%20-%20https://sketch.com%20--%3e%3ctitle%3e3ECBBF50-FD08-412A-943D-2CFC5E025A16%3c/title%3e%3cdesc%3eCreated%20with%20sketchtool.%3c/desc%3e%3cg%20id='Stylesheet'%20stroke='none'%20stroke-width='1'%20fill='none'%20fill-rule='evenodd'%3e%3cg%20transform='translate(-926.000000,%20-668.000000)'%20id='BUTTONS'%3e%3cg%20transform='translate(100.000000,%20597.000000)'%3e%3cg%20id='BTN-/-RIGHTS'%20transform='translate(816.000000,%2065.000000)'%3e%3cg%20id='ICON-/-Rights'%20transform='translate(10.000000,%206.000000)'%3e%3cpath%20d='M13.5752841,0%20L1.34943182,0%20C0.59375,0%200,0.586666667%200,1.33333333%20L0,22.6666667%20C0,23.4133333%200.620738636,24%201.34943182,24%20L17.6505682,24%20C18.40625,24%2019,23.3866667%2019,22.6666667%20L19,5.36%20L13.5752841,0%20Z%20M13.5752841,1.89333333%20L17.0838068,5.36%20L13.5752841,5.36%20L13.5752841,1.89333333%20Z%20M1.34943182,22.6666667%20L1.34943182,1.33333333%20L12.1988636,1.33333333%20L12.1988636,5.36%20C12.1988636,6.10666667%2012.7926136,6.69333333%2013.5482955,6.69333333%20L17.6235795,6.69333333%20L17.6505682,22.6666667%20L1.34943182,22.6666667%20Z'%20id='Shape'%20fill='%23595959'%20fill-rule='nonzero'%3e%3c/path%3e%3ccircle%20id='Oval'%20stroke='%23595959'%20stroke-width='1.5'%20fill='%23595959'%20cx='9.5'%20cy='13.5355339'%20r='5.5'%3e%3c/circle%3e%3cpolyline%20id='Path-2'%20stroke='%23FFFFFF'%20stroke-width='1.5'%20transform='translate(9.500000,%2013.035534)%20rotate(45.000000)%20translate(-9.500000,%20-13.035534)%20'%20points='8%2016.0355339%2011%2016.0355339%2011%2010.0355339'%3e%3c/polyline%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
Disasters, by disturbing the ecological balance worldwide, result in death, injury, and economic, psychological, and social losses. In the year 2020, a total of 389 disasters occurred globally, affecting 98.4 million people, claiming 15 080 lives, and causing an economic loss of 171.3 billion dollars. 1 In Turkey, during the year 2019, a total of 24 400 natural hazards occurred, consisting of 23 646 earthquakes, 245 landslides/rockfalls, 499 floods, and 10 avalanches. 2 According to the Global Risk Index report, which is designed to measure the risks of disasters and humanitarian crises, Turkey ranks 45th among 191 countries. With an index score of 5.0, Turkey is classified in the “high-risk” category. 2
The significance of disaster risk management was emphasized in the Hyogo Framework for Action in 2005, with one of the disaster reduction objectives being to “enhance awareness of the importance of disaster risk reduction policies, thus facilitating and promoting their implementation.” 3 In 2015, the Sendai Framework for Disaster Risk Reduction identified “strengthening disaster risk governance to manage disaster risks” as one of its four priority action areas. 4 In addition, ensuring public participation is crucial for the effective implementation of disaster risk management practices.Reference Titko and Ristvej 5 , Reference Onuma, Shin and Managi 6
Most countries attach significant importance to disaster prevention and risk reduction policies.Reference Zou and Yuan 7 –Reference Selby and Kagawa 10 People’s responses to risks hold central importance in disaster risk reduction planning and policymaking.Reference Sjöberg 11 , Reference Samamdipour, Seyedin and Ravaghi 12 Understanding how individuals perceive risks has become increasingly important in disaster management since it forms the cornerstone of risk management activities. As a result, studies on risk perception in disaster management have gained significance.Reference Birkholz, Muro and Jeffrey 13 , Reference Prabhakar, Srinivasan and Shaw 14 These studies provide decision-makers with a better understanding of public engagement in disaster risk management practices, anticipate the perceptions and behaviors of people at risk, and facilitate effective risk communication.Reference Jóhannesdóttir and Gísladóttir 15
Understanding the perception of individuals at risk is essential to determine disaster risk.Reference Alcántara-Ayala and Moreno 16 Identifying disaster risk perception is a prerequisite for effective disaster risk management and enhancing the safety of citizens.Reference Birkholz, Muro and Jeffrey 13 , Reference Cvetković 17 –Reference Qureshi, Khan and Rana 21 In general disaster planning strategies, information on disaster experiences and risk perceptions is crucial for disaster preparedness, risk reduction, and prevention. Local communities can best identify their urgent needs and coordinate preparations. Therefore, the perception and practices of the local population should not be overlooked and should be considered.Reference Khan, Johar and Ndalai Baba 22
Numerous studies and practical experiences related to risk have demonstrated the importance of risk perceptions for human behavior.Reference Siegrist and Árvai 23 Various models have been developed to measure the relationship between risk perceptions, behaviors, and qualitative characteristics.Reference Slovic 24 In the existing literature, risk perception has been addressed in relation to various types of disasters and disaster-associated topics. However, it is evident that different dimensions are employed in measuring risk perception in disasters across studies. For instance, some studies assess earthquake risk perception using two dimensions (possibility and fear),Reference Qing, Guo and Deng 25 while others employ three dimensions (fear, likelihood of earthquakes, and expected harm from earthquakes).Reference Ahn, Takikawa and Maly 26 A study on landslide risk perception utilized a 14-item scale covering dimensions such as probability, fear, worry, and fatalism.Reference Xu, Peng and Su 27 In contrast, another study on landslide risk perception examined dimensions such as knowledge, local influence, and institutional influence.Reference Ardaya, Evers and Ribbe 28 Similarly, studies conducted in flood-prone areas have measured risk perception using different dimensions.Reference Rana and Routray 29 , Reference Ullah, Saqib and Ahmad 30 This inconsistency in risk perception measurements is not only observed across disaster types but also among different sample groups, such as farmers and school children.Reference Saqib, Kuwornu and Panezia 31 , Reference Shah, Gong and Ali 32 As the current body of research indicates, there is no comprehensive, valid, and reliable tool designed to measure disaster risk perception across multiple disaster types such as earthquakes, floods, and landslides. Therefore, the development of a disaster risk perception scale will significantly contribute to the literature.
