“What Motivates Homeowners to Protect Themselves
from Wildfire Risks in the WUI?”
Holly Bender1
Integrated Resource Solutions, LLC
Ingrid M. Martin
California State University- Long Beach
and
Carol Raish
USDA Forest Service
Rocky Mountain Research Station
DO NOT QUOTE WITHOUT AUTHORS PERMISSION.
1
Contact author: Ingrid M. Martin, Associate Professor of Marketing, Dept. of Marketing, College of
Business Administration, 1250 Bellflower Blvd., California State University-Long Beach, Long Beach, CA
90840. Telephone: 562.985.4767. Email: imartin@csulb.edu.
“What Motivates Homeowners to Protect Themselves from Risks?”
The question of why individuals choose to mitigate risk or downplay and ignore
risk has been a topic of much research over the last 25 years in areas as diverse as
smoking, earthquakes, contraceptive use, alcohol consumption, flooding, and many other
hazards, both natural and human. Wildfire has been a relatively recent focus in the natural
hazards literature, with the last four years of catastrophic fires in the western United
States. The large population migration into this part of the country as well as the desire
of many to live in an area that provides wildland amenities has exacerbated the risks of
large-scale, catastrophic wildfires. This migration has resulted in more people living in
the wildland-urban interface (WUI), which has created many unique problems for
homeowners as well as land managers. The risks of the WUI along with the all the
experiences and biases that homeowners bring to the area have created a high risk
environment. Significant research has demonstrated that a home’s exterior materials and
its immediate surroundings principally determine the home’s ignition potential during
extreme wildfire events (Cohen, 2004). Additionally, the area that determines the home
ignition zone during extreme wildfires largely occurs on private lands (Cohen, 2004).
Yet, many homeowners do not undertake mitigating actions to protect their homes and
potentially their lives from the risks of wildfire.
In this chapter we investigate the cognitive perceptual process that homeowners
experience when faced with the decision to protect themselves from the risk of wildfires.
This decision can be explained through two complementary theories – Protection
Motivation Theory (PMT) and stage theories such as Transtheoretical Model (TTM).
PMT states that some form of risk information can provide the impetus for an individual
to determine the degree of risk severity, their vulnerability, and their ability to reduce that
risk. Stage theories categorize the decision making stages that individuals go through
when exposed to risk information – whether through actual experience with the risk or
through indirect experience (e.g. brochures, videos). We integrated these two theories to
investigate when homeowners are more or less likely to reduce the probability of being
impacted by the risk of wildfire (Block and Keller 1998; Weinstein and Sandman 1992).
To further enhance our understanding of human response to wildfire risk and
effective risk communication, we explore how PMT and TTM can be combined to
2
explain homeowners’ behavior. A discussion of these two theories and some related
theories provides the foundation for a set of hypotheses. This is followed by a description
of the methodology used to test the hypotheses, as well as the data analyses and results.
After a discussion of the results as they relate to our theoretical conceptualization, we
present a set of unanswered questions to provide a basis for a research agenda. Finally,
we discuss the managerial and public policy implications.
Protection Motivation Theory
Protection motivation theory states that stakeholders’ motivations or intentions to
protect themselves from harm are enhanced by four critical cognitions or perceptions,
regarding the severity of the risks, the personal vulnerability to the risks, self-efficacy at
performing the risk-reducing behavior, and the response efficacy of the risk-reduction
behavior (Rogers 1983). It also posits that people’s intentions to protect themselves are
weakened by the perceived costs of the risk-reducing behaviors and the perceived
benefits of the alternative risk-enhancing behaviors. PMT is organized as two mediating
subprocesses that consumers use in evaluating threats (threat-appraisal process) and in
selecting among coping alternatives (coping appraisal). Assessments of threats (severity,
vulnerability, and benefits) and coping factors (self efficacy, response efficacy, and costs)
combine to form a motivation in stakeholders to protect themselves from the risk.
Protection motivation arouses, sustains, and directs activities starting with the
identification of a natural hazard such as wildfire in the WUI (Mulilis and Lippa 1990,
Rustemli and Karanci 1999).
According to PMT, people can be motivated to engage in desirable health
behaviors not only to avoid health risks but also to avoid social or interpersonal risks
(Pechmann, Zhao, Goldberg, Reibling 2003). The research in PMT has focused on the
impact of health information (e.g., antismoking messages) on the elicitation of both the
appraisal of the threat and of the coping techniques (e.g. Mulilis and Duval 1995,
Rippetoe & Rogers 1987, Floyd, Prentice-Dunn, & Rogers 2000). People use the threatappraisal process to evaluate potentially harmful responses such as not doing anything to
protect yourself and your home from the risk of wildfire (maladaptive response). The
3
intrinsic and extrinsic rewards or benefits (e.g. not to protect yourself or others) will
increase the probability of the maladaptive response whereas threat will decrease the
probability of selecting the maladaptive response (e.g. it costs too much to engage in the
needed protective behaviors). In contrast, the severity of the risk and personal
vulnerability to the danger will decrease the probability of the maladaptive response
while the response cost will increase the probability of selecting the adaptive response
(e.g., creating defensible space around one’s home). This theory has been used to explain
the effects of risk communications on health and natural hazards (e.g. Block and Keller
1998, Mulilis and Lippa 1990, Pechmann, et al. 2003, Weinstein 1989).
In the case of wildfire risks, the likelihood of an individual being impacted by the
risk of wildfire is decreased by (a) belief in the severity of the wildfire risk, (b) belief in
one’s vulnerability to the wildfire risk – both physical property and health, (c) belief that
defensible space behaviors are an effective way to avoid the risk of wildfire, and (d)
belief that one can successfully avoid the risk of wildfire. In contrast, the likelihood of
not doing anything about the risk of wildfire is increased by intrinsic rewards (e.g.
enjoying your trees and privacy), extrinsic rewards (e.g. peer approval – neighbors who
believe that same way as you do about the beauty of the forest), and the costs of an
adaptive response (e.g. the costs of creating and maintaining defensible space on your
property). The increased likelihood of an adaptive response (creating defensible space)
depends primarily on the four cognitive perceptions of severity, vulnerability, selfefficacy and response efficacy. The degree of these cognitive perceptions can enhance the
persuasive effects of a risk communication strategy that elicits protection motivation. By
eliciting a desire to protect oneself and one’s property, the risk information can then
arouse, sustain, and direct activities for self-protection such as increasing the number of
defensible space behaviors, working to reduce some of the involuntary risks created by
neighbors, and maintaining these risk mitigating behaviors over time. Strong beliefs that
severity, vulnerability, self efficacy and response efficacy will arouse the motivation to
protect oneself and one’s property should result in a change in their adoption rate of risk
reduction behaviors.
Varying Levels of Readiness to Mitigate Risks
4
The protection motivation theory does not assume that the decision makers are
rational. Each of the cognitive appraisal processes can be biased by a heuristic judgment.
