
lack the money needed to stay in commercial facili-
ties or lack nearby friends or relatives with whom to
One crucial aspect of the social context is the
network of organizations and individuals that com-
prise the warning network.(6,22) An original source
can transmit a message by means of a broadcast pro-
cess directly to ultimate receivers (e.g., households)
and also by means of a diffusion process through in-
termediate sources who, in turn, relay messages to
ultimate receivers.(23) These ultimate receivers might
also transmit messages to each other, thus resulting
in some people receiving multiple warnings, others
receiving only a single warning, and some people re-
ceiving no warnings. The nature of the warning net-
work has a significant impact on protective action de-
cision making because multiple sources often deliver
conflicting messages that require searching for addi-
tional information to resolve the confusion.
The warning network can communicate informa-
tion through a variety of different types of channels.
These include print (newspapers, magazines, and
brochures), electronic (commercial radio and televi-
sion, telephone, route alert (broadcast from a moving
vehicle), tone alert radio, siren, and Internet), and
face-to-face (dyadic conversation or group presenta-
tion). These channels differ in characteristics such as
dissemination rate and precision, penetration of nor-
mal activities, message specificity/distortion, sender
and receiver requirements for specialized equipment,
and feedback/receipt verification (see Ref. 5, chapter
4, for a detailed discussion). Each channel has advan-
tages and disadvantages, with channels that provide
the fastest dissemination often providing the least in-
formation (e.g., mechanical sirens). Moreover, peo-
ple differ in their channel access and preferences.
For example, tornado warnings broadcast over an
English-language radio station missed the population
of Saragosa, Texas that routinely listened to Spanish-
language stations.(24) In addition to the official warn-
ing systems, however, peers typically relay informa-
tion through informal warning systems. Even when
peers do not explicitly transmit warning messages,
their behavior—especially obvious preparations for
evacuation—can serve as social cues for protective
Psychological processes are defined by three sets
of activities—(i) predecisional processes; (ii) core
perceptions of the environmental threat, alterna-
tive protective actions, and social stakeholders; and
(iii) protective action decision making. The three
predecisional processes of exposure (whether people
The Protective Action Decision Model 619
receive information), attention (whether they heed
it), and comprehension (whether they understand
it) are largely automatic processes that take place
outside of conscious processing.(26) The three core
perceptual objects—environmental threats, alterna-
tive protective actions, and societal stakeholders—
can elicit either automatic or reflective judgments,
depending on the degree to which an individual
has schemas that provide readily accessible and co-
herent beliefs about those objects. When some-
one has a schema—a generic knowledge struc-
ture defined by instances, attributes that differen-
tiate these instances, and interrelationships among
the attributes—beliefs about objects encompassed by
that schema are rapidly accessed to produce an over-
all judgment that is congruent with the available in-
formation about the situation. Finally, contrary to
widespread belief, panic rarely occurs.(27) Instead,
protective action decision making is often a reflective
process that assesses the available information about
the threat, alternative protective actions, and social
stakeholders to choose a behavioral response. The
research literature suggests that inappropriate disas-
ter responses are more frequently due to inadequate
information than to defective cognitive processing.
3.1. Predecisional Processes
Regardless of whether information comes from
environmental cues or social warnings, three pre-
decisional processes are necessary to produce a
protective response. Information from the physical
environment will not lead to the initiation of appro-
priate protective actions unless people are exposed
to, heed, and accurately interpret the environmen-
tal cues. Similarly, information from the social envi-
ronment will not lead to the initiation of appropri-
ate protective actions unless people receive, heed,
and comprehend the socially transmitted informa-
tion. Whether or not people heed the available infor-
mation is determined by their expectations, compet-
ing attention demands, and the intrusiveness of the
information.(26) These processes are known to be af-
fected by the warning recipient’s age,(28) but the ef-
fects of other demographic characteristics have not
been reported.
In the case of warnings, people must first receive
information from another person through a warn-
ing channel and attend to this information. Accord-
ingly, the characteristics of the warning channel itself
can have a significant impact on people’s reception
and attention to warning message content. For exam-
ple, in many places along the Oregon coast, moun-
tains prevent people from receiving signals from
National Oceanographic and Atmospheric Adminis-
tration Weather Radio transmitters.(29) Even when a
warning has been received and heeded, some people
will fail to comprehend the available information—
what others called “hearing and understanding.”(30)
The comprehension of warning messages depends
upon whether the message is conveyed in words
they understand. As the Saragossa tornado indicated,
warnings disseminated in English are unlikely to
be understood by those who understand only Span-
ish.(24) In addition, however, comprehension also af-
fected by more subtle factors. A warning message
cannot be comprehended if it uses esoteric terms
that have no meaning for those at risk. For example,
phrases such as “hypocenter” (earthquake), “Saffir-
Simpson Category” (hurricane), “oxidizer” (chem-
ical), and “millirem” (radiological) are specialized
terms that will not be understood by all who hear
3.2. Perceptions of Threats, Protective Actions,
and Stakeholders
Unlike comprehension, which provides the lit-
eral meaning of the words in a warning message, per-
ceptions of the threat, alternative protective actions,
and stakeholders involve broader associations that
are integral parts of the schemas or mental models(31)
within which the threat, protective actions, and stake-
holders are embedded. It should be quite obvious
that people will differ from each other in the com-
prehensiveness of their schemas about these objects.
That is, some people will have highly differentiated
schemas whereas others have poorly differentiated
schemas about an object.
3.2.1. Perceptions of Environmental Threats
The essential attributes of people’s perceptions
of environmental threats are generally considered
to be probability and consequences, but some well-
known approaches to perceived risk include factors
such as dread and unknown risks.(32) These addi-
tional dimensions of perceived risk appear to be
useful in explaining people’s responses to a broad
range of technologies and societal activities, but it
is not completely obvious how they are related to
the perception of, and response to, imminent threats
from environmental hazards.(33) Research on haz-
ards and disasters has emphasized a definition of
620 Lindell and Perry
perceived risk in terms of people’s expectations of
the personal impacts from an extreme environmen-
tal event.(7,8) These expected personal impacts in-
clude death, injury, property damage, and disruption
to daily activities such as work, school, and shop-
ping.(34) Moreover, most research on hazards and dis-
asters has found that risk perception predicts warning
responses such as evacuation(35) and long-term haz-
ard adjustments.(36) These protective responses have
been studied for hazards such as earthquakes,(37)
hurricanes and other coastal storms,(38) floods,(39)
and volcanic eruptions.(40) However, there have been
some studies in which risk perception was unrelated
to hazard adjustment.(34,4042)
In addition to the certainty and severity of the ex-
pected personal impacts, environmental threats can
differ in their degree of intrusiveness, which is the
frequency of “thoughts generated by the distinc-
tive hazard-relevant associations that people have
with everyday events, informal hazard-relevant dis-
cussions with peers, and hazard-relevant information
received passively from the media” (Ref. 13, p. 125).
Hazard intrusiveness is correlated with the adoption
of earthquake hazard adjustments(34,41) and expecta-
tions of participating in hurricane mitigation incen-
tive programs.(43)
Expected personal impacts and hazard intru-
siveness are related to the recency, frequency, and
intensity of people’s personal experience with haz-
ard events.(34,4345) Such experience can involve ca-
sualties or damage experienced by the respondent
him/herself, by members of the immediate or ex-
tended family, or by friends, neighbors, or co-
workers.(34) In turn, hazard experience is often corre-
lated with proximity to earthquake,(46) hurricane,(47)
and flood(48) sources. In addition to the indirect effect
of hazard proximity on risk perception (via hazard
experience), there can also be a direct relation be-
tween hazard proximity and perceived personal risk
that is determined by a perceived risk gradient relat-
ing increasing proximity to increased risk.(49) How-
ever, the resulting risk judgments can be quite inac-
curate because there are cases in which people have
limited ability to identify their location in risk ar-
Information from environmental cues and social
warnings, together with prior beliefs about the haz-
ard agent, produces a situational perception of per-
sonal risk that is characterized by beliefs about the
ways in which environmental conditions will produce
specific personal impacts. In hurricanes, for example,
risk perceptions have been characterized by people’s
beliefs about the degree to which storm surge, inland
flooding, and storm wind will cause their death or
injury, kill or injure their loved ones, destroy their
property, or disrupt their jobs or basic services such
as electric power and water.(25,38)
3.2.2. Perceptions of Hazard Adjustments
A substantial amount of research has focused on
people’s perceptions of natural hazards but there is
also a need for studying their perceptions of natu-
ral hazard adjustments.(52) Such studies are needed
because the theory of reasoned action (TRA) posits
that one’s attitude toward an object (e.g., seismic haz-
ard) is less predictive of behavior than one’s attitude
toward an act (seismic hazard adjustments) relevant
to that object.(53) Thus, to understand the adoption of
hazard adjustments, it is just as important to under-
stand the perceived attributes of the hazard adjust-
ments as the perceived attributes of the hazard itself.
The identity of these perceived attributes can
be surmised from studies on the adoption of haz-
ard adjustments, which have found support for at-
tributes such as effectiveness,(54) cost,(55) required
knowledge,(56) and utility for other purposes.(57)
Lindell and Perry(5,13,52) summarized this line of re-
search by proposing that hazard adjustments can be
defined by hazard-related and resource-related at-
tributes, both of which differentiate among hazard
adjustments.(58) Hazard-related attributes, such as ef-
ficacy in protecting people and property and use-
fulness for other purposes, have been found to be
significantly correlated with adoption intention
and actual adjustment.(41,59,60) Resource-related at-
tributes (cost, knowledge and skill requirements,
time requirements, effort requirements, and required
cooperation with others) generally have the pre-
dicted negative correlations with both adoption in-
tention and actual adjustment, but these have been
small and nonsignificant in studies conducted to date.
The hazard adjustments in these studies have gener-
ally had small resource requirements, so it is unclear
if the lack of support for the significance of these at-
tributes is due to this factor.(58)
3.2.3. Perceptions of Social Stakeholders
Previous research has characterized stakehold-
ers as authorities (federal, state, and local gov-
ernment), evaluators (scientists, medical profession-
als, universities), watchdogs (news media, citizens’,
and environmental groups), industry/employers, and
households.(2,61,62) The interrelationships among
stakeholders can be defined by their power over
The Protective Action Decision Model 621
each other’s decisions to adopt hazard adjustments.