One of the primary objectives of this study is to contribute to the literature by developing a disaster risk perception scale, given the limited number of studies focusing on disaster risk perception. Another aim of the study is to examine the relationship between disaster risk perception and preparedness, and to measure the impact of disaster risk perception on disaster preparedness.
Concepts and Theoretical Background
Concept of Risk
The concept of risk is a central focus in the management of various activities and technologies.Reference Fischhoff and Watson 33 –Reference Lindell and Malmfors 35 Definitions of risk lie at the core of how contemporary society experiences, discusses, and responds to risk.Reference Cordner 36 The concept of risk has been employed in various fields in the literature.Reference Kaplan and Garrick 37 If the future were independent of predetermined or existing human activities, the concept of risk would be meaningless.Reference Rohrmann, Renn, Mumpower and Renn 38 Assuming a distinction between reality and probability, the concept of risk is often associated with the probability of an undesirable reality situation (negative impacts) occurring as a consequence of natural events or human activities.Reference Hilgartner, Short and Clarke 39
The field of disaster risk science, like many other areas, is filled with definitions developed by practitioners, policymakers, and academics.Reference Kelman 40 In disaster research, the concept of risk is generally defined as a disruptive or adverse external event that threatens the well-being of individuals, communities, or societies.Reference Bakić 41 Emotionally, risk expresses our rapid, instinctive, and intuitive responses to danger, while analytically, risk introduces logic, reasoning, and scientific assessment into hazard management.Reference Slovic 42 , Reference Slovic and Peters 43 Risk is characterized by a situation or action in which two or more outcomes are possible, the specific outcome is unknown, and at least one of the probabilities is undesirable.Reference Covello and Merkhoher 44 Risk arises from the interactions between harmful outcomes’ probabilities or expected losses (deaths, injuries, livelihoods, economic losses, etc.), natural or human-induced hazards, and vulnerable conditions. 45
The UNDRR emphasizes that disaster risk is determined probabilistically based on the likelihood of injuries, loss of life, and damage to assets within a community, considering factors such as hazards, vulnerabilities, and capacities. Moreover, disaster risk is categorized into extensive disaster risk and intensive disaster risk depending on severity and frequency. A disaster risk with low severity and high frequency is defined as extensive disaster risk, whereas high severity with medium or low frequency is classified as intensive disaster risk. 46
Risk Perception
Understanding the social structure of social risks requires a holistic, contextual, and community-based approach.Reference Fitchen, Heath, Fessenden-Raden, Covello, Mumpower, Spicker and Stallen 47 Some argue that risk is not merely a representation of real hazards but a social construct, and involving the public in defining tolerable risk levels has been debated.Reference Wynne, Krimsky and Golding 48
The existing literature demonstrates that perceived risk is measurable and predictable.Reference Gardner, Tiemann and Gould 49 –Reference Slovic, Fischhoff, Lichtenstein, Schwing and Albers 51 The use of probabilities, decision trees, and other alleged objective processes to assess risk suggests that risk can be measured in an unbiased manner.Reference Finucane 52 Risk is entirely related to thoughts, beliefs, and constructs.Reference Sjöberg 53 Information, experience, values, attitudes, and emotions all influence individuals’ thoughts and judgments about the seriousness and acceptability of risks.Reference Slovic 54 Trust, involvement, social impact, and self-efficacy influence risk perception, and the need for information is influenced by involvement and social impact.Reference Huurne and Gutteling 55
Since the late 1960s, there has been a significant accumulation of literature concerning environmental risk perception.Reference Slovic 56 –Reference White 58 Although early studies on risk perception date back to the beginning of the 1960s, the concept of “perceived risk” gained prominence in the mid-1970s.Reference Slovic, Fischhoff, Lichtenstein, Schwing and Albers 51 , Reference Slovic, Fischhoff and Lichtenstein 59 The ways individuals perceive their vulnerabilities to natural hazards shape their responses and coping strategies for dealing with these risks.Reference Schumm 60 Slovic (2000),Reference Slovic 24 emphasized that when people perceived a risk as real, their behaviors would be influenced accordingly.
Risk perception is the subjective evaluation of the likelihood of a specific event occurring and how concerned individuals are about its consequences.Reference Sjöberg, Moen and Rundmo 61 It is defined as individuals’ beliefs, attitudes, judgments, and emotions, as well as broader social or cultural values and inclinations that they adopt, regarding hazards and benefits.Reference Pidgeon, Hood and Jones 62 The process by which people’s beliefs and circumstances shape their thoughts about any risk is referred to as risk perception.Reference Smith 63 The process of assessing and understanding why individuals take precautions against a threat is related to the concept of risk perception.Reference Rana, Bhatti and Aslam 64 In its most comprehensive form, risk perception is understood as aiming to describe how people feel about something.Reference Otway and Thomas 65
Risk perception is not solely an emotional perception; it involves attitudes and expectations.Reference Sjöberg 66 Perceived risk is related to how an individual understands and experiences an event.Reference Sjöberg, Moen and Rundmo 61 People assess the risks they encounter to decide how to cope with them, which is entering the realm of risk perception.Reference Visschers, Meertens, Lavino and Neumann 67 Studies on risk perception examine individuals’ expressed opinions when asked to characterize and evaluate hazardous activities and technologies in various ways.Reference Slovic, Fischhoff and Lichtenstein 68 , Reference Rohrmann 69 Risk perception encompasses the process of gathering, selecting, and interpreting signals about the uncertain effects of events, activities, or technologies. Perceptions may vary based on the type of risk, risk context, and social context.Reference Wachinger and Renn 70 It can be argued that when there is nothing that can be called or felt as risk, it is impossible to perceive risk, and there is no risk perception in such situations.Reference Brehmer, Singleton and Hovden 71
Risk perception traditionally relies on psychological (cognitive) or socio-psychological (social-cognitive) theories to explain how individuals characterize and evaluate the risks they encounter and make decisions about taking protective measures.Reference Shreve, Begg and Fordham 72 One of the oldest and most influential models in risk perception is the psychometric paradigm, which utilizes psychophysical scaling and multivariate techniques to explain individuals’ responses to hazards.Reference Slovic 54 , Reference Bodemer and Gaissmaier 73 The psychometric paradigm is an example of an approach that uses explanatory variables semantically close to the risk dimensions it seeks to explain. It uses the directions and attributes of hazards to account for the perceived level of risk and the acceptability of risks.Reference Sjöberg 74 The development of the psychometric paradigm contributed to exploring strategies for examining perceived risks.Reference Fischhoff, Slovic and Lichtenstein 75 This paradigm investigates how individuals make quantitative decisions about the perceived risks in a given situation.Reference Slovic 24 , Reference Jasanoff 76