In the psychological risk paradigm, researchers have demonstrated that people use mental
heuristics to evaluate risk (e.g., Slovic, 1987; Tversky and Kahneman 1973; Viscusi
1990). This research has demonstrated repeatedly the difficulties that individuals have in
understanding probability, biased media coverage, misleading personal experiences, and
more, leading to the overestimation and underestimation of risk perceptions. Slovic
(1987) reports two constructs that appear to be highly correlated with the level of
perceived risk: 1) dread risk—characterized by perceived lack of control, catastrophic
and fatal consequences, inequity in the distribution of costs and benefits, involuntariness;
and 2) unknown risk—characterized by not being observable, risks unknown to science,
the effect of timing is unknown and the effect of outcome delayed. The risks of wildfire
can fall in both the dimension of dread risk and unknown risk. As Slovic (1987) points
out, different publics with different experiences and characteristics will have different
perceptions of risks, which will result in unique beliefs about risk reduction behaviors.
This research has brought attention to the multiple dimensions of risk, and that
individuals respond differently depending on the type of risk, features of the hazard, and
the characteristics of the individual. We introduce a potential factor that may explain the
varying motivations for people to reduce their risks. The degree of readiness to accept
and act on a risk can impact their particular motivational level to protect themselves and
their property. Stage theories have been used to examine health behavior change based on
the assumption that predictor variables will influence different people in different ways
(Horwath 1999). These theories specify an ordered set of categories into which people
can be classified. Based on this categorization, one can identify the factors (e.g.
vulnerability, risk severity) that can induce movement from one stage to the next
(Weinstein, Rothman, and Sutton 1998). The basic premise is that people can be
distinguished based on those who have not yet decided to change their behavior, those
who have decided to change, and those already performing the new behavior. The TTM
proposes six decision making stages that an individual faces when exposed to a health
risk message – precontemplation, contemplation, preparation, action, maintenance, and
5
termination (Prochaska, et al.1994). People are assigned to one of the six stages on the
basis of their behavior and intentions for future actions (e.g. Proschaska, et al. 1994).
Numerous health behavior theories such as PMT and the Health Belief Model
(HBM) specify factors (e.g. response efficacy, vulnerability) which produce transition
between the stages. For example, Block and Keller (1998) found that people are at
different stages of readiness to undertake risk reduction behaviors, which can in turn be
affected by the four cognitive processes in PMT. They integrated three of the six stages
of TTM into the PMT model to provide a more effective way of understanding people’s
responses to risk and risk communication. They found that the degree of risk severity,
vulnerability, self-efficacy and response efficacy are critical motivators as people move
through the stages of precontemplation, contemplation, and action.
Individuals in the precontemplation stage (precontemplators) are resistant to
change and often either do not see or will underestimate the risk of a particular hazard.
For example, a review of the hazard risk perception literature found that overall people
are optimistic about the risks (i.e., the risks are perceived to be lower than they actually
are) (Weinstein, 1987). A number of studies have indicated that residents tend to deny
the risk of natural disasters by discounting the possibility that anything serious will ever
happen to themselves (Burton and Kates, 1964; Mileti et al. 1975). Additionally, people
may be reluctant to acknowledge higher perceived risks if the risk provokes feelings of
anxiety (i.e., the uncontrollability of the risks), thereby denying the existence of the risk
(Lazarus and Folkman 1984). In this stage, homeowners living in the WUI must first
accept that they are vulnerable to the risk of wildfire before they can consider engaging in
any risk reduction behaviors. Thus, at this level of readiness, the focus should be to
increase their perceived vulnerability to the risk of wildfire.
H1: Increasing perceptions of vulnerability will lead to greater intentions to perform risk
reduction behaviors among precontemplators rather than the focus on risk severity,
response efficacy, and self-efficacy.
Individuals in the contemplation stage (contemplatives) have moved forward to
acknowledge that there is a problem but they do not have concrete plans of action in the
next six months or so. This group believes that they are vulnerable to the risk of wildfire.
6
Thus, they are considering risk reduction behaviors. Block and Keller (1998) believe that
consumers in this stage of readiness need a risk communication message that emphasizes
the severity of the hazard outcome to motivate them to take mitigating actions. Research
in health communication focuses on a particular category of risks such as immunization,
sun exposure, breast checks, etc. Additionally, in the natural hazards area, risk, such as
wildfire, have different characteristics that have implications for the means by which
contemplatives will perform recommended behaviors. For example, the hazard literature
indicates that the link between risk perceptions and preparedness behavior is at best,
tenuous (McCaffrey 2004; Tierney 1993). It could be that high awareness levels are
necessary for action, but that other factors are also important in motivating homeowners
toward action. Possibly, increasing homeowners’ feelings of vulnerability along with
increased levels of risk severity will motivate them toward taking more action to protect
themselves. For example, in Mileti and Sorenson (1987), respondents may have been in
the contemplation stage as high perceptions of vulnerability and risk severity from a
natural hazard increased their likelihood of mitigating the risk. We posit that for
homeowners contemplating taking risk mitigating actions, the combination of both
perceptions of vulnerability and severity are needed to motivate them to take action to
protect themselves.
H2: Increasing perceptions of vulnerability and severity will lead to greater intentions to
perform risk reduction behaviors among contemplatives rather than a focus on selfefficacy and response efficacy.
Finally, those in the action stage are in the process or have already taken action to
reduce their risk through specific behaviors. Past research in the health behavior literature
found that self-efficacy increased significantly in the action stage (Rossi, et al. 1994;
Sporny and Contento 1995).Thus, individuals need the confidence to undertake the
critical risk reduction behaviors along with a belief that these risk reduction behaviors
will be effective (self efficacy and response efficacy). This also implies that varying
levels of personal vulnerability and risk severity will not differentially affect
homeowners’ need to protect themselves. At this point, homeowners need the confidence
7
and belief that they can successfully undertake the most effective risk reduction
behaviors.
H3: Increasing perceptions of self-efficacy and response efficacy will lead to greater risk
reduction behaviors in the action stage rather than a focus on perceptions of vulnerability
or severity.
It is possible that for the action homeowners, increasing perceptions of
vulnerability and risk severity perceptions could actually lead to a ‘boomerang effect’
(Pechmann, et al. 2003). In this situation, homeowners who are confronted with
extremely high perceptions of vulnerability and severity may instead revert to denial thus,
reversing their behaviors to mitigate the risks. It becomes important to ensure that the
risks are placed in a context that allows people to feel effective enough to be able to
reduce the risk that they face.
Next, we discuss the methodology used to provide some preliminary support for
the application of PMT and TTM to the context of wildfire in the WUI. This is followed
by the analysis and results of the data collected for this preliminary study. Then we
discuss the questions that still need to be answered concerning why homeowners do or
don’t do what they need to do to mitigate the risks of wildfire. We end with a discussion
of where we go from here in this stream of research.
Methodology
We targeted three communities in the western U.S. as varying types of groups that
have faced recurring wildfires. Two of the three communities were targeted for their
extensive experience with large-scale fires while the third community had a more distant
and not as severe history with wildfires. The homeowner association lists were obtained
for all three communities. All residents of the communities (both part-time and full-time
residents) were sent a survey that included the measures described below. The mail
survey was sent out with a cover letter explaining the project and included a selfaddressed, stamped envelope for residents to return the completed survey. A reminder
postcard was sent out about three weeks after the first mailing. Prior to sending out the
8
survey, we conducted focus groups in each community with a subset of the residents, at
least one special interest group, and representatives from the district ranger’s office.
