14
Risk Perception and Attitudes Towards Mitigation
Güler O. Fişek Ph.D.
Nur Yeniçeri Ph.D.
Serra Müderrisoğlu Ph.D.
Gökçe Özkarar M.A.
Department of Psychology
and
Center for Disaster Management
(CENDIM)
Boğaziçi University
IIASA-DPRI Meeting
Integrated Disaster Risk management: Megacity Vulnerability and Resilience
Laxenburg, Austria
July 29-31, 2002
The research reported here was funded by the Boğaziçi University Research Fund, research code number 00R107.
Risk Perception and Attitudes Towards Mitigation
Güler O. Fişek, Nur Yeniçeri, Serra Müderrisoğlu and Gökçe Özkarar
Department of Psychology
Boğaziçi University
Abstract
This paper is a partial report on the results of the field survey conducted as part of the psychosocial module of the project on an integrated decision support system for disaster management in Turkey. The sample consisted of 254 adults from five different regions of İstanbul, and of varying levels of socioeconomic status. The results indicate that despite low levels of damage in the 1999 earthquake, the respondents were traumatised and not ready to face another major disaster. Level of education and income were significant factors in reported levels of trust in official information, felt readiness and mitigation behaviors. The results indicated a complex relationship between these demographic variables, stated attitudes towards mitigation and actual mitigation behaviors, such that lower SES respondents who expressed more positive attitudes towards mitigation, reported fewer actual mitigation behaviors. While financial constraints may be important, it appears that response biases should be taken into account in evaluating avowed attitudes regarding willingness to mitigate.
Note: The authors would like to thank the Frekans Research Company for conducting the survey for this investigation.
Introduction
The aim of the Psychosocial Module of the Project on an Integrated Decision Support System for Disaster Management in Turkey, was to prepare a guideline for local disaster preparedness, response and recovery plans, involving only the human element of disaster management. To meet this objective, the research was planned on two related issues:
a) Phase I: to assess the population's state of mind regarding risk perceptions, level of knowledge, willingness to engage in mitigation activities and responsiveness to public information.
b) Phase II: to delineate a psychoeducational plan for preparedness at the psychosocial level in view of the data.
It is essential to have information about the form and level of disaster awareness, preparedness and vulnerability of the local population in order to prepare guidelines for the activation and implementation of disaster awareness and preparedness education programs by local authorities. Thus the aim of this study was to collect information regarding the social, and relational structures of local community residents to see how they can be organised and prepared to meet disasters, to study their level of psychological awareness especially as a result of experience with the earthquakes of 1999, to map out the cultural codes through which risk awareness can and is communicated, and to measure local communities' psychological vulnerability and readiness for mitigation regarding disasters.
Phase I. Setting up and Implementing an Earthquake Preparedness Inventory
The preparation of the inventory involved the following essential steps:
(1) carrying out a pilot study to delineate/formulate the assumptions regarding households in the targeted areas, and to evaluate the appropriateness of the questions asked. This step also involved:
a. training the interviewers which would involve imparting to them the skills necessary for successful administration of the inventory to the participants, and
b. deciding on the rules to be followed in sampling the households to be visited during data collection phase of the study.
(2) Evaluating the pilot study results and preparing the final inventory according to the pilot findings ,
a. training a new set of interviewers,
b. setting up a new and final sample of communities to be canvassed.
(3) Analysing and writing up the final report
Method
The Preparation of the Inventory.
The questions that were included in the inventory were selected from a variety of sources, including such surveys conducted in the United States that were accessible (Blanchard-Boehm, 1997), a report prepared by Boğaziçi University colleagues after the August 17 earthquake (Alp, Kabasakal, and Sirman, 1999), reports from the Turkish literature (Sakallı, 2000), and questions posed by colleagues with an interest in the findings. The inventory went through several versions until a final pilot version was considered satisfactory. The completed inventory was composed of seven subsections as follows:
I. Demographic Information: This includes vital statistics regarding the respondent and the members of the household as well as financial information regarding the residence.
II. Previous Earthquake Experience: This includes information on the respondent's and the household's experiences in the 1999 Marmara earthquake, the major aftershock and the Düzce earthquake.