Disaster Risk Perception
Natural hazards are typically perceived as involuntary risks. However, preparing households for an emergency is a voluntary response to an inevitable risk.Reference Palm 77 The risk perception regarding natural hazards is primarily a function of both the characteristics of the hazards and the stakeholders involved.Reference Grossi, Kunreuther, Windeler, Grossi and Kunreuther 78 The role of risk perception is also important when examining measures of vulnerability.Reference Armaş 79
Regarding disaster risk perception, various approaches exist, ranging from rationalist perspectives to evaluating benefits against costs, to defining risk perception as a dynamic practice imposed and shaped by societies.Reference Ardaya, Evers and Ribbe 28 These approaches highlight the need to consider multiple factors when studying disaster risk perception. When identifying factors that affect disaster risk perception, researchers often tend to categorize them based on a multi-level classification.Reference Pazzi, Morelli and Bonati 80 , Reference Beck and Featherstone 81
Much of the literature on disaster risk perception has suggested that perception can be shaped by factors such as experience of hazards or exposure to information related to objective risk.Reference Knuth, Kehl and Hulse 82 Some studies have also indicated that the description and management of disaster risk perception are linked to local cultures and social structures.Reference Dressel 83 , Reference He and Zhai 84 Disaster risk perception directly influences people’s abilities and preparedness levels.Reference Pazzi, Morelli and Bonati 80 , Reference Rana, Jamshed and Younas 85 Therefore, it is crucial for the public to maintain a sufficient level of risk perception effectiveness in order to be prepared for disasters.Reference Kim and Madison 86
Based on the current literature, this study formulated four hypotheses:
Hypothesis 1 (H1): The scale developed to determine citizens’ disaster risk perception will be valid and reliable.
Hypothesis 2 (H2): Participants’ disaster risk perceptions will show significant differences based on their socio-demographic characteristics (gender, age, education level, disaster experience, and disaster volunteering).
Hypothesis 3 (H3): There will be a positive and significant relationship between disaster risk perception and disaster preparedness.
Hypothesis 4 (H4): The five dimensions of disaster risk perception (exposure/impact, probability, uncontrollable, worry/fear, and vulnerability) will have positive and significant effects on disaster preparedness.
Methods
Research Design
In this study, quantitative research methods were used since the aim was to determine whether the scales differentiated based on the variables used in the research and Likert-type scales were employed. The research design involved both causal-comparative and correlational survey designs. Causal-comparative research involves comparing two or more groups to explain existing differences.Reference Johnson 87 Relational research is a non-experimental study that investigates whether there is a relationship between variables within a single group.Reference Seeram 88 In this study, both relational and causal-comparative designs were employed. Initially, the relationships between the subdimensions of the developed disaster risk perception scale were examined, followed by an analysis of their impact on the independent variable of disaster preparedness. Additionally, the study is descriptive in nature, aiming to identify the conditions related to disaster risk perception. The data for the research were collected through face-to-face and online methods between November 19, 2021-December 30, 2021. Most of the data were gathered face-to-face, with only 19 participants completing the survey online during the pilot study and 36 participants during the main study. The collected data were analyzed using descriptive statistics, mean comparison tests, factor analysis, and multiple regression analysis.
Participants
The population of the study consists of citizens aged 18 and over residing in Giresun and Elazig provinces who experienced recent earthquakes, floods, and landslides (Figure 1). According to the Turkish Statistical Institute data from 2021, the total population of Elazig province is 591 098, and Giresun province has a total population of 453 912. 89 Thus, the population of the study is composed of 1 045 010 citizens. Due to the large size of the target population, the researchers utilized the non-probability sampling method of simple random sampling to save time and resources.Reference Büyüköztürk, Kılıç Çakmak and Erkan Akgün 90 However, a disadvantage of this method lies in its cost and time constraints. To draw inferences about the target population and generalize findings from the random sample, the sample size must be sufficient.Reference Taherdoost 91 In this study, the sample size was calculated using the statistical formula proposed by Taherdoost (2017). According to this formula, the sample should consist of at least 384 participants for a 95% confidence interval with a 5% margin of error, and at least 1066 participants for a 97% confidence interval with a 3% margin of error. Additionally, it is recommended that the sample size be at least 10 times the number of items in the scale used.Reference Kline 92 , Reference Burns and Grove 93 During the development of the Disaster Risk Perception Scale (DRPS), a pilot study was conducted with 359 citizens aged 18 and over residing in the provinces of Elazig and Giresun. The main study was subsequently conducted with 786 participants. Therefore, the sample sizes in both the pilot and main studies were adequate for the target population.
Study area. The red lines show active earthquake faults, the yellow area shows Giresun province, and the green area shows Elazig province.
Data Collection Tools
In this study, a survey was used as the data collection tool, which consists of three sections. The first section includes participants’ socio-demographic characteristics, such as gender, age, marital status, education level, income level, residence city, disaster experience, and disaster volunteering. These variables constitute the independent variables of the study. The second section contains the Disaster Risk Perception Scale (DRPS) developed by the authors. The DRPS is a Likert-type 5-point scale (1 = Strongly Disagree, 5 = Strongly Agree). The third section includes the “General Disaster Preparedness Belief Scale (GDPBS)” developed by Inal et al. (2018)Reference Inal, Altintas and Dogan 94 to determine individuals’ level of preparedness for disasters. The Health Belief Model (HBM), which explains current behavior change and its maintenance, is frequently utilized in disaster preparedness studies. For instance, an adapted version of the HBM was applied in a study investigating community preparedness for floods in eastern Ethiopia.Reference Ejeta, Ardalan and Paton 95 The GDPBS is a tool aligned with the HBM, demonstrating scope and construct validity as well as reliability. It comprises 6 subdimensions: self-efficacy, cues to action, perceived susceptibility, perceived barriers, perceived benefits, and perceived severity, distributed across 31 items. The GDPBS has a high reliability with a Cronbach’s Alpha coefficient of 0.93. It employs a 5-point Likert scale (1 = Strongly Disagree, 5 = Strongly Agree). Moreover, the GDPBS was applied to staff at Yalova University in Turkey. The scale is available in both English and Turkish, with the Turkish version used in the referenced study.