The purpose for the focus group interviews was to get an in-depth understanding
of the particular characteristics of the communities, the wildfire(s) that had affected the
area and any other issues or problems that were specific to the community. The
participants in the focus groups were selected because they were seen as more active and
knowledgeable in terms of community history, wildfire history, etc. This provided a
more in-depth understanding of the community. In addition, the participants in these
focus groups were also given the survey to fill out. These respondents were compared to
the mail survey respondents to determine if there was any non-response bias (Alreck and
Settle 1995). We compared each set of focus group survey participants with the
respondents to the mail survey in their respective communities. Based on a set of t-tests
on a subset of the risk perception, self efficacy and response efficacy measures we found
that there was no difference between these two formats.
Sample. The sources of information initiating the cognitive mediating processes
can be environmental or intrapersonal sources. Prior research has used both
environmental factors as well as health communication sources to affect the protection
motivation process and initiate behavioral change (e.g. Rogers 1983, Floyd, et al. 1990;
Pechmann, et al. 2003). In our research, we look at three different fire-affected
communities and use their varied levels of readiness to act on reducing their risks as the
primary factor that affects their motivation to protect themselves from the risks with the
resulting change in behavior. In addition, we also investigate the subjective knowledge
that potentially guides this group of homeowners through the protection motivation
process. All residents in the three communities were sent a mail survey and one reminder
postcard. In total we ended up with 237 out of 423 completed and usable surveys from all
three communities with a response rate of 56%.
Three locations were chosen based on their extensive experience with wildfire.
They included a small community in central Colorado, northern Colorado and central
Oregon. Next, we present the description of each of these communities.
Central Colorado. The first location was in central Colorado and is adjacent to
and surrounded by the Pike National Forest. In June 2002, the Hayman fire burned over
9
138,000 acres north and west of this community. Almost 70% of the homeowners were
evacuated in the Hayman fire when the fire came to within five miles of the community.
The Schoonover fire in May of 2002 was also within a few miles of this community
burning 4000 acres. In the same year, two additional fires, Snaking Gulch fire and Black
Mountain fire, burned approximately 2800 acres in April and May and caused the
evacuation of 3500 homes in a community about 40 miles northwest of the community of
interest. Other fires in previous years have also significantly affected this community
including the Hi Meadow fire in June 2000, the Big Turkey fire in May 1998, and the
Buffalo Creek fire in May of 1996. The Big Turkey fire was within five miles of this
community, while the Buffalo Creek fire was located approximately 20 miles north. This
community in central Colorado has had significant experience with wildfire, including
evacuation, smoke impacts, and frequent exposure. The central Colorado community is
comprised of only full time residents with the majority living in the area for over 10
years. Approximately a quarter of the residents moved into the area less than five years
ago. The homeowners are 35 years and older with over 55% being over 55 years of age,
making it primarily a retirement community. This group is highly educated with 65%
having at least of college degree and the community is fairly affluent with 85% citing a
yearly household income of over $50,000.
Northern Colorado. The second location was in northern Colorado in a
community that is also completely surrounded by the Roosevelt National Forest. In
2002, the Big Elk Fire burned 4400 acres about 15 miles south of this community.
Although many homes were evacuated, this community was not evacuated. In 2000, the
Bobcat Gulch Fire burned 10,600 acres and destroyed 22 structures within approximately
10 miles from this community (although they were not evacuated at this point either). In
1993, the Snowtop Fire burned a couple hundred acres about 10 miles from this
community. Although these homeowners have seen a few fires in the past 10 years, none
of the fires warranted the evacuation of the communities. This northern Colorado
community represents an emerging retirement area, with 63% of the homeowners being
over the age of 55. Sixty percent of the respondents from this community were male,
with 82 percent having attained at least a college degree. This community is also fairly
affluent, with 62 percent earning more than $50,000 per year.
10
Central Oregon. The third community is in central Oregon surrounded by the
Deschutes National Forest.
In August and September, 2003 this community was
evacuated twice during the B&B Complex fire (includes both the Bear Butte and Booth
Fires). This fire burned over 90,000 acres north and west of the community. One report
stated that the fire burned “within feet of dozens of houses”. Almost 90% of the
homeowners who responded to our survey were evacuated in the B&B Complex fire
when the fire came to within 5 miles of their property. The Link Fire also occurred in
2003 very near to this community. Two fires burned in July, 2002: the Cache Mountain
Fire destroyed 2 homes in a subdivision a few miles from this community and 13,000
homes were evacuated; and the Eyerly Fire destroyed 37 structures within 15 miles of
this community. In 1996, the Skeleton Fire destroyed 30 homes about 40 miles southeast
of this community. The Oregon homeowners are also an older sample, with 60 percent
over 55 years of age. They are highly educated with 75% of the respondents having at
least a college degree. Approximately 50 % of the homeowners are part time residents,
and 58% are male. The community is fairly affluent with 57% citing a yearly household
income of over $75,000.
Independent Measures. In order to determine the effect of the cognitive
processes on behaviors and behavioral intentions, we measured homeowners’ risk
perceptions, response and self efficacy, and decision making stages.
Risk Perception Measures. In protection motivation theory risk severity and risk
vulnerability are the critical risk perception measures. Risk severity is defined as the
amount of hardship that would occur if one experienced the risk. We measured risk
severity using two seven-point likert scales by asking “how serious do you feel the
negative consequences of wildfires are to you personally/how severe will the impact of a
wildfire be where you live” anchored by 1=not at all serious/no harm at all to
7=extremely serious/extremely devastating (α=.87). Risk vulnerability is defined as the
likelihood of harm if there is no change in behavior and is measured using two sevenpoint likert scales by asking “how vulnerable do you feel about the possibility of a
wildfire physically affecting you or your family/your property or possession” anchored
by 1=not at all vulnerable to 7=extremely vulnerable (α=.78).
11
Response Efficacy Measures. Response efficacy is defined as the degree to
which a proposed risk mitigation behavior is effective at reducing a particular risk.
Respondents were asked to rate the effectiveness of the risk reduction behaviors using the
following method. They were asked to rate “how effective are the following actions at
helping to reduce the risk of fire impacting your property and lives” followed by the list
of the 11 risk reduction behaviors and each item was anchored by 1=not at all effective
and 7=very effective. In addition, they were asked to evaluate “by doing any of the risk
reduction behaviors on your property, how effective can you be at preventing wildfires
from impacting your personal property and your life, anchored by 1=not at all possible to
7=very possible.
Self-Efficacy Measures. Self-efficacy is defined as the belief that one is capable
or not capable of performing a risk mitigation behavior (Bandura 1977). The measures of
self-efficacy were developed based on measuring how confident respondents felt about
their ability to protect their property and themselves from the risk of wildfire.
Respondents were asked “how confident do you feel in your ability to do the following
(risk reduction behaviors)” as well as “how confident do you feel in general about your
ability to protect yourself and your property from wildfire”, anchored by 1=not at all
confident to 7=very confident. In addition, respondents were asked to rate their
confidence in their ability to protect their property and themselves from the risk of
wildfire anchored by 1=not at all confident to 7=very confident.
Decision Stages. The stage at which each homeowner was categorized was based
on the number of risk reduction behaviors that they undertook to protect themselves.