French and Raven posited that power relationships
can be defined in terms of six bases—reward, co-
ercive, expert, information, referent, and legitimate
power.(63,64) Reward and coercive power are the
principal bases of regulatory approaches, but these
require continuing surveillance to ensure rewards
are received only for compliance and that punish-
ment will inevitably follow noncompliance.(65) Such
surveillance is rarely feasible, so authorities need to
rely on bases of power other than reward and co-
ercion. French and Raven’s conceptions of expert
(i.e., understanding of cause and effect relationships
in the environment) and information (i.e., knowl-
edge about states of the environment) power sug-
gest assessing perceptions of stakeholders’ hazard ex-
pertise. French and Raven’s conception of referent
power is defined by a person’s sense of shared iden-
tity with another,(66) which is related to that per-
son’s trustworthiness. Although trust has been de-
fined many different ways,(67) fairness, unbiasedness,
willingness to tell the whole story, and accuracy are
French and Raven defined legitimate power by
the rights and responsibilities associated with each
role in a social network, which raises questions about
what households consider to be the responsibility
of different stakeholders for protecting them from
seismic hazard.(63) This is reinforced by research
on stakeholders’ perceived protection responsibility,
which dates from research that attributed low rates
of seismic adjustment adoption to respondents’ be-
liefs that the federal government was the stakeholder
most responsible for coping with earthquakes.(69)
Much later, respondents had come to believe earth-
quake preparedness was an individual’s responsibil-
ity.(70) The conclusion that a perception of personal
protection responsibility leads to a higher level of
seismic adjustment adoption is supported by similar
findings on tornado adjustment adoption.(71) Recent
research has found that stakeholders differ signifi-
cantly in these perceived characteristics (expertise,
trustworthiness, and protection responsibility) and,
moreover, these characteristics have significant pos-
itive correlations with hazard adjustment intentions
and actual adjustment adoption.(41,67)
3.3. Protective Action Decision Making
Once the three predecisional processes have
been completed and the three types of core percep-
tions have been activated, cognitive processing turns
to the series of decision stages—risk identification,
risk assessment, protective action search, protective
action assessment, and protective action implemen-
tation. In addition, information-seeking activities in-
clude information needs assessment, communication
action assessment, and communication action imple-
mentation. Each of the decision stages in the PADM
is discussed in detail below.
3.3.1. Risk Identification
According to the PADM, people’s threat percep-
tions lead to risk identification, which is equivalent to
what Lazarus and Folkman call primary appraisal.(72)
In both emergency response and long-term hazard
adjustment, those at risk must answer the basic ques-
tion of risk identification: “Is there a real threat that I
need to pay attention to?”(4,73,74) The importance of
the resulting threat belief is supported by research
showing individuals routinely try to maintain their
definition of the environment as “normal” in the
face of evidence that it is not.(2) Researchers have
found a positive relationship between level of threat
belief and disaster response across a wide range of
disaster agents, including floods,(9,73) volcanic erup-
tions,(15,75) hazardous materials emergencies,(5) hur-
ricanes,(25) earthquakes,(76) and nuclear power plant
3.3.2. Risk Assessment
The next step, risk assessment, refers to the pro-
cess of determining expected personal impacts that
a disaster could cause.(8,74) The process of assessing
personal relevance has also been recognized as an
important factor by persuasion theorists.(66) In the
risk assessment stage, a positive response to the ques-
tion “Do I need to take protective action?” elicits
protection motivation whether the risk involves a dis-
aster response or long-term hazard adjustment.(79,80)
Some of the factors associated with people’s person-
alization of risk include “the probability of the im-
pending event occurring [and] the severity, to the in-
dividual, of such a development” (Ref. 81, p. 104).
As this quote indicates, the immediacy of a threat
is important because warning recipients must under-
stand that the message describes an event whose
consequences are likely to occur in the very near
future. Thus, immediacy is related to forewarning,
which is the amount of time between the arrival of
a warning (or personal detection of environmental
cues) and disaster onset. People tend to engage in
622 Lindell and Perry
activities such as information seeking and expedi-
ent property protection when they believe there is
more time before impact than the minimum neces-
sary to implement protective action.(9,82) Successful
warning confirmation can ultimately increase compli-
ance with recommended protective actions but does,
inherently, delay them. Similarly, the amount of time
that risk area residents devote to expedient property
protection also delays their initiation of personal pro-
tective action. In both cases, the delay in protective
action might be dangerous because the time of dis-
aster impact cannot be predicted with perfect accu-
3.3.3. Protective Action Search
If a threat is judged to be real and some un-
acceptable level of personal risk exists, people are
motivated to engage in protective action search—
which involves retrieving one or more feasible pro-
tective actions from memory or obtaining informa-
tion about them from others. The relevant question
in protective action search is “What can be done to
achieve protection?” and its outcome is a decision
set that identifies possible protective actions. In many
instances, an individual’s own knowledge of the haz-
ard will suggest what type of protection to seek (e.g.,
evacuation from floods or sheltering in the basement
from a tornado).
In addition, those in the risk area might become
aware of feasible protective actions by observing so-
cial cues such as the behavior of others. This occurs,
for example, when neighbors are seen packing cars in
preparation for hurricane evacuation.(84) People also
are likely to consider actions with which they have
had vicarious experience by reading or hearing about
others’ protective actions. Such vicarious experience
is frequently transmitted by the news media and re-
layed by peers. Finally, people also are made aware
of appropriate protective actions by means of dis-
aster warnings and hazard awareness programs that
carry protective action recommendations from au-
thorities. Specifically, a well-designed warning mes-
sage will assist recipients by providing guidance in
the form of one or more protective action recom-
mendations.(8) However, such guidance is often in-
adequate. For example, water contamination advi-
sories often suggest boiling drinking water, but fail
to explain whether boiled water is also required for
making ice, brushing teeth, rinsing vegetables, brew-
ing coffee, cooking pasta, taking showers, or wash-
ing dishes, counters, hands, or clothes. They also fre-
quently fail to mention other protective actions such
as chlorinating the water or using bottled water.(85)
3.3.4. Protective Action Assessment
After people have established that at least one
protective action is available, they pass from pro-
tective action search to the protective action as-
sessment stage. This involves examining alternative
actions, evaluating them in comparison to the con-
sequences of continuing normal activities, and deter-
mining which of them is the most suitable response
to the situation. At this point, the primary question
is “What is the best method of protection?” and its
outcome is an adaptive plan.
Choice is an inherent aspect of emergencies
because those at risk generally have at least two
options—taking protective action or continuing nor-
mal activities. Comparing alternatives with respect to
their attributes leads, in turn, to a balancing or trade-
off of these attributes with respect to their relative
importance to the decisionmaker. Under some con-
ditions, those at risk can only take one action and,
therefore, must make a choice among the alterna-
tives (e.g., either evacuate or shelter in-place; Ref. 5,
p. 155). In other cases, people can take multiple ac-
tions and must choose which ones to implement first.
For example, people in earthquake prone areas are
advised to have a working transistor radio with spare
batteries, at least 4 gallons of water in plastic con-
tainers, a complete first-aid kit, and a 4-day supply of
dehydrated or canned food for themselves and their
families. They are also advised to strap water heaters,
tall furniture, and heavy objects to the building walls
and to bolt the house to its foundation.(57)
The end result of protective action assessment
is an adaptive plan, but people’s adaptive plans vary
widely in their specificity, with some being only vague
goals and others begin extremely detailed. At mini-
mum, a specific evacuation plan includes a destina-
tion, a route of travel, and a means of transporta-
tion.(9) More detailed plans include a procedure for
reuniting families if members are separated, advance
contact to confirm the destination is available, con-
sideration of alternative routes if the primary route
is unsafe or too crowded, and alternative methods of
transportation if the primary one is not available.(86)
3.3.5. Protective Action Implementation
The fifth step, protective action implementation,
occurs when all the previous questions about risk
reduction have been answered satisfactorily. In
The Protective Action Decision Model 623
general, the implementation of protective actions
consumes resources people would prefer to allocate
to other activities, so those at risk frequently delay
implementation until they have determined that the
immediacy of the threat justifies the disruption of
normal activities. Thus, people often ask the ques-
tion: “Does protective action need to be taken now?”
The answer to this question, whose outcome is the
threat response, is crucial because people sometimes
postpone the implementation of protective action
even when there is imminent danger. As noted ear-
lier, recipients of hurricane warnings have often been
found to endanger their safety because many wait un-
til the last minute to begin their evacuations.(82,87)
Unfortunately, they fail to recognize that adverse
weather conditions and a high volume of traffic can
significantly reduce the average speed of evacuat-
ing vehicles, thus running the risk that their evac-
uation will not be completed before the arrival of
storm conditions.(25,8890) The problem of procrasti-
nation is even more severe in connection with long-
term hazard adjustment than it is in disasters with
ample forewarning because hazard awareness pro-
grams cannot specify even an approximate deadline
by which action must be taken. For example, an
earthquake prediction might only be able to indicate
a 67% chance of a damaging earthquake within the
next 30 years.(91)
3.3.6. Information Needs Assessment
At any stage of the protective action decision
process, some of the people who receive a warning
might find that the available information is insuffi-
cient to justify a resource-intensive protective action.
When they think time is available, people cope with
the lack of information by searching for additional in-
formation.(9,15,92) The process of information search
begins with an information needs assessment aris-
ing from an individual’s judgment that the available
information is insufficient to justify proceeding fur-
ther in the protective action decision process. Thus, if
any of the questions cannot be answered with an un-
equivocal yes or no, people will ask “What informa-
tion do I need to answer my question?” so they can
generate an identified information need. People com-
monly need additional information about the cer-
tainty, severity, and immediacy of the threat, and lo-
gistical support for protective action such as suitable
evacuation routes, destinations, modes of transporta-
tion, and arrangements for pets and family members
with major medical needs.
3.3.7. Communication Action Assessment
Identification of an information need does not
necessarily suggest where the needed information
can be obtained. Thus, the next question in the in-
formation seeking process is: “Where and how can
I obtain this information?” Addressing this question
leads to information source selection and informa-
tion channel selection, which constitute an informa-
tion search plan. The sources sought are likely to be
affected by the available channels, which in many dis-
asters precludes the use of the telephone because cir-
cuits are so overloaded that it is impossible to ob-
tain a dial tone for hours or even days.(93) Further,
attempts to reach authorities sometimes prove futile
because emergency response agencies are busy han-
dling other calls. Thus, people are often forced to
rely on the mass media and peers—especially for in-
formation about protective actions.(94) This distinc-
tion between risk area residents’ preferred channels
of information receipt and their actual channels of
information receipt also can be seen in connection
with long-term hazard adjustment. For example, res-
idents of communities downstream from the Mt. St.
Helens volcano revealed some significant disparities
between their preferred and actual channels of in-
formation receipt in the years after the 1980 erup-
tions.(5) Moreover, there also were significant differ-
ences between the two communities of Toutle and
Lexington in both their preferred and actual chan-
nels of information receipt.
3.3.8. Communication Action Implementation
The final step in the information search process
is communication action implementation, which pro-
vides decision information by answering the ques-
tion: “Do I need the information now?” If the answer
to this question is positive, that is, they are threat-
ened by an imminent disaster, people will actively
seek the needed information from the most appro-
priate source through the most appropriate channel.