Data Analysis
In this study, the data analysis was conducted using SPSS (Statistical Package for the Social Sciences) 25 and AMOS (Analysis of Moment Structures) 24 software. Descriptive statistical methods such as frequency (n), percentage (%), mean (M), and standard deviation (SD) were used for the descriptive analysis of the data. The construct validity of the data was examined using factor analysis. The suitability of the data for factor analysis was tested using Kaiser-Meyer-Olkin (KMO) and Bartlett’s sphericity tests in SPSS. The factor loadings of the scale items were determined through exploratory factor analysis (EFA). Principal Component Analysis (PCA) and varimax rotation were preferred for EFA by the researchers. Factors with eigenvalues greater than one were retained. Factors with factor loadings above 0.4 were included in the analysis. After EFA, SPSS was used to determine whether each factor could be grouped under a single factor, that is, PCA and varimax rotation methods were used for each factor. The normality of the data distribution was tested using the Kolmogorov–Smirnov test. Because the data met the assumption of normal distribution, parametric analysis techniques were employed. The Independent Samples t Test was used to determine if there was a statistically significant difference between two independent groups, whereas the ANOVA test was used to determine if there was a statistically significant difference among more than two independent groups. In this study, the LSD post hoc test was used to determine which groups showed significant differences when significant differences were observed among the groups in the ANOVA test. Pearson correlation analysis was performed to identify whether there was a significant relationship between two or more independent variables and to evaluate the strength and direction of any such relationship, if present. Multiple regression analysis was used to test whether the dependent variable was predicted by the independent variable. EFA was performed using SPSS to analyze the data, including mean (M), standard deviation (SD), normality test, and reliability analysis using Cronbach’s alpha (CA), and ANOVA test. Confirmatory Factor Analysis (CFA) was conducted using AMOS to assess the model fit. The significance level for statistical analyses was set at 0.05. Various parameters, such as x2/Degree of Freedom (CMIN/DF), Root-Mean-Square Error Approximation (RMSEA), Standardized Root Mean Square Residual (SRMR), Non-Normed Fit Index (NNFI), Goodness‐of‐fit Index (GFI), and Comparative Fit Index (CFI), were evaluated to assess the model fit between the observed data and the hypothesized model.
Item Pool Creation
The first step in item generation, commonly referred to as “item pool development,” is usually supported by theoretical considerations.Reference Morgado, Meireles and Neves 96 There are two primary approaches to item generation. The first approach is called “deductive” or “top-down” reasoning. The second approach is known as “inductive” or “bottom-up” reasoning. Deductive scale development uses a classification schema before collecting data. This approach requires a comprehensive review of the literature to understand the construct of interest and develop a theoretical definition of the studied structure.Reference Hinkin 97 In this study, the deductive method was employed. The item generation process involved an extensive review of the international literature. The literature search was conducted through academic databases using keywords such as disaster risk perception, risk perception according to types of disasters (earthquake, flood, landslide, etc.), risk perception, and disaster risk. The criteria, methods, models, and approaches used in studies determining disaster risk perceptions were examined. Numerous studies were utilized to create the item pool.Reference Havenaar, De Wilde and Van den Bout 98 , Reference Xu, Peng and Liu 99 , Reference Trumbo, Peek and Meyer 108 –Reference Peters, Slovic and Hibbard 113 , Reference De Dominicis, Fornara and Ganucci Cancellieri 100 –Reference Peng, Lin and Liu 107 The scale items were not organized for specific disaster types. They were prepared with a comprehensive perspective for earthquakes, floods, and landslides and were not tailored to specific risk profiles of different cities.
Determining the number of items that should be included in the item pool initially is challenging. More items are usually needed than the intended number of items to be included in the final scale. At this stage of scale development, because the nature of relationships between items is generally unknown, having a large number of items serves as a kind of insurance against weak internal consistency. It is recommended to start with an item pool that is three or four times larger than the final scale.Reference DeVellis 114 Considering the factors that should be considered when creating the item pool, a total of 53 items were generated.
Validity
A study is considered valid if it measures what it claims to measure and does not contain logical errors in drawing conclusions from the data.Reference Garson 115 Content validity explains the extent to which each item in the scale represents the intended construct.Reference Cronbach and Meehl 116 To assess the content validity of a scale, expert opinions regarding the representativeness of the scope of all items in the scale need to be sought.Reference Allen and Yen 117 , Reference Lawshe 118 The content validity and face validity of the items in the scale were evaluated by obtaining expert opinions and calculating the Content Validity Index (CVI). The compatibility of each item with the overall scale was determined, and some items were revised, while others were excluded from the scale. Looking at the calculation of CVI, it can be seen that there are different calculation methods in the literature. In this study, the technique proposed by Davis (1992)Reference Davis 119 was preferred, and 20 experts provided their opinions for each item ("1: Relevant," “2: Item needs slight revision,” “3: Item needs substantial revision,” and “4: Item is not relevant/should be excluded”). The CVI for each item was calculated by dividing the number of experts who marked “1: Relevant” and “2: Item needs slight revision” options by the total number of experts. Items with CVI lower than 0.80 need to be removed from the scale.
In some studies where the calculation of CVI is explained, it is not explicitly specified whether it refers to the CVI for each item or the CVI for the entire scale. This situation is referred to in the international literature as “I-CVI, item-level content validity index” and “S-CVI, scale-level content validity index.”
In the Davis (1992)Reference Davis 119 technique, the results were not statistically calculated, and a direct value of 0.80 was accepted as the criterion. Another expert opinion evaluation method is the Lawshe (1975)Reference Lawshe 118 technique. According to this method, the content validity ratios (CVR) are calculated based on a statistical formula related to whether the items are included in the scale or not, based on content validity.Reference Lawshe 118
According to the number of experts, the Content Validity Ratio (CVR) value varies. At the 0.05 significance level, the minimum/critical values of CVRs, i.e., the CVR value when the number of experts is 20, is 0.50.Reference Ayre and Scally 120
Based on the evaluation of I-CVI results, two items with CVR values below 0.80 and CVI values below 0.50 were removed from the scale. All other items had CVR values above 0.80 and CVI values above 0.50.