There were a total of 11 potential risk reduction behaviors that homeowners could
undertake (see Appendix). Each behavior could be selected based on whether they had
already undertaken the action or were considering undertaking it in the future or were not
considering undertaking the action for a particular reason. A person was categorized as
being in the “Action” stage if they had already completed 7 or more behaviors. A person
was categorized as “Precontemplative” if they have 5 or more behaviors that they
answered as “probably will not do”. All those who did not fall into the “Action” or the
“Precontemplative” categories were classified as “Contemplatives”.
12
Additional Measures. Some additional measures were taken to be able to get a
clearer picture of what may motivate people to protect or not protect themselves against
the risk of wildfire. These measures were designed to provide further support for the
results of our initial analyses. These included measures of the biggest impediment to
protecting oneself (response costs) and demographic information. We asked respondents
to explain what the biggest impediment that they faced when deciding to take some
protective action on their property. This was an open-ended question that allowed
respondents to write down one or more reasons that they felt were critical in determining
whether to protect their property. Two researchers coded the responses into categories
independently. The coders were blind to our hypotheses and the result was a set of 18
categories of cited impediments. Finally, we ended the survey with a set of demographic
questions which asked respondents to identify their age, gender, highest level of
education, their primary residence, how long they had lived at their primary residence,
and their income level. To determine if respondents had a second home in the respective
communities, they were asked to state how much time they spent at the part-time
residence and how long they have lived there. This information is included in the
descriptions above of each of the three communities.
Dependent Measure. The dependent measure of interest in PMT and our
research is behavioral intention or the amount of risk reduction behaviors that
homeowners engage in to protect themselves from the risk of wildfire.
Behavioral Intention and Actual Behavior Measures. The primary measure of
adaptive coping was the intentions of homeowners to change from not undertaking any
risk reduction behaviors to undertaking a set of risk mitigation behaviors. We measured
the behavioral intentions and actions that respondents could take to protect themselves
against the risk of wildfire. There were 11 defensible space actions that were taken from
the Firesafe Council of California (www.firesafecouncil.org) (see Appendix for the
complete list of behaviors). Each behavior was measured using the following scale: 1=
already done, 2=will do in next month, 3= will do in next 3-6 months, 4= will do next
year, and 5= probably will not do. For the last option, respondents were asked to explain
why they would not take that action to protect their property. The 11 measures were
combined into a composite measure of behaviors for the purpose of this analysis.
13
Analyses and Results
Two hundred and thirty-seven homeowners in the three communities participated
in the experiment. To test H1 through H3, respondents were categorized into one of the
three decision stages based on the categorization scheme described above. We will
discuss the results of the analysis next.
Hypotheses Testing (H1 through H3). To evaluate the hypotheses we tested a set
of dummy variable regressions for each decision stage (see Table One). We used
vulnerability, severity, self-efficacy and response efficacy as the independent variables
and a composite measure of the behaviors as the dependent variable. As seen in Table
One, the result for the first regression for precontemplators demonstrates that the only
significant predictor of behavioral intentions (coping response) was perceived
vulnerability consistent with H1. For the second regression for contemplatives, there was
support for H2 with the significant predictors of behavioral intentions (coping response)
being perceived vulnerability and perceived risk severity. For both precontemplatives
and contemplatives, the increasing feelings of vulnerability led to greater intentions to
engage in risk reduction behaviors. What differentiated precontemplatives from
contemplatives was the importance that risk severity has with the latter group. Consistent
with H3, the significant predictors of self- and response efficacy confirm their influence
on behavioral intention (coping response) for the action homeowners (see Table One).
The unexpected result for the action phase was that risk severity was also significant.
Overall, these results demonstrate that the three groups of homeowners are differentiated
based on what is critical to move them from one decision stage to the next decision stage.
In addition, we ran an overall regression equation using behavioral intention as
the dependent variable and the same four independent variables. The regression analysis
indicated significant beta coefficients for all four predictor variables (last column of
Table One). These results are consistent with previous studies in PMT (e.g., Pechmann, et
al. 2003) although they do not follow a similar pattern that some studies in PMT have
displayed (e.g., Block & Keller 1998; Maddux & Rogers 1983). For example, Block and
Keller (1998) found that only perceived vulnerability and self efficacy were significant in
the overall regression equation.
14
To get a better understanding of why homeowners in different decision stages
were differentially likely or unlikely to undertake risk reduction strategies, we asked
them to state factors that affected their decision on whether or not to undertake risk
mitigation actions (see Table Two).
TABLE ONE
Beta Coefficients and P-values for Behavioral Intentions Regressed on
Vulnerability, Severity, Self-Efficacy and Response Efficacy
________________________________________________________________________
Decision Stage
______________________________________________________
Precontemplatives
Contemplatives
Action
Overall
________________________________________________________________________
Vulnerability
9.37**
18.94**
2.14
29.26**
Severity
1.23
3.38*
13.96**
23.45**
**
Self-Efficacy
3.24
2.85
12.48
24.02**
**
Response Efficacy
0.27
0.02
5.51
21.11**
N
58
85
94
237
R2
.21
.24
.28
.30
________________________________________________________________________
Note: *p<.01 and **p<.001. Degrees of freedom: 4, 232 for the overall model and 4, 57; 4, 84; 4,93 for each
of the main effects, respectively.
TABLE TWO
Biggest Impediment By Decision Stage (number of times this response was stated)
________________________________________________________________________
Decision Stage
Precontemplatives Contemplatives
1. Cost
14
2. Time
3
3. Others Pose Risks
12
4. Only So Much One Can Do
6
5. FS Mismanagement Poses Risks 12
6. Aesthetics
7
7. Wildlife is Uncontrollable
6
8. No Impediments
9
9. Mother Nature
7
10. Age and Physical Ability
5
11. Vacation Home
1
8
11
2
5
2
1
3
2
2
3
3
Action
Total
12
6
5
6
2
7
3
1
1
1
3
34
20
19
17
16
15
12
12
10
9
7
The most notable difference in the biggest impediment themes between the action
and precontemplative stages is the aspect of involuntary risk (see Table Two).
15
Homeowners in the precontemplative decision stage were more apt to note the
involuntary aspects of risk as impediments for personal protective action. The
impediments that were different for the two decision stages were 1) campers and
neighbors pose risks, 2) Forest Service mismanagement poses the risk, 3) mother nature
poses risk, 4) and wildfire is uncontrollable. This is consistent with the PMT model, as
homeowners in the precontemplative stage are not motivated by efficacy measures; these
homeowners believe that external, involuntary factors are driving the majority of the
wildfire risk, and that their actions will not significantly reduce the risks of wildfire
affecting their homes and lives. These homeowners need to understand the vulnerability
of their homes and lives before they can make the link to the effectiveness of the
protective actions and confidence in their ability to carry out the actions.
To confirm if PMT integrates with the Transtheoretical model, a trend analysis
was performed to determine if a positive linear trend existed for the four predictor
variables. This would verify the progression from the precontemplative to the
contemplative to the action decision stages. The results demonstrate that there is a
significant positive linear trend for all four predictor variables. Consistent with Block &
Keller (1998), we found further support for the positive linear trend for vulnerability and
self-efficacy. In contrast to their findings, though, we also found a significant positive
linear trend for severity and response efficacy. This allows us to conclude that
homeowners move from the precontemplation stage to the action stage as their perceived
vulnerability, severity, self-efficacy and response efficacy increase. These results signify
that actors have the most confidence in their ability and the means by which they need to
protect themselves while precontemplators have the least confidence. In addition, actors
have the highest level of perceived vulnerability and severity and the precontemplators
have the least. A closer investigation of Table Three also shows that actors have the
highest level of perceived vulnerability and severity even though they are engaged in the
most risk reduction behaviors.