People will go to great lengths, contacting many peo-
ple over a period of minutes to hours, if the prospect
of an imminent disaster needs to be confirmed.(93,95)
Indeed, half of the respondents in one study re-
ported monitoring the news media hourly or more
frequently.(96) However, information seeking will be
less frequent and less active if the location is specific
but the time of impact is ambiguous. Many residents
of the area around Mt. St. Helens monitored the ra-
dio four or more times a day after the initial ash and
624 Lindell and Perry
steam eruptions.(97) By contrast, the absence of lo-
cational specificity and time pressure inherent in a
hazard awareness program provides little need for
those at risk to obtain immediate answers, so they
are likely to forego active information seeking in fa-
vor of passive monitoring of the situation. Unfortu-
nately, the absence of a deadline for action means
this passive monitoring is likely to continue until an
imminent threat arises (as in the case of hurricanes
and floods) or until a disaster strikes (as in the case
of earthquakes).
The actual implementation of behavioral re-
sponse depends not only on people’s intentions to
take those actions but also on conditions in their
physical and social environment that can impede ac-
tions that they intended to take or that can facili-
tate actions that they did not intend to take.(98) In
most cases, the lack of correspondence between in-
tentions and behavior seems to be caused by imped-
iments rather than unexpected facilitators; there are
many instances in which people have wanted to evac-
uate but lacked a safe place to go and a safe route to
travel.(9) Other impediments include a lack of access
to a personal vehicle (e.g., those who are routinely
transit dependent or families in which one spouse has
the only car during the workday) or a lack of personal
mobility due to physical disabilities.(99,100) In cases in-
volving rapid onset disasters, the separation of fam-
ily members can also be an evacuation impediment.
Until family members have been reunited or sepa-
rated family members can establish communication
contact and agree upon a place to meet, evacuation
is unlikely to occur.(101,102)
The final stage in the PADM is a feedback loop
directed by the communication action assessment
that returns to the initial inputs—environmental and
social cues, and information sources, channel access
and preferences, and warning messages. This feed-
back loop is extremely common in sudden onset dis-
asters because people seek to confirm or contradict
any warnings they have received, typically by con-
tacting a different source using a different channel.(9)
Alternatively, some people might be looking to ob-
tain additional information about the threat, about
which protective action to select, or about how to im-
plement a protective action that they have selected.
Finally, people might relay warnings to others or dis-
cuss the implications of the information they have
received.(2) As is the case with the adoption of long-
term hazard adjustments, a perceived lack of urgency
in obtaining information about a hazard or protective
actions can lead to procrastination.
As the previous sections indicate, there has been
a substantial amount of research that has been con-
ducted within the framework of the PADM or on
variables that are part of the PADM (e.g., other re-
search on the relationship of risk perception to evac-
uation). In addition, there has been some limited
discussion of the relationship between the PADM
and the findings from research in domains other
than hazards and disasters—including persuasion, so-
cial conformity, behavioral decision theory, attitude-
behavior theory, and information seeking. However,
there has only been limited discussion of the PADM
in relation to other theories of protective action.
6.1. PADM Comparison to Other Theories of
Protective Action
The PADM identifies a series of information-
processing stages relevant to household adoption
of protective actions and—for each stage—the typ-
ical activity performed, question asked, and out-
come. The characterization of the PADM as a stage
model of protective action warrants a comment
in light of Weinstein, Rothman, and Sutton’s cri-
tique of stage models of health behavior.(103) Like
Janis and Mann’s(4) conflict model, the PADM dif-
fers from stage models such as the Transtheoreti-
cal Model (TTM)(104) and the Precaution Adoption
Process Model (PAPM)(105) in three important ways.
First, risk area residents implement emergency re-
sponse actions such as evacuation and sheltering in-
place temporarily rather than permanently. Thus, the
PADM lacks an equivalent to the TTM’s mainte-
nance stage and the idea of a “termination stage”
has a qualitatively different meaning for the situa-
tions addressed by the PADM. In the TTM, termi-
nation is “the stage at which individuals have zero
temptation”(104) to revert to their previous behavior.
Thus, the TTM is clearly directed toward (health)
behaviors that must be maintained indefinitely
through continuing effort. In the PADM, by contrast,
The Protective Action Decision Model 625
“termination” of an inherently temporary emergency
response would be the stage at which it is safe for
individuals to revert to their previous (“normal”)
Second, emergency response actions have signifi-
cantly different time scales for their decision-making
processes than do health behavior changes. During
emergencies, people might pass through all of the
stages of the PADM in a matter of minutes, whereas
the TTM and PAPM address stages of change that
are more likely to unfold over months or years. Fi-
nally, the PADM is typically applied to situations
in which emergency managers are transmitting in-
formation concurrently to large numbers of people
who are responding to a single “focusing event”(106)
rather than situations in which health professionals
conduct personal interventions that are tailored to
individuals in different stages of a behavioral change
process. Nonetheless, the PADM can be applied to
long-term hazard adjustment as well as emergency
response actions. In such cases as deciding to insure
houses, elevate them above flood level, or bolt them
to their foundations, there are more notable overlaps
between the situations addressed by the PADM, on
the one hand, and the TTM and PAPM, on the other
hand. However, the extent of the overlap needs to be
examined in future research.
In addition to examining the PADM’s compat-
ibility with stage models of health behavior, fu-
ture research also needs to systematically examine
its theoretical similarities to and differences from
other psychological models. For example, hazard
adjustment attributes in the PADM are equiva-
lent to TRA’s attitude toward the act. However,
although disaster research has found that confor-
mity with the behavior of others can be a sig-
nificant influence on people’s evacuation decisions,
there has been no PADM research that has ex-
plicitly addressed the subjective norm as it has
been conceptualized in the TRA. Moreover, the
PADM’s premise that people can vary in the ex-
tent to which they base their protective action deci-
sions upon thoughtful consideration of information
about environmental cues, social context, warning
source characteristics, and warning message charac-
teristics or, alternatively, upon unquestioning com-
pliance with an authority’s recommendation is sim-
ilar to ideas about central versus peripheral routes
to persuasion in the Elaboration Likelihood Model
(ELM).(107) The findings of PADM research that are
compatible with the ELM need to be replicated and
Finally, there are some similarities as well as
distinct differences between PADM and protection
motivation theory (PMT).(108110) PADM’s concep-
tion of hazard-related attributes (protection of per-
sons, protection of property, and utility for other
purposes) is similar to, but broader than, PMT’s re-
sponse efficacy. However, PADM’s conception of
resource-related attributes (cost, time and effort
requirements, knowledge and skill, and required co-
operation) is distinctly different from PMT’s self-
efficacy because the latter seems most closely re-
lated to the knowledge and skill component of the
resource-related attributes. Moreover, the resource-
related attributes are characteristics of a protective
action whereas self-efficacy is a characteristic of the
person. An advantage of the PADM formulation is
that assessing the perceived characteristics of differ-
ent protective actions makes it possible to determine
if people have erroneous perceptions of the alterna-
tive actions and correct those misperceptions if they
exist. In addition, identifying the characteristics of
different protective actions makes it possible to iden-
tify which actions are most likely to be adopted and
to initially focus a risk communication program on
increasing their adoption. Once the most resource-
effective hazard adjustments have been adopted (the
ones that provide the greatest efficacy in protect-
ing persons and property for a given level of re-
sources), the risk communication program can be
redirected toward the more resource-intensive haz-
ard adjustments. Nonetheless, a focus on (personal)
self-efficacy might be more useful than a focus on
(task) resource requirements in some situations. For
example, an emphasis on self-efficacy might be useful
when one must implement a single protective action
(e.g., weight loss) whereas a focus on task demands
(resource requirements) might be more useful when
someone can choose among multiple protective ac-
6.2. Identifying the Core Perceptions
that Determine Protective Actions
One major implication of the literature cited in
the previous section is that, despite extensive theo-
rizing and data collection, it still is not entirely clear
what motivates people to take protective action.
One of the major contributions of PADM research
is the finding that, though risk perception is usu-
ally an important determinant of protective action,
there are other perceptions that are also important—
and sometimes even more important than risk
626 Lindell and Perry
perception. There is a need to consider other as-
pects of threat perception, such as hazard intrusive-
ness, and also to consider perceptions of protective
actions and stakeholders. Nonetheless, it is not en-
tirely clear whether the PADM’s core perceptions of
threat, protective actions, and stakeholders are them-
selves sufficient to account for an adequate percent-
age of the variance in people’s hazard adjustment
This issue is particularly important when hazard
experts, such as seismologists and earthquake engi-
neers, try unsuccessfully to communicate event prob-
abilities and consequences to risk area residents in
an effort to increase the adoption of hazard adjust-
ments. Lindell(111) addressed this problem by char-
acterizing risk area residents’ perceptions of three
hazards (a volcanic eruption of Mt. St. Helens, a
toxic chemical release from a railroad tank car, and
a radiological materials release from a nearby nu-
clear power plant) in terms of four categories of haz-
ard characteristics that were consistent with previous
theorizing about technological hazards(32) and dis-
aster impacts.(1) Respondents rated the three haz-
ards in terms of hazard agent characteristics (like-
lihood of a major release, ease of risk reduction,
and likelihood of release prevention), impact charac-
teristics (speed of onset, existence of environmental
cues, scope of impact, and duration of impact), ex-
pected personal impacts (immediate death, delayed
cancer, genetic effects, and total property loss), and
affective/behavioral reactions (dread, frequency of
thought about the hazard, and frequency of discus-
sion about the hazard). The results suggested that
perceived impact characteristics mediated the rela-
tionship between characteristics of the hazard agent
characteristics and expected personal impacts. Simi-
lar results emerged from a study of household evac-
uation decision making in response to Hurricane
Ike.(84) Perceived storm characteristics (local land-
fall, major intensity, and rapid onset) partially me-
diated the effects of coastal proximity and hurri-
cane experience on expected personal impacts (surge
damage, inland flood damage, storm wind damage,
and casualties), which, in turn, had a direct effect on
evacuation decisions.
Another line of research that is likely to prove
fruitful in understanding people’s conceptions of haz-
ards is the mental models approach.(31) Such studies
interview individuals to elicit the elements of their
beliefs about a specific hazard domain and the in-
terrelationships among the elements of that domain.
After interviewing a number of participants, the indi-
vidual mental models can be aggregated into a con-
sensus model. In a somewhat similar manner, the
hazard beliefs approach(30,112) presents some very
broad questions about a hazard to a pretest group,
identifies common themes in the responses, and de-
velops an inventory of fixed-response items that as-
sess both accurate and erroneous beliefs. Both of
these inductive approaches are likely to prove most
useful in characterizing people’s beliefs about new
technologies such as nanotechnology.(113)
One of the major impediments to assessing how
people think about hazards is that they vary in the
differentiation of their beliefs about a domain and,
among those with differentiated beliefs, there is vari-
ation in the degree to which those beliefs are in-
tegrated.(114) At one extreme are those who have
few or no beliefs about the domain whereas, at the
other extreme, are those who have highly differenti-
ated belief systems characterized by a large number
of elements that vary in the strength of their link-
ages.(115,116) Consistent with these analyses in other
domains, there is evidence that risk area residents
vary in the degree to which they have differentiated
schemas or mental models of environmental haz-
ards. This variation in people’s hazard models has
led some researchers to measure respondents’ reac-
tions to hazards simply in terms of a single question-
naire item of concern about the problem.(117) If risk
area residents have only very diffuse conceptions of
seismic threat, then a global construct such as con-
cern might be a more accurate characterization of
their beliefs than the specific dimensions assumed
by PADM. Further research is needed to determine
what proportions of the population have specific be-
liefs, global beliefs, and no beliefs at all about the en-
vironmental hazards to which they are exposed and
the actions they can take to protect themselves.