In this study, opinions were obtained from experts in the fields of education sciences, mental health and psychiatry, clinical psychology, linguistics, and disaster management. The expert opinions were evaluated using both the Davis (1992)Reference Davis 119 technique and the Lawshe (1975)Reference Lawshe 118 technique, and the results obtained from both methods were found to be consistent. The obtained S-CVI value (0.96) being greater than the CVR value (0.50) indicates that the content validity of the scale items is statistically significant.Reference Lawshe 118
Results
Pilot Study
In the pilot study, 359 participants were reached. Table 1 presents the characteristics of the participants. 61.28% (220) of the participants were female, and 38.72% (139) were male. Among the participants, 61.28% (220) were in the age group of 18-24, 23.41% (84) were in the age group of 25-34, 9.19% (33) were in the age group of 35-44, 3.06% (11) were in the age group of 45-54, and 3.06% (11) were 55 years and above. Regarding marital status, 27.02% (97) were married, and 73.98% (262) were single. In terms of education, 5.85% (21) of the participants had primary education, 30.36% (109) had high school education, 24.79% (89) had associate degree, 36.21% (130) had a bachelor’s degree, and 2.79% (10) had postgraduate education. In relation to monthly average income, 49.58% (178) had an income of 2000TL and below, 23.68% (85) had an income of 2000-4000TL, 15.88% (57) had an income of 4000-6000TL, 8.08% (29) had an income of 6000-8000TL, and 2.78% (10) had an income of above 8000TL. Concerning the place of residence, 35.65% (128) of the participants were living in Elazig, and 64.35% (231) were living in Giresun. Moreover, 54.60% (196) of the participants had experienced a disaster before, and 86.07% (309) expressed their willingness to volunteer in case of any disaster occurrence.
Individual characteristics of participants (pilot study)
In the pilot study, factor analysis was conducted initially, and the KMO value was found to be 0.929, indicating the adequacy of the data for factor analysis. The significant result of the Bartlett’s Test of Sphericity (P = 0.000) also confirmed the suitability of the data for factor analysis. After confirming the suitability of the data with KMO and Bartlett’s Test, an EFA was performed to determine the construct validity of the DRPS. The most commonly used method for determining the factors is PCA.Reference Kleinbaum, Kupper and Muller 121 , Reference Büyüköztürk 122 PCA is a method used to reduce a multidimensional data to a lower dimension while maximizing the variance.Reference Tabachnick, Fidell and Ullman 123 Hence, PCA was preferred in the EFA. To facilitate the identification and interpretation of the related items, a varimax rotation was applied, which helps the related items to align and have a clear structure.
In the EFA, the Pattern Matrix revealed that the 51 items of the DRPS were grouped under 5 factors, explaining a total variance of 51.69%. The accepted threshold for the total explained variance is 40% and above. The factor loadings of the items were reported to be at an acceptable level of 0.40 and above.Reference Fabrigar and Wegener 124 In the pilot study, the cutoff point for the factor loadings of the items was set at 0.40.
During the EFA, it is possible for an item to load on multiple factors. In such cases, if an item is to be included in one specific factor, the difference between its factor loading in that factor and the factor loading in the other factor should be at least 0.10. If the difference is not at least 0.10, the item is considered a cross-loading item.Reference Cudeck 125 In this research, the cross-loading items were not included in the scale.
Based on the results of the EFA, items with low factor loadings and cross-loadings were systematically excluded, resulting in a revised analysis. Ultimately, 25 items were retained in the scale.
According to the results of the Exploratory Factor Analysis (EFA) in Table 2, it was determined that the remaining 25 items in the scale were placed in their respective dimensions, that their factor loadings in the corresponding dimensions were above 0.40, and that the difference between their factor loadings in their respective dimensions and other dimensions was less than 0.10. The total explained variance in the scale increased to 64.07%, which is higher than the 51.70% obtained with 51 items. The variance ratios for the dimensions were determined as follows: 20.67% (exposure/impact), 14.09% (worry/fear), 11.29% (vulnerability), 9.67% (uncontrollable), and 8.35% (probability). The overall Cronbach’s Alpha coefficient for the scale is 0.92, and the Cronbach’s Alpha coefficients for the sub-dimensions are 0.90 (exposure/impact), 0.75 (worry/fear), 0.84 (vulnerability), 0.86 (uncontrollable), and 0.83 (probability). The item-total correlations for all items were found to be higher than 0.30 (ranging from 0.39 to 0.69).
The pattern matrix of the Disaster Risk Perception Scale (pilot study)
CA = Cronbach’s alpha; VE = Variance Explained; TVE = Total Variance Explained.
The developed DRPS was subjected to CFA using the maximum-likelihood approach to test whether the remaining 25 items fit the 5-dimensional structure obtained from the EFA. The CFA results confirmed the 5-factor structure with good fit indices, as recommended by Hu and Bentler (1999).Reference Hu and Bentler 126 The fit indices were as follows: x 2/df = 1.951, SRMR = 0.063, GFI = 0.900, NNFI = 0.937, CFI = 0.946, RMSEA = 0.052, and the inter-factor correlations ranged from 0.30 to 0.62. These results indicate that the data fit well with the five-factor model.
In the pilot study phase, preliminary support was provided for the five-factor DRPS. Based on the results of both EFA and CFA, the initial pool of 53 items was reduced to 25 items. Thus, the DRPS was developed within the context of Turkish culture. With its 5-factor and 25-item structure, the DRPS was found to be valid and reliable according to the results of both AFA and DFA (H1 hypothesis was accepted).
Main Study
In the scale development process, a 25-item 5-factor DRPS was developed based on the results of the pilot study with 359 participants. In the main study, a total of 786 participants were reached. Table 3 presents the characteristics of the participants.