TABLE THREE
Means, Standard Deviations and F-Statistics for Linear Trend
For PMT Variables Across the Four Decision Stages
________________________________________________________________________
Decision Stage
______________________________________________________
16
Precontemplatives
Contemplatives
Action
Linear Term
F-Statistic
________________________________________________________________________
Vulnerability
4.63
5.22
5.59
F=3.62*
(1.52)
(1.51)
(1.40)
Severity
4.68
5.43
5.50
F=2.84*
(1.42)
(1.46)
(1.19)
Self-Efficacy
5.21
5.7
6.02
F=2.76*
(1.01)
(.90)
(.73)
Response Efficacy
5.11
5.69
6.22
F=2.40*
(1.02)
(1.04)
(.63)
N
58
85
94
_______________________________________________________________________
Note: *p<.01
In conclusion, this research supports the work of numerous researchers to
integrate protective motivation theories into stage theories to develop a more processoriented understanding of why some individuals attend to risks and others discount risk
(Block & Keller 1998; Weinstein, et al. 1998). Our results confirm that homeowners who
are in the precontemplative stage are motivated by perceived vulnerability, which leads
them to contemplate potential further action if given the right type of ‘vulnerabilitypromoting’ information. Likewise, those contemplating whether they should change their
behavior to mitigate the risk are motivated primarily by perceptions of risk severity. This
is partially consistent with the extant literature which finds that contemplators were more
affected by risk severity than perceived vulnerability. Given that contemplators’
perceptions of vulnerability and risk severity are significantly higher than those of the
precontemplators, this group has already undertaken more actions than the
precontemplative stage, and they are more apt to take additional precautions if given
effective ‘risk severity-promoting’ information to facilitate that decision. An additional
difference between the contemplators and the other two groups is related to their
determination that “time” is the biggest impediment to whether they will undertake an
risk mitigating behavior. Finally, the precontemplatives often attribute their inaction to
involuntary risks that are beyond their control (see Table Two). These homeowners need
to understand that effective risk mitigation strategies must include multiple public and
private landowners, and that homeowners’ actions are integral part of reducing the risks.
Our model, which provides additional support for the literature, would indicate that
17
increasing these homeowners’ perceptions of vulnerability will increase their self
protective behaviors.
Finally, homeowners currently engaged in risk reduction behaviors to some
degree already believe that they are vulnerable to the negative consequences and that
these negative consequences are severe enough to encourage them to act to protect
themselves. For action homeowners, increasing the perceptions of self-efficacy and
response efficacy rather than focusing on perceptions of vulnerability or severity will lead
to greater risk reduction behaviors.
Additional Considerations and Some Unanswered Questions
We have alluded to the importance of risk communication throughout this
chapter. Now we will discuss this in the context of future research as well as public
policy implications. It is important to understand how risk communication can and
should be tested in the integrative PMT-TTM approach to take this research the next step.
By understanding how to communicate in an effective way to different segments of
homeowners (precontemplatives, contemplatives, actors), we can help provide solutions
to improving the probability that each segment will engage in risk mitigating behaviors.
Traditionally, the health communication literature has used the PMT-TTM
through a pre-posttest experimental design framework to examine the effects of various
types of information on risk reduction behaviors. This study has demonstrated that
homeowners at different levels of readiness respond to and are motivated by a series of
cognitive processes. To extend the findings in this study, the next step would be to
introduce risk information that addresses these cognitive processes. We would predict
that the risk information that focuses on increasing perceptions of vulnerability would be
effective at moving precontemplatives into the contemplative stage. Next, information
that focuses on vulnerability and risk severity would move contemplatives to the action
stage. Finally, providing information on the means by which homeowners could reduce
their risk would result in action homeowners undertaking more mitigating behaviors..
This research focused on three of the six TTM decision stages: precontemplatives,
contemplatives, and actors. The complete Transtheoretical model proposes six decision
18
making stages that an individual faces when exposed to a health risk message –
precontemplation, contemplation, preparation, action, maintenance, and termination
(Prochaska, Norcross, and DiClemente 1994). This research needs to extend the
segments of homeowners to include an examination of the two of the three remaining
decision stages - preparation and maintenance. Termination is the final stage of this
model but given the ongoing nature of the risk of wildfires and living in the WUI, this is
not a meaningful decision stage for homeowners. Investigating homeowners’
understanding of and motivation for maintaining their defensible space actions over time
(i.e., the maintenance decision stage) would be an obvious extension of this research.
The impact of subjective knowledge on the PMT model needs to be investigated
to address the prevalent belief in the natural hazards and wildfire risk literature that
“people are knowledgeable about the high risks of wildfire but still they do little or
nothing to mitigate it”. Homeowners’ subjective knowledge (i.e., self rated knowledge
level) is based on their respective direct experience (whether they were evacuated, lost
property, etc. due to a fire) and indirect experience (e.g. reading information, experts,
word-of-mouth) with a particular category of interest such as the risk of wildfire (Alba &
Hutchinson 1987). Personal experience can have a powerful impact on the recognition of
risk and the willingness to protect oneself from the risk (Weinstein 1989). An
understanding of homeowners’ personal experiences can provide input into effective risk
communication. Insight into the factors that motivate stakeholders to protect themselves
against risks is dependent on an individual’s direct and indirect experience, which is often
the basis for their subjective knowledge on the topic. These past experiences become the
foundation for individuals’ beliefs in their own knowledge about a risk (Alba &
Hutchinson 1987). The extant literature in knowledge has consistently demonstrated that
experts have much greater awareness and knowledge about a particular risk and
alternative strategies to deal with the risk (Mitchell & Dacin 1996). In contrast, less
knowledgeable consumers tend to recall less information related to the risk and their
knowledge tends to be episodic (e.g. causal relationships between defensible space
actions and reducing their vulnerability and the severity of the risk are missing).
Additionally, a significant number of the homeowners in our sample are part-time
or “second home” residents. In our focus groups, discussions with primarily full-time
19
homeowners resulted in a common complaint about the part-time residents. These
“second home” residents were viewed as creating much of the involuntary risk that fulltime residents face as they were seen as less likely to undertake mitigating actions on
their properties, thus putting the fulltime residents at a greater risk (Integrated Resource
Solutions 2004). Further, many of the communities at risk from the impact to wildfires
are experiencing rapid population growth, which in some locations includes a recent
significant increase in part-time home ownership. One report analyzing the impacts and
demographics of the Rocky Mountain Region of Colorado (five counties in central
Colorado) found that approximately 60 percent of homes in this region are second homes
(see Lichtenstein 2004). Incorporating this dynamic into the PMT-TTM process could
offer risk communication insights for land managers, community organizers, and
homeowner association associations in working with this segment of homeowners.