The PADM has mostly been applied in three
areas—risk communication programs, evacuation
modeling, and long-term hazard adjustment. Each of
these is discussed below.
7.1. PADM Application to Developing Community
Risk Communication Programs
Research on which the PADM is based has
led to the development of an extensive set of
recommendations for risk communication, particu-
larly in communities with ethnic minorities. These
The Protective Action Decision Model 627
recommendations for community risk communica-
tion programs are based upon the distinct differences
between risk communication activities undertaken
during the continuing hazard phase (the time be-
tween incidents) and those taken during an escalating
crisis (when there is adequate forewarning of disas-
ter impact) or emergency response (when forewarn-
ing is absent). Risk communication during the con-
tinuing hazard phase is directed toward encouraging
long-term hazard adjustments such as hazard mitiga-
tion, emergency preparedness, and hazard insurance
purchase. Risk communication during an escalating
crisis or emergency response is directed toward en-
couraging appropriate disaster responses. However,
risk communication during both phases requires the
development of an effective risk communication pro-
gram. Such programs need to be carefully developed
during the continuing hazard phase because the hu-
man and financial resources available for environ-
mental hazard management are usually limited until
a crisis occurs. Although resources are more read-
ily available during a crisis or emergency response,
time is often severely limited so improvised ef-
forts at risk communications can produce spectacular
There are five basic functions that should be ad-
dressed in the continuing hazard phase. These are
strategic analysis, operational analysis, resource mo-
bilization, program development, and program im-
plementation (see Table I). The purpose of strategic
analysis is to identify community constraints and set
appropriate goals for the overall risk communication
program. The purpose of operational analysis is to
use the elements of the classic communication model
(source, channel, message, receiver, effect, and feed-
back) to identify the community resources that are
available for the risk communication program. The
purpose of resource mobilization is to enlist the sup-
port of stakeholders in the community who are likely
to share an interest in using risk communication to
reduce hazard vulnerability. Program development
for all phases involves the use of available commu-
nity resources to develop a workable system that can
implement risk communication in both the continu-
ing hazard phase and also in an escalating crisis or
emergency response. As soon as the risk communi-
cation program has been developed, it is possible to
immediately begin program implementation for the
continuing hazard phase. This function involves con-
ducting the activities that will encourage risk area
residents to adopt long-term hazard adjustments. It
also involves conducting the activities that will allow
Table I. Tasks for the Continuing Hazard Phase
Strategic analysis
Conduct a community hazard/vulnerability analysis
Analyze the community context
Identify the community’s prevailing perceptions of the hazards
and hazard adjustments
Set appropriate goals for the risk communication program
Operational analysis
Identify and assess feasible hazard adjustments for the
community and its households/businesses
Identify ways to provide incentives, sanctions, and
technological innovations
Identify the available risk communication sources in the
Identify the available risk communication channels in the
Identify specific audience segments
Resource mobilization
Obtain the support of senior appointed and elected officials
Enlist the participation of other government agencies
Enlist the participation of nongovernmental (nonprofit) and
private sector organizations
Work with the mass media
Work with neighborhood associations and service
Program development for all phases
Staff, train, and exercise a crisis communications team
Establish procedures for maintaining an effective
communication flow in an escalating crisis and in emergency
Develop a comprehensive risk communication program
Plan to make use of informal communication networks
Establish procedures for obtaining feedback from the news
media and the public
Program implementation for the continuing hazard phase
Build source credibility by increasing perceptions of expertise
and trustworthiness
Use a variety of channels to disseminate hazard information
Describe community or facility hazard adjustments being
planned or implemented
Describe feasible household hazard adjustments
Evaluate program effectiveness
Source: Lindell & Perry (2004).
authorities to determine if the risk communication
program has been effective.
The format of Table I might seem to imply that
the five risk communication functions form a simple
linear sequence, but some tasks will be performed
concurrently and the entire process will frequently be
iterative. For example, some resource mobilization
tasks might take place concurrently with the opera-
tional analysis, or tasks conducted during the opera-
tional analysis phase might be suspended temporar-
ily in order to return to the strategic analysis and
refine it.
628 Lindell and Perry
Once authorities have determined that they are
in an escalating crisis or emergency response, they
need to implement the predetermined risk communi-
cation actions that were developed during the contin-
uing hazard phase. These include activating the crisis
communication team promptly, determining the ap-
propriate time to release sensitive information, and
selecting the communication channels appropriate to
the situation. An escalating crisis or emergency re-
sponse also requires authorities to maintain source
credibility with the news media and the public, pro-
vide timely and accurate information to the news
media and the public, and evaluate performance
through postincident critiques. For further details on
the application of the PADM to the development
of community risk communication programs, see
Ref. 13, chapter 5.
7.2. PADM Application to Evacuation Models
Most of the research on the PADM and similar
models has focused on the prediction of whether peo-
ple will engage in a particular behavior. Although
this is an important application, there are cases in
which it is important to predict when people will en-
gage in a particular behavior. For example, the tim-
ing of a household’s evacuation departure is a critical
input to transportation analysts’ evacuation models
because evacuation routes can become overloaded
if too many vehicles enter the evacuation route sys-
tem at the same time. Thus, local officials seeking to
manage the evacuation need to understand the rate
at which people receive an initial warning and the
time it takes them to prepare to evacuate (see Refs.
119 and 120 for further discussion of evacuation time
components). Each warning mechanism (siren, tone-
alert radio, radio, television, and telephone) has a
characteristic rate of exposure over time.(5,23,121) In
addition, warning mechanisms vary in terms of their
ability to attract attention and provide comprehensi-
ble messages that will change risk area residents’ core
perceptions of threat, protective actions, and stake-
holders in the desired directions.
The time households spend in preparing to
evacuate appears to involve two components, men-
tal preparation and logistical preparation. Mental
preparation is the time it takes to collect and pro-
cess the information needed to decide an evacua-
tion is necessary. The existence of a period of men-
tal preparation has been documented in research
indicating that people engage in milling, during
which time they seek confirmation that a danger ex-
ists, obtain further information about the threat and
alternative protective actions, and relay warnings to
peers.(2) The broad outlines of this process are under-
stood,(13,14,122) but there appears to be no research on
the duration of psychological preparation for protec-
tive action.
Logistical preparation is the time it takes for
a household to assemble the people and resources
needed to implement protective action. Thus, evac-
uees need to perform tasks such as assembling fam-
ily members, packing bags for the trip, and securing
the house before leaving.(82) The time required for
mental preparation and logistical preparation both
contribute to total preparation time, but these com-
ponents do not seem to be additive because the two
components cannot be presumed to be mutually ex-
clusive. That is, some household members might be
searching for information at the same time that oth-
ers are packing bags and still others installing hur-
ricane shutters. Available evidence has documented
that households’ preparation times vary as a function
of incident characteristics; for example, preparation
times for evacuation from the Mt. St. Helens eruption
were noticeably different from those for Hurricane
Lili.(82) Because of these and other differences, pre-
dictions of evacuation departure times have been less
successful than predictions of households’ evacuation
decisions. Some researchers have tried to improve
prediction by adopting estimation models such as
neural networks, sequential logit, and survival anal-
ysis but these have had modest success.(87,123125)
Other research has extended the scope of the
PADM to examine evacuation logistics—people’s
actions between the time they decide to evacuate
and the time they arrive at their evacuation destina-
tions(86) as well as the return entry process.(126) The
latter process is, in some respects, just the reverse of
the evacuation problem. However, it is more com-
plicated because of the logistical problems author-
ities experience in communicating reentry informa-
tion to evacuees that might be scattered over dozens
of towns in multiple states.
7.3. PADM Application to Hazard
Adjustment Adoption
A recent review of research on household adop-
tion of hazard adjustments concluded that risk per-
ception is consistently related to the adoption of
hazard adjustments, but perceptions of stakehold-
ers and hazard adjustments are also relevant and
deserve greater attention.(36) There is considerable
The Protective Action Decision Model 629
evidence that hazard experience increases hazard ad-
justment adoption, but hazard proximity and hazard
intrusiveness also appear to play significant roles and
should be the focus of additional research. Finally,
demographic variables continue to be unreliable pre-
dictors of hazard adjustment adoption but should re-
ceive continuing attention to assess their effects on
risk perception, stakeholder perceptions, and hazard
adjustment perceptions, as well as hazard experience,
hazard proximity, and hazard intrusiveness.
One of the major limitations of hazards and dis-
asters studies to date is that the PADM predicts
that some of the variables should form causal chains.
For example, hazard proximity is predicted to cause
hazard experience, hazard experience is predicted
to cause risk perception, and risk perception is pre-
dicted to cause hazard adjustment adoption. Another
example is that perceptions of information sources’
expertise, trustworthiness, and protection responsi-
bility should affect risk area residents’ risk percep-
tions, which, in turn, should affect their adoption
of hazard adjustments. That is, each successive vari-
able is hypothesized to completely mediate the re-
lationship between the variable that precedes it and
the variable that follows it so that, for example,
the estimated effect of stakeholder perceptions on
hazard adjustment adoption becomes nonsignificant
when controlling for risk perception. Most hazards
and disasters studies to date have used single equa-
tion models to predict a single dependent variable
such as household hazard adjustment or evacuation
so their results cannot shed any light on the me-
diation hypotheses. Moreover, the few studies that
did report tests of the PADM’s mediation hypothe-
ses have found only partial mediation where com-
plete mediation was expected.(34,44) Nonetheless, the
results did suggest that risk area residents’ aware-
ness of the available hazard adjustments and accurate
perceptions of their attributes do mediate the rela-
tionship between hazard experience and hazard ad-
justment.(59,127) In addition, Lindell and Prater found
evidence that hazard intrusiveness had an effect on
hazard adjustment adoption that was independent of
perceived personal risk.(34)
In some cases, the failure of a complete me-
diation hypothesis has a theoretically logical expla-
nation. For example, the inability of risk percep-
tion or hazard intrusiveness to completely mediate
the relationship between perceived stakeholder char-
acteristics and hazard adjustment adoption is con-
sistent with Petty and Cacioppo’s(128) distinction be-
tween central and peripheral routes to persuasion
and Chaiken’s(129) distinction between systematic
and heuristic processes. Both theories assert that in-
cidental factors such as stakeholder perceptions af-
fect attitudes and behavior without affecting salient
beliefs or subjective norms (to use terms from TRA).
Further research is needed to determine if the fail-
ures to find complete mediation are due to additional
variables specified by the PADM, other variables not
specified by the PADM, or are simply manifestations
of the effects of peripheral/heuristic processing.