Individual characteristics of participants (main study)
Of the 786 participants, 49.49% (389) were female and 50.51% (397) were male. Regarding age distribution, 46.82% (368) were between 18-24 years old, 32.70% (257) were between 25-34 years old, 13.61% (107) were between 35-44 years old, and 6.87% (54) were 45 years old and above. In terms of marital status, 37.02% (291) were married, while 62.98% (495) were single. Regarding educational background, 16.92% (133) had primary education, 33.84% (266) had high school education, 29.77% (234) had associate degree, and 19.47% (153) had a bachelor’s degree. In terms of monthly income, 34.99% (275) of the participants earned 2000TL and below, 38.30% (301) earned between 2000-4000TL, 17.30% (136) earned between 4000-6000TL, and 9.41% (74) had a monthly income above 6000TL. As for the living location, 49.11% (386) of the participants resided in Elazig, while 50.89% (400) were from Giresun. Additionally, 48.85% (384) of the participants had previously experienced a disaster, and 86.90% (683) expressed their willingness to volunteer in the event of any disaster.
After determining the 25-item 5-factor structure of the DRPS in the pilot study, we conducted DFA with 786 participants using the maximum-likelihood approach in the main study. The DFA conducted in the main study confirmed the model structure that we identified in the pilot study, with goodness-of-fit values supporting the model’s adequacy. According to the results of DFA, which were evaluated based on the criteria proposed by Hu and Bentler (1999)Reference Hu and Bentler 126 the data demonstrated a good fit to the 5-factor model of the DRPS (x 2/df = 3.269, SRMR = 0.056, GFI = 0.924, NNFI = 0.900, CFI = 0.914, RMSEA = 0.054).
Upon examination of Table 4, it can be observed that the item with the highest M value among the DRPS items is “I can suffer economically in my region if a disaster occurs” (M: 4.26, SD: 0.935), which belongs to the EI (Exposure/Impact) dimension. On the other hand, the item with the lowest M value is “I feel vulnerable due to my physical inadequacy when a disaster occurs in my region” (M: 3.29, SD: 1.359), which is under the V (Vulnerability) dimension.
Findings of the Disaster Risk Perception Scale (main study)
**P<0.01; r = item-total correlation; CA = Cronbach’s alpha; M = Mean; SD = Standard Deviation; SE = Standardized Estimates
According to the results of the DFA, it can be observed that the factor loadings of the 25 items in the scale to their respective 5 dimensions are higher than 0.40, and that all the items’ t-values are significant. The overall Cronbach Alpha coefficient for the entire scale is 0.87, and the Cronbach Alpha coefficients for the dimensions are 0.86 (Exposure/Impact), 0.71 (Probability), 0.79 (Uncontrollable), 0.77 (Worry/Fear), and 0.73 (Vulnerability) respectively. The item-total correlations for all items are higher than 0.20 (ranging from 0.30 to 0.53). Based on the findings of the validity and reliability analysis, the 25-item, 5-dimension (Exposure/Impact, Probability, Uncontrollable, Worry/Fear, Vulnerability) structure of the DRPS is deemed reliable and valid.
The main study phase provided further evidence on the reliability and validity of the DRPS. The results of the main study demonstrated that the DRPS had satisfactory reliability and validity performance in a diverse sample of 786 individuals (H1 hypothesis was supported).
GDPBS Findings, Comparative Analysis According to Demographic Variables and Multiple Regression Analysis
The validity and reliability of the 31-item, 6-dimensional GDPBS (General Disaster Preparedness Belief Scale) developed by Inal et al. (2018) were subjected to analysis. In the initial results of the DFA, it was found that the factor loadings and model fit index values were not within acceptable ranges (x 2/df = 9.261, SRMR = 0.154, GFI = 0.699, NNFI = 0.528, CFI = 0.575, RMSEA = 0.103). Therefore, the items with low factor loadings were gradually removed from the scale, and the analyses were repeated. In the final step of the DFA, it was determined that 21 items remained in the scale, and with the implementation of recommended covariance connections (Item 3 with 5, Item 20 with 21), the fit index values reached good and very good \levels (x 2/df = 3.271, SRMR = 0.054, GFI = 0.935, NNFI = 0.905, CFI = 0.922, RMSEA = 0.054, Factor loadings = 0.48/0.84).
It can be observed that after the DFA stages, the 21 remaining items in the scale had factor loadings of 0.40 or higher in their respective 6 dimensions, and that all items had significant t-values. The overall Cronbach’s Alpha coefficient for the scale is 0.78, while the Cronbach’s Alpha coefficients for the subdimensions are 0.70 (SE), 0.72 (CA), 0.71 (PSu), 0.82 (PBa), 0.70 (PBe), and 0.81 (PSe), respectively. The item-total correlations for all items are above 0.20 (ranging from 0.20 to 0.48). Based on the findings of the validity and reliability analyses, it can be concluded that the disaster preparedness scale with its 21 items and 6-dimensional structure is a reliable and valid measurement tool.
When examining the M and SD values of the items in the GDPBS, it can be observed that the item with the highest M value is “Thinking about the possibility of an emergency/disaster scares me” under the PSe subdimension (M: 4.08, SD: 1.256), and the item with the lowest M value is “I have much more important responsibilities than preparing for emergencies/disasters” under the PBa subdimension (M: 3.00, SD: 1.376).
The DRPS and the GDPBS, along with their dimensions and subdimensions, were subjected to comparative analyses to examine whether there were significant differences based on demographic variables. The Kolmogorov-Smirnov test indicated that the data showed a normal distribution; hence, the Independent Samples t test and ANOVA were used. The Independent Samples t test results revealed that there were no significant differences between gender and Exposure/Impact (t = -0.54; P = 0.589), Probability (t = -0.63; P = 0.531), Uncontrollable (t = 1.29; p = 0.199), Worry/Fear (t = 0.65; p = 0.513), Perceived benefits (t = -0.09; P = 0.930), Disaster risk perception (t = 1.41; P = 0.159), and Disaster preparedness (t = 0.76; P = 0.450). However, there were significant differences between gender and Vulnerability (t = 3.29; P = 0.001), Self-efficacy (t = -2.00; P = 0.045), Cue to actions (t = 1.98; P = 0.048), Perceived susceptibility (t = -1.98; P = 0.048), Perceived barriers (t = 2.19; P = 0.029), and Perceived severity (t = 5.01; P = 0.000). The female participants had higher mean scores in Vulnerability, Cue to actions, Perceived barriers, and Perceived severity compared to the male participants. On the other hand, the male participants had higher mean scores in Self-efficacy and Perceived susceptibility compared to the female participants.