The timing of a catastrophic natural hazard can affect perceptions of risk
vulnerability and severity. For example, hazard research has found that perceptions of
risk vulnerability and severity are higher after an extreme event and then wane over time
(Sims and Baumann 1983, Weinstein 1989). Two of our three communities were
sampled approximately six months after a large-scale wildfire came into their
community. Also, recurring fires were experienced in these areas. Future research should
explore the temporal and spatial dimensions of wildfire experience, subjective knowledge
and decision stage on the PMT model, which would help to address the question of what
happens “when the rains come” and the current drought is not as apparent.
Involuntary aspects and impediments for reducing risks is another theme that is
pervasive in our qualitative research (Bender et al. 2004; Integrated Resource Solutions
2004). Involuntary risks (i.e., as part of ‘dread risks’) have been demonstrated as a
dimension of risk that can potentially influence human cognitive processes for
responding to natural hazards (Slovic 1987). Neighbors’ or agencies’ actions (or
inactions) are an important factor in the decision-making process for homeowners’ in
taking precautionary measures (e.g., peer pressure, frustration with lack of action).
Additionally, other involuntary factors such as drought, weather patterns, the volatility of
mother nature (e.g., “no matter what I do, if the big one comes, I am toast”), living in a
dense forest location “surrounded by a tinderbox” are all additional reasons given by
20
homeowners for not undertaking risk reduction actions. These involuntary aspects of risk
and the related concept of responsibility need to be further explored and explicitly
examined to model why segments of homeowners are more or less likely to protect
themselves from the impact of natural hazard risks such as wildfires.
Finally, the effect of social and cultural variables on risk perception, risk
communication, and response requires further examination. This topic is discussed in the
natural hazards literature, as well as in the body of work on risk analysis (Flynn et al.
1994, McCaffrey 2004, Lindell et al. 1980, Plough and Krimsky 1990 [1987]). There is
empirical evidence that ethnic minorities vary systematically in ways that affect the
process and outcomes of risk communication, although the number of studies is small
(Lindell and Perry 2004). There also is very little work on the effect of culture, ethnicity,
and race on risk perception and response when it comes to wildfire. This topic is indeed
an important one that must be pursued in future research. As has been pointed out, future
studies in this area must go beyond simple reporting of differences that appear to be
cultural in origin. Model-based, cross-cultural studies seeking underlying causal variables
are needed (Weber and Hsee 1999). Education, income, age, and years of residence in the
U.S., for example, have been suggested as variables of interest (Winter and Cvetkovich
2003, Lindell and Perry 2004). The increased understanding provided by determining
which aspects of cultural variation may be causing possible differences in perceptions
and behaviors can provide managers with an enhanced capacity to understand and work
with distinct cultural, ethnic, and racial groups. For example, accommodation theory
would suggest that the more effort managers put into trying to communicate risk
effectively with an ethnically diverse community the more positive the response should
be from that community (Koslow, Shamdasani, & Touchstone 1994).
Management Implications
Management implications from this research center on understanding the
readiness stages of the affected publics and developing means and methods to move them
into the action stage. This involves communicating risk vulnerability and severity, as well
as providing recognizable means of effectively responding to the defined risks. People
21
must feel they have the knowledge, ability, and resources (self-efficacy) to deal with the
risk at hand and that the actions they take will effectively reduce the risk (response
efficacy), before they are ready to move into the action stage of risk reduction. Managers
responsible for communicating the risks of environmental hazards must be able to present
these messages to a variety of culturally, ethnically, and racially distinct groups. The
literature on natural hazards provides information on risk communication messages for
both majority and minority groups, much of which is applicable to wildfire. Essentially,
risk communication should center around homeowners’ vulnerability to the risk of
wildfire, the severity of the risk, and the tools (knowledge, ability, means, and resources)
that homeowners have to effectively deal with the risk. Various segments of the public in
different levels of risk readiness respond to the risk communication messages differently.
Information on how, when, and what to communicate to move the public to the
stage of risk reduction action is of considerable importance to public land managers
charged with implementing fire risk reduction programs. Some information seems to
apply to a wide variety of cultural/ethnic groups, while other information is apparently
more specific to certain segments of the population. Selected examples of how, when,
and what to communicate follow. The emphasis on risk communication methods and how
to communicate risk stems from the emphasis in the literature.
It is empirically known that there is a difference in the way American majority
and minority groups assess the credibility of hazard warning sources and respond to those
warnings. There are also differences within minority groups themselves (Lindell and
Perry 2004). As mentioned previously, precisely why there are differences requires
considerably more research into potential underlying causal variables.
For all groups, however, sources must be considered credible and trustworthy so
that people trust the source of the information and the content of the message (Lindell
and Perry 2004). Thus, it is important that managers know their affected groups and who
those groups consider to be credible sources (respected community leaders or opinion
leaders, for example). It is also important that managers increase their own credibility by
direct interaction and appropriate communication with their constituent groups (Lindell
and Perry 2004). Interactive communication allows for immediate feedback from the
public, which can indicate to managers when a message is not being understood or does
22
not meet the needs of the segment. Trustworthiness and expertise seem to be important
considerations in fostering a group’s willingness to respond (Lindell and Perry 2004).
Research in the hazard literature has also shown the importance of peers in
transmitting and influencing the adoption of hazard adjustments, or mitigation measures
(Lindell and Perry 2004). In this case, research by Lindell and Perry (2004) found that
Mexican Americans were likely to consider peers (friends, relatives, and neighbors) as
credible information sources concerning the risks of natural hazards. This group was
more likely to consider peers as credible sources than were Whites or African-Americans.
Additionally, McCaffrey (2002) has found that indirect wildfire experiences through
neighbors and friends have more of an impact on influencing greater adoption of
mitigating behaviors than did direct wildfire experience on homeowners. Why this may
be the case, and if this trend occurs in other studies, are important topics for future
research on underlying causal variables. Lindell and Perry (2004) also found that
authorities such as police and fire departments are considered credible sources by the
majority of Whites, African-Americans, and Mexican Americans in their study, except in
cases where the police have targeted minorities for crime enforcement in certain areas.
Land managers should communicate with these other agencies, neighborhood groups, and
community organizers about the various risk messages that might be effective for
homeowners in different segments of risk readiness.
Various aspects of the mass media are considered major channels for
communicating hazard and risk information to the public, which can be useful for
managers in certain instances. As a first statement, using sources in the appropriate
language is critical when dealing with many ethnically diverse groups, especially those
who are newly arrived immigrants. Language-specific television and radio stations, as
well as newspapers, are helpful in this regard (Lindell and Perry 2004). In their study,
Lindell and Perry (2004) found that Whites and African-Americans tended to be more
skeptical of the media than Mexican Americans. Again, why this may be the case
requires further exploration. It is unknown, for instance, if this may be related to use of
Spanish-language media sources by Mexican American groups.
In addition, it has been found that different ethnic groups seem to prefer different
media sources or communication channels. Some research has shown that Whites prefer
23
print media such as brochures and newspapers, while Mexican Americans prefer oral
media such as local radio and television stations and neighborhood meetings. AfricanAmericans, on the other hand, prefer local radio, newspapers, and brochures to speakers
at organizations or neighborhood meetings (Lindell and Perry 2004, Nelson and Perry
1991, Perry and Nelson 1991). For these preferences to have wide application for
managers, their rationale, extent of occurrence, and relationship to communications
concerning wildfire risk require in-depth examination. To clarify this point, recent
research on fire risk communication in the Southwest among Whites, as well as other
groups, found that direct, face-to-face communication in the form of presentations at
locally important meetings and door-to-door discussions with knowledgeable authorities
(such as fire department personnel) were considered to be very important (Integrated
Resource Solutions 2004).