III. Reports of Post Traumatic Stress Complaints: This section refers to the respondent's report of stress symptoms after these earthquake experiences. The items were taken from The Impact of Events Scale (Horowitz, Wilner and Alvarez, 1979).
IV. Information about Earthquakes: This section evaluates the level of knowledge of the respondent about the causes and effects of earthquakes,
V. Risk Perception and Expectations about Future Earthquakes: This section includes questions about the respondent's assessment about the level of risk to his/her residence, locality and city, predictions about what kind of damage might result from a future earthquake etc.,
VI. Subjective Appraisals and Attitudes towards Mitigation: This section covers the respondent's attitudes about the value of mitigation behaviours and subjective evaluation of his/her ability to reduce the potential risks posed by a future earthquake,
VII. Actual or Planned Mitigation Behaviours: This sections covers what mitigation activities were engaged in, if any by the individual, family or residents of an apartment, the reasons for engaging or not engaging in such activities.
The Pilot Study
Since the main purpose of the pilot study was to test the appropriateness of the inventory and the respondents' willingness and ability to answer the questions posed to them, it was decided to test a small sample. Nevertheless for purposes of comparison two localities in Istanbul were chosen for their specific characteristics. One was the locality of Avcılar, which suffered some damage in the 1999 Marmara earthquake and is known to be located on insecure ground; the other was Rumelihisarüstü, a locality situated on relatively rocky ground, and which did not suffer any significant effects from that earthquake, despite the fact that construction tends to be poor there. A random sample of 25 respondents was interviewed in each locality, which were demographically similar. Results indicated that both localities showed similar levels of traumatisation from the 1999 earthquakes; had similar low levels of information and trust in information; and similar attitudes towards risk and mitigation. The main difference was in the fact that the more traumatised respondents showed more clear evidence of processes involving defensiveness, helplessness and inconsistencies in their responses (Fişek, et al, 2001).
The final sample consisted of 254 respondents from five different localities in Istanbul. Avcılar (N=54), Fatih (N=50), and Erenköy (N=50) represented relatively poor soil conditions and Beşiktaş (N=50) and Beykoz (N=50) represented relatively better soil conditions. The socioeconomic status of the respondents was similarly varied, with Erenköy and Beşiktaş representing a higher SES level in general. The demographic characteristics of the sample is presented below.
Procedure
The interviewers approached each respondent with an explanation of the purpose of the inventory, the connection to the university and asked if the respondent were willing to participate. There were no refusals and the interview took about one hour.
Results
The results will be presented in two parts: descriptive statistics on the demographic characteristics of the respondents and some of their responses, and the findings from some analyses of the relations between risk perception, demographics and mitigation variables.
Demographic Characteristics of the Respondents.
The respondents in the five localities are rather similar demographically as Table 1 indicates. Only about a quarter of the respondents were male, as the interviews were conducted at home in the daytime. The mean age of the respondents was 39. The localities differed in terms of education level, Beşiktaş and Erenköy scoring significantly higher than Avcılar, Fatih and Beykoz (Table 2). Table 3a shows that 85% of Avcılar and Fatih fall below the poverty level for Istanbul, while this rate is about 55% for Beşiktaş and Beykoz and 32% for Erenköy. About three fourths of the Beykoz sample own their own house, while this rate is about two-thirds for the rest (Table 3b). This is most likely due to the fact that the homes in Beykoz tend to be gecekondu (squatter) homes built by the people illegally on government land.
Risk Perception.
All respondents express an expectation of relatively high risk to their dwellings in a future earthquake. As seen in Table 4a, Fatih with relatively poor soil expects significantly higher risk than Beykoz, with the best soil conditions. The primary criterion they use to evaluate risk is their level of trust in the construction of their buildings, followed by the level of trust in the zone, and trust in God (Table 4b).
The criteria were ranked in terms of their abstractness versus realism, with a higher score indicating less concrete/more abstract criteria, more remote from the realistic. As seen in Table 4c trust in the building was the most concrete criterion and don’t know answers the least. Erenkoy, with higher levels of education and income had the most concrete criteria in evaluating risk.