There was no significant difference between disaster experience and probability (t = 1.05; P = 0.295), uncontrollable (t = -1.53; P = 0.128), cue to actions (t = -1.88; P = 0.060), and disaster risk perception (t = 0.79; P = 0.433). However, there was a significant difference between disaster experience and exposure/impact (t = 7.19; P = 0.000), worry/fear (t = 4.05; P = 0.000), vulnerability (t = -4.96; P = 0.000), self-efficacy (t = 2.90; P = 0.004), perceived susceptibility (t = 3.78; P = 0.000), perceived barriers (t = -8.92; P = 0.000), perceived benefits (t = 4.43; P = 0.000), perceived severity (t = -8.51; P = 0.000), and disaster preparedness (t = 2.13; P = 0.034). The participants with disaster experience had higher average scores in exposure/impact, worry/fear, self-efficacy, perceived susceptibility, perceived benefits, and disaster preparedness compared to those without disaster experience. On the other hand, the participants without disaster experience had higher average scores in vulnerability, perceived barriers, perceived severity, and disaster preparedness compared to those with disaster experience.
There was no significant difference between the willingness to volunteer in any disaster and probability (t = 1.69; P = 0.091), uncontrollable (t = 0.15; P = 0.878), vulnerability (t = 0.10; P = 0.920), self-efficacy (t = 1.80; P = 0.073), perceived barriers (t = -1.38; P = 0.169), and disaster risk perception (t = 1.88; P = 0.060). However, there was a significant difference between the willingness to volunteer in any disaster and exposure/impact (t = 2.20; P = 0.028), worry/fear (t = 3.22; P = 0.001), cue to actions (t = 4.02; P = 0.000), perceived susceptibility (t = 2.40; P = 0.016), perceived benefits (t = 2.49; P = 0.013), perceived severity (t = 2.39; P = 0.017), and disaster preparedness (t = 4.06; P = 0.000). The mean scores for exposure/impact, worry/fear, cue to actions, perceived susceptibility, perceived benefits, perceived severity, and disaster preparedness were higher in those participants willing to volunteer in any disaster compared to those who are not willing to volunteer.
The ANOVA results indicate that there was no significant difference between age and probability (F = 0.08; P = 0.972), uncontrollable (F = 2.18; P = 0.089), vulnerability (F = 0.52; P = 0.670), perceived benefits (F = 1.18; P = 0.316), and disaster risk perception (F = 1.90; P = 0.128). However, there was a significant difference between age and exposure/impact (F = 7.94; P = 0.000), worry/fear (F = 3.29; P = 0.020), self-efficacy (F = 5.99; P = 0.000), cue to actions (F = 4.97; P = 0.002), perceived susceptibility (F = 5.91; P = 0.001), perceived barriers (F = 6.98; P = 0.000), perceived severity (F = 6.15; P = 0.000), and disaster preparedness (F = 11.02; P = 0.000). To determine which groups showed significant differences, LSD post hoc tests were conducted. The participants aged 35 and above had higher mean scores in exposure/impact, self-efficacy, perceived susceptibility, and disaster preparedness compared to the participants aged 34 and below. Those participants aged 35-44 have higher mean scores in worry/fear compared to those participants aged 25-34. The participants aged 35-44 had higher mean scores in cue to actions, perceived severity, and disaster preparedness compared to the participants aged 34 and below. The participants aged 44 and below had higher mean scores in perceived barriers and disaster preparedness compared to the participants aged 45 and above.
There was no significant difference between educational level and probability (F = 1.52; P = 0.209), self-efficacy (F = 2.21; P = 0.086), perceived susceptibility (F = 1.14; P = 0.333), perceived benefits (F = 2.02; P = 0.110), and disaster preparedness (F = 0.91; P = 0.434). However, there was a significant difference between educational level and exposure/impact (F = 3.92; P = 0.009), uncontrollable (F = 2.92; P = 0.033), worry/fear (F = 2.96; P = 0.032), vulnerability (F = 8.24; P = 0.000), cue to actions (F = 3.86; P = 0.009), perceived barriers (F = 8.15; P = 0.000), perceived severity (F = 10.90; P = 0.000), and disaster risk perception (F = 3.05; P = 0.028). LSD post hoc tests were conducted to determine which groups showed significant differences. The participants with associate and bachelor’s degrees had higher mean scores in exposure/impact and worry/fear compared to the participants with high school education. The participants with associate’s degrees had higher mean scores in uncontrollable and disaster risk perception compared to the participants with high school and bachelor’s degrees. The participants with primary school, high school, and associate’s degrees had higher mean scores in vulnerability compared to the participants with bachelor’s degrees.
In general, there were significant differences in disaster risk perception, educational level, and disaster volunteering status among the participants, while there was no significant difference based on gender, age, and disaster experience. According to the results of the independent samples t-tests and ANOVA, H2 hypothesis was inconsistent.
Since the data showed a normal distribution, the Pearson correlation analysis was conducted. There was a positive and significant relationship between the DRPS and the GDPBS (r = 0.29; P<0.01) (H3 hypothesis was accepted). There were positive and significant relationships between the DRPS and the Exposure/Impact (r = 0.56; P<0.01), Probability (r = 0.69; P<0.01), Uncontrollable (r = 0.68; P<0.01), Worry/Fear (r = 0.64; P<0.01), and Vulnerability dimensions (r = 0.68; P<0.01). Additionally, there were positive and significant relationships among the dimensions of the DRPS. The internal consistency performance of the DRPS dimensions was acceptable, as the alpha levels of all dimensions were close to the recommended 0.70.Reference Nunnally and Bernstein 127 In other words, all correlation coefficients among the factors of the DRPS were significantly positive, indicating that these five factors constitute a higher-level structure; in essence, they formed the DRPS.
According to the results of the multiple regression analysis, we examined the impact of disaster risk perception, as the independent variable, on disaster preparedness, as the dependent variable. Since the uncontrollable and vulnerability dimensions in the DRPS did not show a significant relationship with the GDPBS, they were not included in the multiple regression analysis conducted to assess the effect of disaster risk perception on disaster preparedness. The initial model was found to be appropriate (F(5.782) = 94.90; P = 0.000), and there was no issue of multicollinearity among the independent variables (Tolerance>0.20; VIF<10).Reference Hair, Black and Babin 128 Together, the variables of disaster risk perception explained approximately 27% of the variance in disaster preparedness (Adjusted R 2 = 0.267). According to the model, the disaster risk perception factors that have a significant impact on disaster preparedness were exposure/impact and worry/fear, respectively. The exposure/impact (β = 0.34; t = 9.16; P = 0.000) and worry/fear (β = 0.24; t = 6.58; P = 0.000) variables were found to have positive and significant effects on disaster preparedness, while the probability variable was not found to have a significant effect (P = 0.836) (H4 hypothesis was not supported). The higher averages of exposure/impact and worry/fear lead to higher levels of disaster preparedness.