When to communicate risk information to the public can be as important as the
content of the information and how it is communicated (Lindell and Perry 2004). A few
suggestions from the literature on natural hazards are equally appropriate for
communicating wildfire risk. Fire events, as with other natural disasters, can create
opportunities to raise awareness and stress risk vulnerability and severity. These events
are “windows of opportunity” or “teachable moments.” These brief windows opened by a
focusing event put vulnerability to environmental hazards, such as wildfire, on the
political agenda and on the minds of those affected by the event or in proximity to it.
Research shows, however, that the timing of the catastrophe and the dissemination of the
information is critical as interest wanes as time passes (Sims and Baumann 1983;
Weinstein 1989; Lindell and Perry 2004). Indirect experience with fire can be helpful in
increasing awareness, which can increase feelings of vulnerability and comprehension of
risk severity. These indirect experiences can include footage of fires and their results
viewed in the media, testimonials of victims, and information from affected friends and
relatives. Indirect experience can also create teachable moments for recent immigrants to
the U.S. when learning of disasters that have occurred in their homeland or country of
origin (Stallings 1986).
What to communicate addresses the crucial topic of moving members of the
public to the action stage of addressing fire risk. Presenting information in the appropriate
24
manner and at the appropriate time that addresses questions of self-efficacy and response
efficacy is required to accomplish this task. Managers should not oversimplify the
problem. This caution is especially important when attempting to encourage those in the
action stage, as this segment understands their vulnerability and needs to know the
specifics of implementing mitigating measures, the associated costs and resulting
benefits. Managers should be specific about how each mitigation activity will decrease
risk and how its benefits will outweigh the costs (Tierney 1993; Kates 1999).
Research on fire risk in the Southwest has shown that physically demonstrating what
should be done, indicating that help is available, and providing that help in a timely and
reliable manner considerably increases feelings of self-efficacy and response efficacy and
can move people to take action to mitigate fire risks on their property. Such actions have
included preparing demonstration houses/properties that show what fire risk mitigation,
or defensible space, should look like. Other activities include providing free pick-up and
hauling for thinned materials, free or low cost chipping for thinned materials, and burn
pits for those materials, among others (Integrated Resource Solutions 2004; Steelman and
Kunkel 2004). The town of Ruidoso, New Mexico has been cited as an excellent example
of a community that is proactively addressing its wildfire threat (Integrated Research
Solutions 2004; Steelman and Kunkel 2004).2 Additionally, policy changes to provide
funding incentives, grants, and state programs that provide financial assistance for
landowners to clear their property have encouraged people to undertake fire risk
mitigation activities that they might have viewed as too costly without outside assistance
(Steelman and Kunkel 2004). These are examples of the ways in which managers can
actually encourage desired actions by working with groups to accomplish those actions.
Perhaps the best advice for all types of land managers is to define the different
segments of stakeholders in a public land manager’s district or region. By developing an
in-depth description and understanding of these potentially diverse segments within the
communities surrounding public lands, managers are better equipped to develop a
strategy to effectively communicate with these stakeholders. Many times this
information is available through housing associations, thus, land managers need to
2
Information on the actions of this community can be found at the following websites:
www.ncsu.edu/project/wildfire/ruidoso and http://www.voruidoso.com/Forestry.html .
25
determine the gatekeepers in these communities. Based on developing an understanding
of a segment’s decision stage, level of knowledge, and other relevant information, the
manager can target risk communication information in the most effective way.
Traditionally, the health communication literature has used the PMT-TTM
framework through a pre-posttest experimental design framework to examine the effects
of various types of information on risk reduction behaviors. This study has demonstrated
that homeowners at different levels of readiness respond to and are motivated by a series
of cognitive processes. To extend the findings in this study, the next step would be to
introduce risk information that addresses these cognitive processes. We would predict
that the risk information that focuses on increasing perceptions of vulnerability would be
effective at moving precontemplatives into the contemplative stage. Next, information
that focuses on vulnerability and risk severity would move contemplatives to the action
stage. Finally, providing information on the means by which homeowners can reduce
their risk can increase the number of risk reduction actions that they choose to undertake.
References
Alba, J. W. and J. W. Hutchinson (1987). Dimensions of Consumer Expertise, Journal of
Consumer Research, 13(March), 411-54.
Alreck, P.L. and R. B. Settle (1995). The Survey Research Handbook: Guidelines and
Strategies for Conducting a Survey, second edition.
Bender, H.W., B. Kent, I.M. Martin, W.E. Martin, C. Raish (2003). Wildfire Risk
Perceptions, Behaviors, and Treatment Preferences for Homeowners in the WildlandUrban Interface, working paper.
Block, L.G. and P.A. Keller (1998). Beyond Protection Motivation: An Integrative
Theory of Health Appeals, Journal of Applied Social Psychology, 28(17), 1584-1608.
Burton, I. and R.W. Kates (1964). The Perception of Natural Hazards in Resource
Management, Natural Resources Journal, 3, 412-441.
Cohen, J. 2004. Understanding Wildland-Urban Fire Disasters. Rocky Mountain
Research Station, Fire Science Laboratory, Missoula, MT, unpublished document.
Faupel, C.E., S.P. Kelly, and T. Petee (1992). The Impacts of Disaster Education on
Household Preparedness for Hurricane Hugo, International Journal of Mass
Emergencies and Disasters, 10, 5-24.
26
Floyd, D., S. Prentice-Dunn, and R. Rogers (2000). A Meta-Analysis of Research in
Protection Motivation Theory, Journal of Applied Social Psychology, 30(2), 407-29.
Flynn, J., P. Slovic, and C.K. Mertz (1994). Gender, Race, and Perception of
Environmental Health Risks, Risk Analysis, 14(6), 1101-1108.
Kahneman, A. and A. Tversky (1979). Prospect Theory: An Analysis of Decision under
Risk, Econometrica, 47(2), 263-91.
Horwath, C.C. (1999). Applying the Transtheoretical model to eating behavior change:
challenges and opportunities, Nutrition Research Reviews, 12, 281-317.
Integrated Resource Solutions (2004). Draft Interim Report and Phase One of Public
Input Concerning Fire and Fuels Management in the Southwestern Region: Results
form the Lincoln National Forest, Unpublished Report.
Kates, R.W. (1999). Natural Hazard in Human Ecological Perspective: Hypotheses and
Models, In Environmental Risks and Hazards, ed. S.L. Cutter, 78-93. Englewood Cliffs,
NJ: Prentice Hall.
Koslow, S., P. Shamdasani, and E.E. Touchstone (1994). Exploring Language Effects in
Ethnic Advertising: A Sociolinguistic Perspective, Journal of Consumer Research,
20(4), 575-85.
Lazarus, R.S. and S. Folkman (1984). Stress, Appraisal and Coping. New York, NY:
Springer.