This article has described an updated version
of the PADM and summarized the research that
supports its principal components. This review in-
cludes articles that were cited in Lindell and Perry,(13)
but also includes other research conducted prior to
and subsequent to that book’s publication. There
are varying degrees of support for different com-
ponents of the PADM. Some components have ex-
tensive support, whereas others have conflicting or
counter-intuitive findings, and still others are largely
untested. Future research should seek to resolve the
conflicting findings, explain the counterintuitive find-
ings, and examine the untested propositions.
In addition to its predictive validity, any theoret-
ical model should be useful. Accordingly, this arti-
cle has described the ways in which the PADM has
been applied to three different areas—risk commu-
nication, evacuation modeling, and long-term hazard
adjustment. Although these applications are in their
early stages, the evidence to date suggests that the
PADM is continuing to evolve into a useful frame-
work for managing societal response to environmen-
tal hazards. Applications of the model would not
need to produce large changes in behavior to be valu-
able. Updating an argument made in an earlier arti-
cle,(41) major disasters are so costly—the economic
impact of Hurricane Katrina was estimated to range
from $81 billion(130) to $140 billion(131)—that even a
1% decrease in losses would save hundreds of mil-
lions if not billions of dollars.
This work was supported by the National Sci-
ence Foundation under Grants SES 0527699, SES
0838654, and CMM 0927739. None of the conclusions
expressed here necessarily reflects views other than
those of the authors.
630 Lindell and Perry
1. Barton A. Communities in Disaster. New York: Doubleday,
2. Drabek TE. Human System Responses to Disaster: An
Inventory of Sociological Findings. New York: Springer-
Verlag, 1986.
3. Fritz CE. Disaster. In Merton RK, Nisbet RA (eds). Con-
temporary Social Problems. New York: Harcourt, Brace &
World, 1961.
4. Janis I, Mann L. Decision Making: A Psychological Analysis
of Conflict, Choice and Commitment. New York: Free Press,
5. Lindell MK, Perry RW. Behavioral Foundations of Com-
munity Emergency Planning. Washington, DC: Hemisphere
Press, 1992.
6. Mileti DS, Drabek T, Haas JE. Human Systems in Extreme
Environments. Boulder, CO: University of Colorado Insti-
tute of Behavioral Science, 1975.
7. Mileti DS, Peek L. The social psychology of public response
to warnings of a nuclear power plant accident. Journal of
Hazardous Materials, 2000; 75:181–194.
8. Mileti, D S, Sorensen JH. Why people take precautions
against natural disasters. Pp. 296–320 in Weinstein N (ed).
Taking Care: Why People Take Precautions. New York:
Cambridge University Press, 1987.
9. Perry RW, Lindell MK, Greene MR. Evacuation Planning in
Emergency Management. Lexington, MA: Heath Lexington
Books, 1981.
10. Tierney K, Lindell, MK, Perry RW. Facing the Un-
expected: Disaster Preparedness and Response in the
United States. Washington, DC: Joseph Henry Press,
11. Burton I, Kates, R, White GF. The Environment as Hazard,
2nd edition. New York: Guildford Press, 1993.
12. Lindell MK, Alesch D, Bolton PA, Greene MR, Larson
LA, Lopes R, May PJ, Mulilis J-P, Nathe S, Nigg JM, Palm
R, Pate P, Perry RW, Pine J, Tubbesing SK, Whitney DJ.
Adoption and implementation of hazard adjustments. Inter-
national Journal of Mass Emergencies and Disasters Special
Issue, 1997; 15:327–453. Available at:, Ac-
cessed 1 May 2010.
13. Lindell MK, Perry RW. Communicating Environmental
Risk in Multiethnic Communities. Thousand Oaks, CA:
Sage, 2004.
14. Gladwin CH, Gladwin H, Peacock WG. Modeling hurri-
cane evacuation decisions with ethnographic methods. Inter-
national Journal of Mass Emergencies and Disasters, 2001;
19:117–143. Available at:, Accessed 1 May
15. Perry RW, Greene, M. Citizen Response to Volcanic Erup-
tions. New York: Irvington, 1983.
16. Lindell MK, Prater CS, Perry RW. Fundamentals of
Emergency Management. Emmitsburg, MD: Federal
Emergency Management Agency Emergency Management
Institute, 2006. Available at:
EMIWeb/edu/fem.asp, Accessed 1 May 2010.
17. Lasswell H. The structure and function of communication
in society. Pp. 43–71 in Bryson L (ed). Communication of
Ideas. New York: Harper, 1948.
18. McGuire WJ. The nature of attitudes and attitude change.
Pp. 329–334 in Lindsey G, Aronson E (eds). Handbook
of Social Psychology, 2nd edition. Reading, MA: Addison-
Wesley, 1969.
19. McGuire WJ. The nature of attitudes and attitude change.
Pp. 233–256 in Lindsey G, Aronson E (eds). Handbook of
Social Psychology, 3rd edition. New York: Random House,
20. O’Keefe D. Persuasion: Theory and Research, 2nd edition.
Thousand Oaks, CA: Sage, 2002.
21. Petty RE, Wegener DT. Attitude change: Multiple roles for
persuasion variables. Pp. 323–390 in Gilbert DT, Fiske ST,
Lindzey G (eds). Handbook of Social Psychology, 4th edi-
tion. New York: McGraw-Hill, 1998.
22. Lindell MK, Prater CS, Peacock WG. Organizational com-
munication and decision making in hurricane emergencies.
Natural Hazards Review, 2007; 8:50–60.
23. Rogers GO, Sorensen JH. Diffusion of emergency warnings.
Environmental Professional, 1988; 10:185–198.
24. Aguirre BE. The lack of warnings before the Saragosa tor-
nado. International Journal of Mass Emergencies and Disas-
ters, 1988; 6:65–74. Available at:, Accessed 1
May 2010.
25. Baker EJ. Hurricane evacuation behavior. International
Journal of Mass Emergencies and Disasters, 1991; 9:287–310.
Available at:, Accessed 1 May 2010.
26. Fiske ST, Taylor SE. Social Cognition: From Brains to Cul-
ture. Boston: McGraw-Hill, 2008.
27. Committee on Disaster Research in the Social Sciences.
Facing Hazards and Disasters: Understanding Human Di-
mensions. Washington, DC: National Academy of Sci-
ences/National Research Council, 2006.
28. Mayhorn CB. Cognitive aging and the processing of hazard
information and disaster warnings. Natural Hazards Review,
2005; 6:165–170.
29. Lindell MK, Prater CS. Tsunami preparedness on the Ore-
gon and Washington coast: Recommendations for research.
Natural Hazards Review, 2010; 11:69–81.
30. Turner R, Nigg J, Heller-Paz D. Waiting for Disaster. Los
Angeles: University of California Press, 1986.
31. Morgan MG, Fischhoff B, Bostrom A, Atman CJ. Risk Com-
munication: A Mental Models Approach. Cambridge: Cam-
bridge University Press, 2002.
32. Slovic P, Fischhoff B, Lichtenstein S. How safe is safe
enough: A psychometric study of attitudes toward techno-
logical risks and benefits. Pp. 80–103 in Slovic P (ed). The
Perception of Risk. London: Earthscan, 1980.
33. Terpstra T, Lindell MK, Gutteling JM. Does communicat-
ing (flood) risk affect (flood) risk perceptions? Results of
a quasi-experimental study. Risk Analysis, 2009; 29:1141–
34. Lindell MK, Prater CS. Household adoption of seismic haz-
ard adjustments: A comparison of residents in two states.
International Journal of Mass Emergencies and Disasters,
2000; 18:317–338. Available at:, Accessed 1
May 2010.
35. Sorensen JH. Hazard warning systems: Review of 20 years
of progress. Natural Hazards Review, 2000; 1:119–125.
36. Lindell MK. North American cities at risk: Household re-
sponses to environmental hazards. In. Rossetto T, Joffe H,
Adams J (eds). Cities at Risk: Living with Perils in the 21st
Century. Dordrecht: Springer, in press.
37. Mileti DS, Fitzpatrick C. The causal sequence of risk com-
munication in the Parkfield earthquake prediction experi-
ment. Risk Analysis, 1992; 12:393–400.
38. Dash N, Gladwin H. Evacuation decision making and behav-
ioral responses: Individual and household. Natural Hazards
Review, 2007; 8:69–77.
39. Terpstra T, Gutteling JM. Households’ perceived respon-
sibilities in flood risk management in the Netherlands. In-
ternational Journal of Water Resources Development, 2008;
40. Perry RW, Lindell MK. Volcanic risk perception and adjust-
ment in a multi-hazard environment. Journal of Volcanology
and Geothermal Research, 2008; 172:170–178.
41. Lindell MK, Whitney DJ. Correlates of seismic hazard ad-
justment adoption. Risk Analysis, 2000; 20:13–25.
42. Mileti DS, Darlington JD. The role of searching in shaping
reactions to earthquake risk information. Social Problems,
1997; 44:89–103.
The Protective Action Decision Model 631
43. Ge Y, Peacock WG, Lindell MK. Florida households’ ex-
pected responses to hurricane hazard mitigation incentives.
Risk Analysis, in press.
44. Lindell MK, Hwang SN. Households’ perceived personal
risk and responses in a multi-hazard environment. Risk
Analysis, 2008; 28:539–556.
45. Weinstein ND. Effects of personal experience on self-
protective behavior. Psychological Bulletin, 1989; 105:31–
46. Palm R, Hodgson M, Blanchard, RD, Lyons D. Earthquake
Insurance in California. Boulder, CO: Westview Press, 1990.
47. Peacock WG, Brody SD, Highfield W. Hurricane risk per-
ceptions among Florida’s single family homeowners. Land-
scape and Urban Planning, 2005; 73:120–135.
48. Preston V, Taylor SM, Hedge DC. Adjustment to natural
and technological hazards: A study of an urban residential
community. Environment and Behavior, 1983; 15:143–164.
49. Lindell MK, Earle TC. How close is close enough: Public
perceptions of the risks of industrial facilities. Risk Analysis,
1983; 3:245–253.
50. Arlikatti S, Lindell MK, Prater CS, Zhang Y. Risk area ac-
curacy and hurricane evacuation expectations of coastal res-
idents. Environment and Behavior, 2006; 38:226–247.
51. Zhang Y, Prater CS, Lindell MK. Risk area accuracy and
evacuation from Hurricane Bret. Natural Hazards Review,
2004; 5:115–120.
52. Lindell MK, Perry RW. Household adjustment to earth-
quake hazard: A review of research. Environment and Be-
havior, 2000; 32:590–630.
53. Fishbein MF, Ajzen I. Predicting and Changing Behavior:
The Reasoned Action Approach. New York: Psychology
Press, 2010.
54. Mulilis J-P, Duval TS. Negative threat appeals and earth-
quake preparedness: A person-relative-to-event PrE model
of coping with threat. Journal of Applied Social Psychology,
1995; 25:1319–1339.