In terms of expectations of damage to the zone they lived in, as would be expected Avcilar and Fatih see more risk then Beşiktaş and Beykoz (Table 5a). The criteria for damage expectation were ranked in the same manner and can be seen in Table 5b. Table 6 indicates that all respondents expect a higher destructive potential in the next quake, compared to the 1999 earthquakes. Finally Table 7 indicates that overall risk perception is highest in Avcılar and Fatih.
In summary we see that the perceptions of risk and expectations of damage are expectable and realistic, considering the different zones. The criteria for risk evaluation are mostly realistic, involving primarily structural concerns. Finally overall risk perception is high.
Preparedness and Mitigation.
Table 8a indicates that more than one third of the respondents claim they have a plan about what to do at the time of an earthquake. Erenköy leads those who have practiced or drilled on their plan by about two thirds, diminishing to one fourth of the respondents in Beşiktaş. In terms of preventive measures, again Erenköy heads the group both inside the home and for their buildings, with fatih reporting the lowest level of mitigation with 10% (Table9a). As for reasons for undertaking mitigation, safety needs and concern about the family dominate (Table9b). It is interesting to see in this table that Beşiktaş and Erenköy, the most educated groups, did not endorse scientists’ warmings as a reason to mitigate but the other groups did.
Table 9c presents reasons for not undertaking any mitigation measures. There are many reasons, topped by trust in their buildings and the expenses involved. Table 9d presents the list of preventive measures undertaken. Most claim they learned what to do during an earthquake. Taking out insurance ranges from 72% in Erenköy to24% in Beykoz. This is a repeated trend: Erenköy and Beşiktaş indicate higher percentages in the various preventive measures, presumably due to higher education and income. This presumption is strengthened when we look at Table 9e, presenting the overall level of completed prevention, with Erenköy showing higher percentages than the rest.
Asked about reasons for not taking out insurance, the most common response involved the expense, followed by lack of trust in the government (Table 9f). Finally Table 10 shows that the overall level of felt readiness follows an expectable pattern. Erenköy feels the most ready and Avcılar and Fatih the least.
In summary these results seem to indicate that mitigation is more a matter of income and education than anything else. While Avcılar and Fatih realistically perceive higher risk for their localities, they do not seem to have followed up that perception with appropriate efforts at mitigation.
Relationships between the Variables.
This section presents some of the analyses done to explore relationships between the variables. Gender was explored and women were found to be more traumatised, use more abstract criteria for risk perception and feel less ready (Table 11). Table 12 presents the effects of income on some variables, and the presumptions mentioned above seem to be borne out. Income has no effect on overall risk perception. However those with higher income use more concrete criteria for risk assessment, report more completed prevention, a more positive attitude towards insurance, more total trust in various agencies and more felt readiness. Thus it appears that while income does not affect the assessment of risk it does affect how one feels about it and what one does.
A number of regression equations were run to see the predictors of some variables. The only variable significantly predicting the level of overall risk perception was the post traumatic stress reactions score (Table 13). Education, income, gender, criterion for risk assessment and previous experience of damage were excluded from the equation. Trust was significantly predicted only by income (Table 14), while education, income, gender and experienced damage were excluded. Lastly the predictors of completed prevention and level of felt readiness were evaluated. In addition to income and abstractness of risk assessment criteria, some attitudinal factors derived from a series of statements involving the earthquake experience, were included as predictors. “Subjective norms” refers to appropriate preventive behaviour according to significant others, “attitude towards preparation” refers to one’s belief in the value of preparedness, “knowledge” refers to avowals of knowing what to do in earthquakes, and “perceived behavioural control” refers to avowals of self efficacy regarding earthquake preparedness. In all cases a higher score is a more positive response.
Tables 15 and 16 present the results. Income predicts both completed prevention and felt readiness positively. The interesting finding seen in these tables is that a more positive score on the attitude factors predicts less completed prevention and less felt readiness. Thus it appears that avowals of all sorts of positive attitudes about earthquake preparedness does not translate into mitigation behaviour and the resulting sense of being prepared for the next big one.