Lichtenstein, G. (2004). Part-Time Paradise, High Country News, 36 (20), 7-12.
Lindell, M.K and R.W. Perry (2004). Communicating Environmental Risk in Multi
Ethnic Communities, Sage Publications.
Lindell, M.K., R.W. Perry, and M.R. Greene (1980). Race and Disaster Warning
Response, NSF Grant ENV-77-23697, Battelle Human Affairs Research Centers.
Maddux, J.E. and R.W. Rogers (1983). Protection motivation and self-efficacy: A revised
theory of fear appeals and attitude change. Journal of Experimental Social Psychology,
19, 469-79.
McCaffrey, S. (2004). Thinking of Wildfire as a Natural Hazard, Society and Natural
Resources, 17, 509-15.
McCaffrey, S. (2002). For Want of Defensible Space a Forest is Lost: Homeowners and
the Wildfire Hazard and Mitigation in the Residential Wildland Intermix at Incline
Village, Nevada, Doctoral Dissertation, University of California, Berkeley.
27
Mileti, D.S. and J.H. Sorenson (1987). Natural Hazards and Precautionary Behavior in
Taking Care: Understanding and encouraging self-protective behavior, N. Weinstein
(ed), Cambridge University Press, 189-207.
Mileti, D.S., T.E. Drabek, and J.E. Haas (1975). Human Systems in Extreme
Environments: A Sociological Perspective. Boulder: Institute of Behavioral Science,
University of Colorado.
Mitchell, A.A. and P.A. Dacin (1996). The Assessment of Alternative Measures of
Consumer Expertise, Journal of Consumer Research, 23(December), 219-39.
Mulilis, JP and T.S. Duval (1995). Negative Threat Appeals and Earthquake
Preparedness: A Person-Relative-to-Event (PrE) Model of Coping with Threat, Journal
of Applied Social Psychology, 25(15), 1319-39.
Mulilis, JP and R. Lippa (1990). Behavioral Change in Earthquake Preparedness Due to
Negative Threat Appeals: A Test of Protection Motivation Theory, Journal of Applied
Social Psychology, 20(8), 619-38.
Nelson and, L. and R.W. Perry (1991). Organizing Public Education for Technological
Emergencies, Disaster Management, 4, 21-26.
Pechmann, C., G. Zhao, M.E. Goldberg, & E.Thomas Reibling (2003). What to Convey
in Antismoking Advertisements for Adolescents: The Use of Protection Motivation
Theory to Identify Effective Message Themes, Journal of Marketing, 67(April), 1-18.
Perry, R.W.and L. Nelson (1991). Ethnicity and Hazard Information Dissemination,
Environmental Management, 15, -581-587.
Plough, A. and S. Krimsky (1990). The Emergence of Risk Communication Studies:
Social and Political Context, in Readings in Risk, T.S. Glickman and M. Gough (eds.).
Washington, DC: Resources for the Future, 223-31.
Prochaska, J.O., J.C. Norcross, and C.C. DiClemente (1994). Changing for Good, New
York, NY: William Morrow.
Rippetoe, P. and R. Rogers (1987). Effects of Components of Protection Motivation
Theory on Adaptive and Maladaptive Coping with a Health Threat, Journal of
Personality and Social Psychology, 52(3), 596-604.
Rogers, R. (1983). Cognitive and Physiological Processes in Fear Appeals and Attitude
Change: A Revised Theory of Protection Motivation, Social Psychophysiology, 153-76.
28
Rossi, S.R., G. Reed, G.W. Greene, J.S. Rossi, J.O. Prochaska, and W.F. Velicer (1994).
Development of a Self-efficacy Questionnaire for Dietary Fat Reduction, paper presented
at the 14th annual convention of the Society for Behavioral Medicine, Boston, MA.
Rustemli, A. and N.A. Karanci (1999). Correlates of Earthquake Cognitions and
Preparedness Behavior in Victimized Populations, The Journal of Social Psychology,
11(4), 767-780.
Sims, J.H. and D.D. Baumann (1983). Educational Programs and Human Response to
Natural Hazards, Environment and Behavior, 15, 165-189.
Slovic, P. (1987). Perception of Risk, Science, 236, 28085.
Slovic, P. (2002). Cigarette Smokers: Rational Actors or Rational Fools? In Smoking:
Risk, Perception, and Policy, P. Slovic (ed.), Sage Publications, Thousand Oaks, CA.,
97-124.
Sporny, L.A. and I.R. Contento (1995). Stages of Change in Dietary Fat Reduction:
Social Psychological Correlates, Journal of Nutrition Education, 27, 191-99.
Steelman, T.A. and G.F. Kunkel. (2004). Effective Community Responses to Wildfire
Threats: Lessons from New Mexico, Society & Natural Resources, 17(8), 679-699.
Stallings, R.A. (1986). Reaching the ethnic minorities: earthquake public education in the
aftermath of foreign disasters, Earthquake Spectra, 2(4), 695-702.
Tierney, K.J. (1993). Socio-Economic Aspects of Hazard Mitigation. University of
Delaware, Disaster Research Center.
Tversky, A. and D. Kahneman (1973). Availability: A Heuristic for Judging Frequency
and Probability, Cognitive Psychology, 4, 207-32.
Viscusi, W.K. (1990). Do Smokers Underestimate Risks? Journal of Political Economy,
98, 1253-69.
Weber, E.U. and C.K. Hsee (1999). Models and Mosaics: Investigating Cross-cultural
Differences in Risk Perception and Risk Preference. Psychonomic Bulletin & Review,
6(4), 611-617.
Weinstein, N.D. (1989). Effects of Personal Experience on Self-Protective Behavior,
Psychological Bulletin, 105(3), 31-50.
Weinstein, N.D. (1987). Unrealistic optimism about Susceptibility to Health Problems:
Conclusions from a Community-wide Sample, Journal of Behavioral Medicine, 10, 481500.
29
Weinstein, N.D.and P.M. Sandman (1992). A model of the precaution adoption process:
evidence from home radon testing, Health Psychology, 11, 170-80.
Weinstein, N.D., Rothman, A.J., and Sutton, S.T. (1998). Stage Theories of Health
Behavior: Conceptual and Methodological Issues, Health Psychology, 17, 290-99.
Winter, P.L. and G.T. Cvetkovich (2003). A Study of Southwesterners’ Opinions on the
Management of Wildland and Wilderness Fires: Fire Management Version. Riverside,
CA: US Department of Agriculture Forest Service, Pacific Southwest Research Station.
Appendix
Risk Reduction Behaviors











Creating a minimum 30-foot defensible space around your home
Planting low-growing, fire resistant plants around your home
Putting a fire resistant roof on your home
Putting fire resistant undersides to any decks and balconies on your home
Removing any dead branches from your home’s roof and around the chimney
Making sure that your home is easily identifiable and accessibly from a main road
Making sure that all the trees on or near your property are away from structures
Making sure that all the trees on or near your property are away from utility lines
Working with neighbors to clear common areas and prune areas of heavy vegetation
Stacking firewood and scrap wood piles at least 30 feet from any structure
Contacting your local fire department to get a personal fire safety inspection at your
home and property.
Source: Firesafe Council of California website: www.firesafecouncil.org
30