55. Kunreuther H, Ginsberg R, Miller L, Sagi P, Slovic P,
Borkan B, Katz N. Disaster Insurance Protection: Public
Policy Lessons. New York: John Wiley, 1978.
56. Davis MS. Living along the fault line: An update on earth-
quake awareness and preparedness in Southern California.
Urban Resources, 1989; 5:8–14.
57. Russell L, Goltz JD, Bourque LB. Preparedness and hazard
mitigation actions before and after two earthquakes. Envi-
ronment and Behavior, 1995; 27:744–770.
58. Lindell MK, Arlikatti S, Prater CS. Why people do what they
do to protect against earthquake risk: Perceptions of hazard
adjustment attributes. Risk Analysis, 2009; 29:1072–1088.
59. Lindell MK, Prater CS. Risk area residents’ perceptions and
adoption of seismic hazard adjustments. Journal of Applied
Social Psychology, 2002; 32:2377–2392.
60. Terpstra T, Lindell MK. Citizens’ Perceptions of Flood Haz-
ard Adjustments: An Application of the Protective Action
Decision Model. College Station, TX: Texas A&M Univer-
sity Hazard Reduction & Recovery Center, 2009.
61. Pijawka KD, Mushkatel AH. Public opposition to the siting
of the high-level nuclear waste repository: The importance
of trust. Policy Studies Review, 1991; 10:180–194.
62. Lang JT, Hallman WK. Who does the public trust? The case
of genetically modified food in the United States. Risk Anal-
ysis, 2005, 25:1241–1252.
63. French JRP, Raven BH. The bases of social power. Pp. 150–
167 in Cartwright D (ed). Studies in Social Power. Ann Ar-
bor, MI: Institute for Social Research, 1959.
64. Raven B. Social influence and power. Pp. 371–382 in Steiner
I, Fishbein M (eds). Current Studies in Social Psychology.
New York: Holt, Rinehart & Winston, 1965.
65. Raven BH. The bases of power: Origins and recent develop-
ments. Journal of Social Issues, 1993; 49:227–251.
66. Eagly AH, Chaiken S. The Psychology of Attitudes. Ft.
Worth, TX: Harcourt, Brace College Publishers, 1993.
67. Arlikatti S, Lindell MK, Prater CS. Perceived stakeholder
role relationships and adoption of seismic hazard adjust-
ments. International Journal of Mass Emergencies and Dis-
asters, 2007; 25:218–256. Available at:, Ac-
cessed 1 May 2010.
68. Meyer P. Defining and measuring credibility of newspapers:
Developing an index. Journalism Quarterly, 1988; 65:567–
574, 588.
69. Jackson EL. Response to earthquake hazard: The West
Coast of North America. Environment and Behavior, 1981;
70. Garcia EM. Earthquake preparedness in California: A sur-
vey of Irvine residents. Urban Resources, 1989; 5:15–19.
71. Mulilis JP, Duval TS. The PrE model of coping with threat
and tornado preparedness behavior: The moderating effects
of felt responsibility. Journal of Applied Social Psychology,
1997; 27:1750–1766.
72. Lazarus RS, Folkman S. Stress, Appraisal, and Coping. New
York: Springer, 1984.
73. Mileti DS. Natural hazards warning systems in the United
States. Boulder, CO: University of Colorado Institute of Be-
havioral Science, 1975.
74. Perry RW. Evacuation decision making in natural disaster.
Mass Emergencies, 1979; 4:25–38.
75. Perry RW, Hirose H. Volcano Management in the United
States and Japan. Greenwich, CT: JAI Press, 1991.
76. Blanchard-Boehm RD. Understanding public response to
increased risk from natural hazards. International Journal of
Mass Emergencies and Disasters, 1998; 16:247–278. Avail-
able at:, Accessed 1 May 2010.
77. Houts PS, Cleary PD, Hu TW. The Three Mile Island Cri-
sis: Psychological, Social and Economic Impacts on the Sur-
rounding Population. University Park, PA: Pennsylvania
State University Press, 1988.
78. Perry RW. Comprehensive Emergency Management: Evac-
uating Threatened Populations. Greenwich, CT: JAI,
79. Fritz CE, Marks, E. The NORC studies of human be-
havior in disaster. Journal of Social Issues, 1954; 10:26–
80. Perry RW. Environmental hazards and psychopathology.
Environmental Management, 1983; 7:543–552.
81. Withey S. Reaction to uncertain threat. Pp. 93–123 in Baker
G, Chapman D (eds). Man and Society in Disaster. New
York: Basic Books, 1962.
82. Lindell MK, Lu JC, Prater CS. Household decision making
and evacuation in response to Hurricane Lili. Natural Haz-
ards Review, 2005; 6:171–179.
83. Lindell MK, Prater CS. A hurricane evacuation manage-
ment decision support system (EMDSS). Natural Hazards,
2007; 40:627–634.
84. Huang SK, Lindell MK, Prater CS, Wu HC, Siebeneck LK.
Household Evacuation Decision Making in Response to
Hurricane Ike. College Station, TX: Texas A&M University
Hazard Reduction & Recovery Center, 2010.
85. Lindell MK, Mumpower J, Huang S-K, Wu H-C. Percep-
tions and Expected Responses to a Water Contamination
Emergency. College Station, TX: Texas A&M University
Hazard Reduction & Recovery Center, 2010.
86. Lindell MK, Kang JE, Prater CS. The logistics of household
evacuation in Hurricane Lili. Natural Hazards, in press.
87. Fu H, Wilmot CG, Zhang H, Baker EJ. Modeling the hur-
ricane evacuation response curve. Transportation Research
Record, 2007; 2022:94–102.
88. Baker EJ. Hurricane evacuation in the United States. Pp.
98–108 in Pielke R Jr, Pielke R Sr (eds). Storms, Vol .1. Lon-
don: Routledge, 2000.
632 Lindell and Perry
89. Dow K, Cutter SL. Crying wolf: Repeat responses to hurri-
cane evacuation orders. Coastal Management, 1998; 26:237–
90. Dow K, Cutter SL. Emerging hurricane evacuation issues:
Hurricane Floyd and South Carolina. Natural Hazards Re-
view, 2002; 3:12–18.
91. Southern California Earthquake Center. Uniform California
Earthquake Rupture Forecast (UCERF), 2011. Available at:, Accessed 27 January 2011.
92. Hanson S, Vitek JD, Hanson PO. Natural disaster: Long-
range impact on human response to future disaster threats.
Environment and Behavior, 1979; 11:268–284.
93. Drabek TE. Social processes in disaster: Family evacuation.
Social Problems, 1969; 16:336–347.
94. Lindell MK, Perry RW. Risk area residents’ changing per-
ceptions of volcano hazard at Mt. St. Helens. Pp. 159–166 in
Siccardi F, Nigg J, Nemec J (eds). Prediction and Perception
of Natural Hazards. Amsterdam: Kluwer Academic Publish-
ers, 1993.
95. Drabek TE, Stephenson J. When disaster strikes. Journal of
Applied Social Psychology, 1971; 1:187–203.
96. Morss RE, Hayden MH. Storm surge and “certain death”:
Interviews with Texas coastal residents following Hurricane
Ike. Weather, Climate, and Society, 2010; 2:174–189.
97. Perry RW, Lindell MK, Greene MR. Threat perception and
public response to volcano hazard. Journal of Social Psychol-
ogy, 1982; 116:199–204.
98. Triandis HC. Values, attitudes, and interpersonal behavior.
Pp. 195–259 in Howe H, Page M (eds). Nebraska Symposium
on Motivation, Vol. 27. Lincoln, NE: University of Nebraska
Press, 1980.
99. Heath SE, Kass PH, Beck AM, Glickman LT. Human and
pet-related risk factors for household evacuation failure dur-
ing a natural disaster. American Journal of Epidemiology,
2001; 153;659–667.
100. Van Willigen M, Edwards T, Edwards B, Hessee S. Riding
out the storm: Experiences of the physically disabled during
Hurricanes Bonnie, Dennis, and Floyd. Natural Hazards Re-
view, 2002; 3:98–106.
101. Killian LM. The significance of multi-group membership in
disaster. American Journal of Sociology, 1952; 57:309–314.
102. Drabek TE, Boggs K. Families in disaster: Reactions and rel-
atives. Journal of Marriage and the Family, 1968; 30:443–451.
103. Weinstein ND, Rothman AJ, Sutton SR. Stage theories
of health behavior: Conceptual and methodological issues.
Health Psychology, 1998; 17:290–299.
104. Prochaska JO, Johnson S, Lee P. The transtheoretical model
of behavior change. Pp. 59–84 in Gochman DS (ed). Hand-
book of Health Behavior Research I: Personal and Social
Determinants. New York: Plenum Press, 1997.
105. Weinstein ND. The precaution adoption process. Health
Psychology, 1988; 7:355–386.
106. Birkland TA. After Disaster: Agenda Setting, Public Policy
and Focusing Events. Washington, DC: Georgetown Uni-
versity Press, 1997.
107. Petty RE, Wegener DT. The elaboration likelihood model:
Current status and controversies. Pp. 37–72 in Chaiken,
Trope Y (eds). Dual-Process Theories in Social Psychology.
New York: Guilford Press, 1999.
108. Floyd DL, Prentice-Dunn S, Rogers RW. A meta-analysis
of research on protection motivation theory. Journal of Ap-
plied Social Psychology, 2000; 30:407–429.
109. Neuwirth K, Dunwoody S, Griffin RJ. Protection motivation
and risk communication. Risk Analysis, 2000; 20:721–734.
110. Rogers RW. A protection motivation theory of fear appeals
and attitude change. Journal of Psychology, 1975; 91:93–114.
111. Lindell MK. Perceived characteristics of environmental haz-
ards. International Journal of Mass Emergencies and Dis-
asters, 1994; 12:303–326. Available at:, Ac-
cessed 1 May 2010.
112. Whitney DJ, Lindell MK, Nguyen DH. Earthquake beliefs
and adoption of seismic hazard adjustments. Risk Analysis,
2004; 24:87–102.
113. Bostrom A, Lofstedt RE. Nanotechnology risk communica-
tion past and prologue. Risk Analysis, 2010; 30:1645–1662.
114. Scott WA, Osgood DW, Peterson C. Cognitive Structure:
Theory and Measurement of Individual Differences. New
York: Wiley, 1979.
115. Converse PE. The nature of belief systems in mass publics.
Pp. 206–261 in Apter DE (ed). Ideology and Discontent.
New York: Free Press, 1964.
116. Schuman H, Kalton G. Survey methods. Pp. 635–698 in
Lindzey G, Aronson E (eds). The Handbook of Social Psy-
chology, 3rd edition, Vol. 1. Reading, MA: Addison-Wesley,
117. Dooley D, Catalano R, Mishra S, Serxner S. Earthquake pre-
paredness: Predictors in a community survey. Journal of Ap-
plied Social Psychology, 1992; 22:451–470.
118. Seeger MW, Sellnow TL, Ulmer RR. Communication and
Organizational Crisis. Westport, CT: Praeger, 2003.
119. Lindell MK. EMBLEM2: An empirically based large-scale
evacuation time estimate model. Transportation Research,
2008; A 42:140–154.
120. Lindell MK, Prater CS. Critical behavioral assumptions in
evacuation analysis for private vehicles: Examples from hur-
ricane research and planning. Journal of Urban Planning and
Development, 2007; 133:18–29.
121. Lindell MK, Perry RW. Warning mechanisms in emer-
gency response systems. International Journal of Mass
Emergencies and Disasters, 1987; 5:137–153. Available at:, Accessed 1 May 2010.
122. Gladwin H, Peacock WG. Warning and evacuation: A night
for hard houses. Pp. 52–74 in Peacock WG, Morrow BH,
Gladwin H (eds). Hurricane Andrew: Gender, Ethnicity and
the Sociology of Disasters. London: Routledge, 1997.
123. Dixit VV, Pande A, Radwan E, Abdel-Aty M. Understand-
ing the impact of a recent hurricane on mobilization time
during a subsequent hurricane. Transportation Research
Record, 2008; 2041:49–57.
124. Fu H, Wilmot CG. Survival analysis–based dynamic travel
demand models for hurricane evacuation. Transportation
Research Record, 2006; 1964:211–218.
125. Wilmot CG, Mei B. Comparison of alternative trip genera-
tion models for hurricane evacuation. Natural Hazards Re-
view, 2004; 5:170–178.
126. Siebeneck LK, Cova TJ. An assessment of the return-entry
process for Hurricane Rita 2005. International Journal of
Mass Emergencies and Disasters, 2008; 26:91–111. Available
at:, Accessed 1 May 2010.
127. Norris FH, Smith T, Kaniasty K. Revisiting the experience-
behavior hypothesis: The effects of Hurricane Hugo on haz-
ard preparedness and other self-protective acts. Basic and
Applied Social Psychology, 1999; 21:37–47.
128. Petty RE, Cacioppo JT. Communication and Persua-
sion: Central and Peripheral Routes to Attitude Change.
New York: Springer-Verlag, 1986.
129. Chaiken S. The heuristic model of persuasion. Pp. 3–39 in
Zanna MP, Olson JM, Herman CP (eds). Social Influence:
The Ontario Symposium, Vol. 5. Hillsdale NJ: Erlbaum,
130. Knabb RD, Rhome JR, Brown DP. Tropical Cyclone
Report: Hurricane Katrina 23–30 August 2005. Mimi, FL:
National Hurricane Center, 2006. Available at: http://www.
nhc.noaa.govpdf/TCR-AL122005 Katrina.pdf, Accessed 25
January 2011.
131. Burton L, Hicks MJ. Hurricane Katrina: Preliminary Esti-
mates of Commercial and Public Sector Damages. Hunting-
ton, WV: Marshall University for Economic and Business
Research, 2005. Available at:,
Accessed 1 January 2011.
... Members of the public use weather warnings as one source of information to determine their behavioural response. Lindell and Perry (2012) describe how the decision-making process follows a series of ideal stages: ...
... Environmental and social cues are key sources of information that can influence the receiver's perceptions and actions (Lindell & Perry 2012). Environmental cues, such as a gathering storm, a funnel cloud, a roaring sound or heavy rain, are often considered when determining a response, such as during the Joplin, Missouri and US tornado warnings (Kuligowski et al. 2014). ...
... This can be via a range of channels, including social media, phone calls, text messages and face-to-face. The availability of assistance, including the provision of transport and shelters, also provides cues (Lindell & Perry 2012). ...
Full-text available
In this chapter, we introduce early warning systems (EWS) in the context of disaster risk reduction, including the main components of an EWS, the roles of the main actors and the need for robust evaluation. Management of disaster risks requires that the nature and distribution of risk are understood, including the hazards, and the exposure, vulnerability and capacity of communities at risk. A variety of policy options can be used to reduce and manage risks, and we emphasise the contribution of early warnings, presenting an eight-component framework of people-centred early warning systems which highlights the importance of an integrated and all-society approach. We identify the need for decisions to be evidence-based, for performance monitoring and for dealing with errors and false information. We conclude by identifying gaps in current early warning systems, including in the social components of warning systems and in dealing with multi-hazards, and obstacles to progress, including issues in funding, data availability, and stakeholder engagement.
... Members of the public use weather warnings as one source of information to determine their behavioural response. Lindell and Perry (2012) describe how the decision-making process follows a series of ideal stages: ...
... Environmental and social cues are key sources of information that can influence the receiver's perceptions and actions (Lindell & Perry 2012). Environmental cues, such as a gathering storm, a funnel cloud, a roaring sound or heavy rain, are often considered when determining a response, such as during the Joplin, Missouri and US tornado warnings (Kuligowski et al. 2014). ...
... This can be via a range of channels, including social media, phone calls, text messages and face-to-face. The availability of assistance, including the provision of transport and shelters, also provides cues (Lindell & Perry 2012). ...
Full-text available
In this chapter, we explore the challenges of achieving a level of awareness of disaster risk, by each person or organisation receiving a warning, which allows them to take actions to reduce potential impacts while being consistent with the warning producer’s capabilities and cost-effectiveness considerations. Firstly we show how people respond to warnings and how the nature and delivery of the warning affects their response. We look at the aims of the person providing the warning, the constraints within which they must act and the judgement process behind the issue of a warning. Then we address the delivery of the warning, noting that warning messages need to be tailored to different groups of receivers, and see how a partnership between warner and warned can produce a more effective result. We include illustrative examples of co-design of warning systems in Argentina and Nepal, experience in communicating uncertainty in Germany and the Weather-Ready Nation initiative in the USA. We conclude with a summary of aspects of the warning that need to be considered between warner and decision-maker when designing or upgrading a warning system.
... The PADM is one of a few influential theories that explain how decisions are made in risky situations [25]. Other theories, such as the theory of planned behavior [26,27] and the theory of protection motivation [28], focus more on individual risks and personal health contexts. ...
... Other theories, such as the theory of planned behavior [26,27] and the theory of protection motivation [28], focus more on individual risks and personal health contexts. Designed for and applied to collective risks and mass emergencies, the PADM is particularly useful for understanding behaviors in a nuclear accident scenario, which is the context in which it was first applied [13,25]. However, this model has also been used in studies on other types of collective risks, such as wildfires [29], chemical release from the petrochemical industry [30], city smog [31], flood hazards [32], and the 'not-in-my-backyard' attitude to nuclear power plants [33]. ...
... More than one measure may seem reasonable. If a nuclear accident occurs that threatens a community, the stakeholders not only include the authorities and the media, but also the independent experts, the company that operates a nuclear plant, and all of the affected households [25]. Taken together, these considerations are assumed to shape protective action decisions. ...
Full-text available
Studies of the aftermath of nuclear power plant accidents show that affected citizens assess higher risks and adopt more risk-avoidant behaviors than authorities expect. This results in differences between the planned recovery and actual outcomes. Based on this knowledge, this study examined the factors that affect citizens’ preference to continue living in a decontaminated area. Testing the key aspects of the protective action decision model (PADM), this study analyzed Swedish survey data (N = 2291) regarding such an accident scenario. Several aspects of the PADM, from the layperson’s view of threats and protective actions, to stakeholders and situational factors, were strongly supported. The most influential variables affecting settlement choices are perceptions of radiation risk, perceptions of decontamination effectiveness, government information, living with certain restrictions, and attachment to an area because of one’s work. A novel contribution of this study is that it ranked the significance of such effects on behavioral intentions in an emergency scenario. Regarding the policy recommendations, this study concluded that a recovery program must facilitate most aspects of people’s lives and provide trustworthy information on decontamination efficiency. As some people will avoid potential health risks and leave a decontaminated area, planning to implement one solution for everyone would likely not be optimal.
... The complicated relationship between disaster evacuation decision-making and the information provided to and perceived by (potential) evacuees is a topic of significant interest in the domains of human behavior modeling, information processing, and evacuation planning (Murray-Tuite & Wolshon, 2013;Peers et al., 2021;Terpstra et al., 2009). Prior studies have established that information provided by public agencies can substantially influence the evacuation rates for hurricanes (Baker, 1979;Lindell & Perry, 2012;Sorensen, 2000) as well as other evacuation characteristics such as mobilization time (Sadri et al., 2013), departure time (Gehlot, Sadri, et al., 2019a), and travel mode (Murray-Tuite & Wolshon, 2013). Although householders consider their situational factors in making the final decision, the information they receive related to hurricane evacuation, especially evacuation notices, considerably increases their certainty of the decision-making process (Dash & Gladwin, 2007;Murray-Tuite & Wolshon, 2013). ...
... Research on information processing models relevant to disaster evacuation decision-making addresses general risk perception (Slovic, 1986), persuasion (Tormala, 2016), and crisis management (Egelhoff & Sen, 1992). Risk perception is considered an important facilitator of decision-making in hurricane evacuation as well as post-disaster return (Lindell & Perry, 2012;Riad et al., 1999). Many studies of information processing build upon the psychometric paradigm, which offers a framework for examining and measuring risk-related judgments and explaining why some risks are perceived as greater than others (Fischhoff et al., 1978;Slovic et al., 1986). ...
... The HSM provides the groundwork for two other relevant models addressing information processing. First, in the context of disasters, Lindell and Perry's protective action decision model (PADM) (Lindell & Perry, 2004, 2012, which uses some key principles of the HSM, provides a sound theoretical basis for modeling risk perception and decision-making. Studies following the PADM have observed that impacted individuals' previous experience with disasters and perceptions of their personal risk, along with the amount of disaster-related information they have, influence their information-seeking behaviors (Terpstra et al., 2009;Yang & Zhuang, 2020) and evacuation decision-making (Sarwar et al., 2018). ...
Conventional evacuation studies typically do not gauge the development of participants’ certainty about evacuation-related decisions with the updates in the information provided to them. This study uses an online survey that provides three kinds of progressively varied information about the current status of a hypothetical hurricane for five days leading to its landfall and collects respondents’ certainty of their situational comprehension and evacuation-related decisions each day. Most participants (84%) made a final decision (60% evacuate) after seeing information of just one day (four days before the landfall), indicating a tendency of swift decision-making. Modeling shows that the time spent looking at information, especially uncertainty cone forecast maps, positively influences the understanding of the hurricane’s status, which in turn helps in increasing the certainty of making evacuation-related decisions, with an increasing temporal effect. This study contributes to the understanding of the public perception of information and its association with evacuation-related decision-making.
... Local tsunamis from the Cascadia subduction zone (CSZ) are a significant threat to the US Pacific Northwest coast, so residents of coastal communities need to have accurate information about this hazard and appropriate protective actions (NAS 2011;NTHMP Review Committee 2017). One framework for assessing coastal residents' hazard knowledge and emergency preparedness is the protective action decision model (PADM-Lindell 2018;Lindell and Perry 1992, 2004, 2012. According to the PADM, protective action decisions begin with environmental cues, social cues, and social warnings. ...
... In response to the National Tsunami Hazard Mitigation Program's need to evaluate tsunami evacuation brochures being distributed in the five Pacific Coast states, we drafted a composite brochure that included the most important topics in those states' brochures. This composite brochure focused on topics emphasized in the PADM (Lindell 2018;Lindell and Perry 2012) and the RISP (Dunwoody and Griffin 2014) model, which identify the most important variables for evaluating coastal residents' tsunami hazard knowledge and beliefs about appropriate response actions if one strikes. ...
Full-text available
This study surveyed 227 residents in three US Pacific Coast communities that are vulnerable to a Cascadia subduction zone tsunami. In the Brochure condition, information was presented online, followed by questions about tsunamis. Respondents in the Comparison condition received the same questionnaire by mail but did not view the brochure. Respondents in the Brochure condition had higher levels of perceived information sufficiency than those in the Comparison condition about three of the five tsunami topics. Both conditions had generally realistic expectations about most tsunami warning sources. However, they had unrealistically high expectations of being warned of a local tsunami by social sources, such as route alerting, that could not be implemented before first wave arrival. They also had unrealistically high expectations being warned of a distant tsunami by ground shaking from the source earthquake, whose epicenter would be too far away for them to feel. Moreover, respondents in both conditions expected higher levels of personal property damage and family casualties than is the case for most hazards, but their levels of negative affective response were not especially high. Overall, only 10% of the sample accessed the tsunami brochure even when sent repeated contacts and the brochure demonstrated modest effects for those who did access it. These results suggest that state and local officials should engage in repeated personalized efforts to increase coastal communities’ tsunami emergency preparedness because distribution of tsunami brochures has only a modest effect on preparedness.
... Common factors identified as the key variables affecting homeowner mitigation decisions include property type and value, financial availability, mortgage situation, risk perception and attitude, previous experience with wildfire and other natural hazards, etc. For example, there is considerable evidence that people who have experienced natural disasters are more inclined toward taking risk mitigation actions (Peacock 2003;Ge et al. 2011;Lindell and Perry 2012). McGee and Russell (2003) suggested that several attributes of homeowners, including previous experiences with wildfires, involvement in agriculture and with the local fire brigade, and their social network, contributed to mitigation behavior, while wildfire preparedness within the community was affected by "a culture of self-reliance, experience with fires as part of farming, and social cohesion." ...
Full-text available
Wildfires have become increasing threats to residents, built environments, and ecosystems in the USA. Individual responsibility plays an important role in reducing structure-level ignitability, and in turn, overall community vulnerability. When wildfire cannot be completely prevented through risk reduction efforts, homeowners insurance can serve as the second line of defense by allowing homeowners to transfer risk. To understand homeowner decisions on wildfire-related proactive actions and the effects of such decisions on the housing recovery, this study conducted an online survey of homeowners living at high to extreme risk of wildfire in the Western United States and collected data related to two types of proactive actions: individual-level risk reduction actions and homeowners insurance. First, a regression model for each proactive action was estimated to identify key characteristics of homeowners and house/property that had the greatest impact. The results indicated that homeowner age and household income were the two common factors affecting their decisions about home hardening and insurance policies, while the only statistically significant factor in homeowner decisions about defensible space was satisfaction with the surrounding environment (e.g., scenic beauty, privacy). Moreover, the effects of each proactive action on the housing recovery process were evaluated. The results showed that home hardening was a more effective action in reducing wildfire damage to a house than defensible space was, which was consistent with homeowner perception. The survey results also indicated that homeowners with insurance were less likely to experience post-wildfire financial hardship, and subsequently were more likely to repair their damaged houses.
... Social norms are "a common behavior or practice" and "an average outcome or output standard" (Miller & Prentice, 2016, p. 240), which can impact individuals' perceptions of social expectations and behaviors (Ajzen, 2011;Cialdini, 2012). In the disaster warning literature, observing others' behaviors (i.e., social norms) is predictive of message compliance (Lindell & Perry, 2012;Wood et al., 2018). To date, researchers have not examined how a variety of social norms can be integrated into messages to motivate individuals' disaster mitigation behaviors. ...
Full-text available
Preparing for natural disasters and adapting to climate change can save lives. Yet, minimal research has examined how governments can motivate community members to prepare for disasters (e.g., purchasing flood insurance or installing water barriers in homes for floods and hurricanes). Instead, studies have focused on how to communicate actions individuals should take during disasters, rather than before disasters. This study develops messages targeting social norms, which are promising approaches to motivate community members to adopt disaster risk preparedness and mitigation behaviors. Specifically, we developed a variety of messages integrating descriptive norms (i.e., what others do), injunctive norms (i.e., what others believe should be done), and a social norms-based fear appeal, or social disapproval rationale (i.e., a negative social result of [not] taking behaviors). Then, we tested these messages through two between-subject factorial online experiments in flood- and hurricane-prone U.S. states with adult samples (N = 2,286). In experiment 1 (i.e., purchasing flood insurance), the injunctive norms message using weather forecasters and the social disapproval rationale message significantly increased social norms perceptions, which in turn influenced behavioral intentions. In experiment 2 (i.e., installing water barriers), the injunctive norms message using weather forecasters, the injunctive norms message using neighbors, and the social disapproval rationale message significantly increased social norms perceptions, which in turn influenced mitigation intentions. However, the descriptive social norms message was not effective in increasing social norms perceptions. We provide some of the first empirical evidence on how organizations’ risk communication can empower community members to prepare and mitigate the impact of disasters.
Full-text available
The frequency of extreme weather events has increased in recent decades due to climate change, and the demand for both more accurate weather forecasts and early warnings surges in risk management of weather-related disasters. However, the gap between weather forecast issuance and proper usage remains considerable, and we still see individual and household inactive to precautionary actions across lead times. A lack of effective risk management and the inaction of individuals urge us to review and inspect household usage of weather information and the decision-making processes, so as to find effective strategies and ensure smarter risk management. Using a survey designed by NOAA National Severe Storm Laboratories, USA, and conducting nationwide surveys in the USA and Taiwan, respectively, we aim to inspect the determinants of household evacuation decisions. Particularly, we hope to know whether biased risk perception, cognitive biases on risk information, and cultural biases exist, thereby causing inactive household evacuation. This research finds both Taiwanese and US respondents have problems of optimistic bias at earlier lead times, until they observe the hazards. They also easily suffer from cognition biases while determining whether to evacuate. Our research suggests real-time weather risk management is essential and we demonstrate the following: firstly, strategies that help correct biased risk perception, including education on weather forecast interpretation, training courses of probabilistic forecasts, customized formats of weather risk information, and AI assistance in real-time weather risk assessment, should be initiated and developed to inform risk situations for different demographic groups. Secondly, smart scientific risk communication and risk education should be accurately implemented to capture user lifestyle about risk management and meet local needs and communication preferences.
The United States experienced losses exceeding 305 billion USD in property damage and relief costs in 2017. While examining the impacts of disasters remains a pressing area of study, it is elemental to understand whether and how social and behavioral contexts relate to the ability to withstand a disaster. Informed by Albert Bandura's theory of self-efficacy and concepts of social vulnerability, we examine socio-cognitive and contextual factors associated with disaster preparedness in the U.S. through a quantitative analysis of cross-sectional data from the Federal Emergency Management Agency's 2018 National Household Surveys. Building on previous scholarship, disaster preparedness is the dependent variable operationalized as a cumulative score and as two indicator variables: adequate and minimal preparedness. Weighted descriptive statistics describe the sample and ordinary least squares and logistic regression analyses test the association of preparedness with socio-cognitive measures of disaster related efficacy. Response efficacy was significantly associated with preparedness across cumulative and adequate preparedness, but not with minimal preparedness. Confidence in carrying out action, on the other hand, was consistently associated with being better prepared across all three operationalizations. Women householders were less prepared than men overall. African American and Asian respondents were both less likely than their White counterparts to have taken the steps to be considered at least minimally prepared. For disaster policies to remain equitable, administrators and organizers need to ensure resources are devoted to communities that have been historically disenfranchised. Our findings advance knowledge that has the potential to inform policy, practice, and research on pre-disaster interventions.
Full-text available
Hazard risk perceptions and protective behaviors are examined for wildfires, earthquakes and volcanic activity. Data were gathered in two northern California (USA) communities that are exposed to all three hazard types. It was found that resident risk perceptions approximated the risks calculated by experts. Personal risks associated with fires were significantly lower than property risks associated with the same threat. The discrepancy between person and property risks for earthquakes and volcanic activity was much smaller. In general, it was found that the number of protective adjustments undertaken for each hazard was small (averaging about half of the possible number measured). When combined in a regression analysis, risk perception was not a statistically significant predictor of number of adjustments for any of the three hazards. Resident's sense of responsibility for self-protection and experience with property damage were significant predictors of adjustment for all three hazards. Information seeking behavior was significantly related to protective actions for earthquakes and volcanic activity, but not for fire hazards. In general, an insufficient number of residents reported experience with personal injury or harm to make meaningful assessments of the effect of this variable on adjustments.
This series is dedicated to serving the growing community of scholars and practitioners concerned with the principles and applications of environ­ mental management. Each volume is a thorough treatment of a specific topic of importance for proper management practices. A fundamental ob­ jective of these books is to help the reader discern and implement man's stewardship of our environment and the world's renewable resources. For we must strive to understand the relationship between man and nature, act to bring harmony to it, and nurture an environment that is both stable and productive. These objectives have often eluded us because the pursuit of other in­ dividual and societal goals has diverted us from a course of living in balance with the environment. At times, therefore, the environmental manager may have to exert restrictive control, which is usually best applied to man, not nature. Attempts to alter or harness nature have often failed or backfired, as exemplified by the results of imprudent use of herbicides, fertilizers, water, and other agents. Each book in this series will shed light on the fundamental and applied aspects of environmental management. It is hoped that each will help solve a practical and serious environmental problem. Robert S. DeSanto East Lyme, Connecticut Acknowledgments Compilation of the materials reviewed in this inventory was facilitated greatly by several staff members of the Disaster Research Center, University of Delaware (formerly at The Ohio State University) and the Natural Haz­ ards Research and Applications Information Center, University of Colorado.
Hurricane Katrina was an extraordinarily powerful and deadly hurricane that created catastrophic damage and inflicted large loss of life. The complex genesis of Katrina involved the interaction of a tropical wave, the middle tropospheric remnants of Tropical Depression Ten, and an upper tropospheric trough. Katrina generated an intense burst of deep convection over the low level center during the afternoon of August 25, 2005, while positioned over the northwestern Bahamas. It made it first appearance in the US as a Category 1 hurricane on the Saffir-Simpson Hurricane scale, with maximum sustained winds of 70 knots, near the border of Miami-Dade County and Broward County. A precise measurement of the storm surge produced by Katrina along the northern Gulf Coast is complicated by many factors, including the widespread failures of tide gauges. Estimates of the insured property losses by Katrina ranges between $20 billion and $60 billion.