Discussion
Decision-makers in disaster risk management should investigate the risk cultures and risk perceptions of citizens. Understanding people’s behaviors in risk situations is the most crucial aspect of the risk management process. This way, more effective damage-reducing measures can be taken, and by preparing for disasters, people can feel safer against disasters. The earthquakes in Elazig and the floods and landslides in Giresun in 2020 highlighted the importance of disaster preparedness. In such disasters, it is essential to separately investigate people’s levels of disaster preparedness and their perceptions of disaster risk.
Inconsistent with hypothesis H2, disaster risk perception did not significantly differ based on individuals’ gender. However, this finding does not align with the existing literature. For instance, Lovekamp and Tate (2008)Reference Lovekamp and Tate 129 found that female college students had significantly higher levels of fear related to hurricanes and earthquakes compared to male college students. Additionally, studies showed that females had higher disaster risk perceptions than males concerning floods,Reference Kellens, Zaalberg and Neutens 104 , Reference Perić and Cvetković 130 earthquakes,Reference Tekeli-Yeşil, Dedeoǧlu and Braun-Fahrlaender 131 terrorism,Reference Bourque, Regan and Kelley 132 , Reference Eisenman, Wold and Fielding 133 and natural hazards.Reference Bronfman, Cisternas and López-Vázquez 134 When examining the literature containing findings that contradict our research results, it is evident that these studies are from earlier years. Previous research often categorized women as a more sensitive and vulnerable group. However, the findings of this study highlight the importance of avoiding generalizations, demonstrating that not all women should be considered a vulnerable group and that they do not all face the same disadvantages. Therefore, future research should aim to explore disaster risk perception theory more comprehensively and in greater detail specifically for women.
Inconsistent with hypothesis H2, disaster risk perception did not significantly differ based on individuals’ age groups. However, there are conflicting results in the literature, particularly regarding flood risk perception, where some studies found similar outcomes and others found different outcomes. For instance, Qasim et al. (2015)Reference Qasim, Nawaz Khan and Prasad Shrestha 135 did not find a significant relationship between flood risk perception and participants’ age and income status. Some studiesReference Grothmann and Reusswig 136 –Reference Zaalberg, Midden and Meijnders 138 showed no significant correlation between age and flood risk perception, while othersReference Kellens, Zaalberg and Neutens 104 , Reference Botzen, Aerts and Den Bergh 139 found a significant relationship. However, it is well-known that individuals aged 65 and above are more vulnerable to disasters and emergencies.Reference Granger, Jones and Leiba 140 Studies conducted on earthquake risk perception in Bucharest revealed that people aged 60 and above were the most vulnerable.Reference Armaş 79 , Reference Armaş 141 In previous studies, participants’ age ranges were somewhat broader. In our study, the age variable was divided into five groups: 18-24 years, 25-34 years, 35-44 years, 45-54 years, and 55 years and above. The discrepancy between our study’s results and the literature might be attributed to the narrower age intervals used in our analysis. Additionally, it was observed that most studies, including this one, had very few participants aged 65 and above. In this research, only 3.1% of the participants were aged 55 and older. Therefore, future studies on disaster risk perception are recommended to focus specifically on older adults (55 years and above).
Consistent with hypothesis H2, disaster risk perception significantly varied based on individuals’ educational levels, which is in line with findings in the literature. Indeed, people’s level of education is closely related to their awareness of disaster risks. Individuals with higher levels of education tend to have higher levels of disaster risk awareness.Reference Ainuddin, Kumar Routray and Ainuddin 142 This result is highly consistent with previous studies.Reference Armaş 79 , Reference Paul and Bhuiyan 143 , Reference Armaş and Avram 144 However, it should be noted that there is also a counter-argument in some studies, suggesting a weak relationship between certain socioeconomic factors, such as educational level, and risk perception.Reference Armaş 141
Inconsistent with hypothesis H2, disaster risk perception did not significantly differ based on individuals’ disaster experiences. However, the exposure/Impact dimension did show statistically significant variations depending on individuals’ disaster experiences. These findings align with mixed results in the literature. Severe and frequent disaster experiences leave a lasting and profound impression on individuals, leading them to overestimate the likelihood of future disasters. For instance, those who have successfully escaped a landslide will have greater confidence in their abilities and believe that the impact of disasters can be controlled through human efforts.Reference Yang, Tan and Peng 110 People often base their thoughts on similar experiences they have encountered. Additionally, some research suggests that experience is a significant factor influencing individuals’ risk perceptions and hazard mitigation behaviors.Reference Jackson 145 , Reference O’Connor, Yarnal and Dow 146 For example, a community that has experienced flooding, especially those working in water systems, is more likely to perceive higher probabilities of future floods.Reference O’Connor, Yarnal and Dow 146
Inconsistent with hypothesis H2, disaster risk perception did not significantly differ based on individuals’ willingness to volunteer in any disaster scenario. This finding aligns with similar results in the literature. Miceli et al. (2008)Reference Miceli, Sotgiu and Settanni 137 reported that there was no significant relationship between participation in civil defense activities and flood risk perception. It is important to consider that various factors, such as the level of disaster knowledge, education, and participation in disaster drills, can influence disaster risk perception among disaster volunteers. In the survey, the participants were asked if they would like to volunteer in a potential disaster scenario. However, this study did not differentiate between individuals who were actual disaster volunteers and those who expressed a willingness to participate in voluntary activities. This indicates the need to distinguish between disaster volunteers and individuals who express a willingness to participate in volunteer activities and to conduct further detailed research to explore potential differences between these groups.
Consistent with hypothesis H3, there was a significant and positive relationship between disaster risk perception and disaster preparedness. This finding is in line with similar results in the literature. High levels of disaster preparedness contribute to individual and community resilience, enabling better response and reducing negative outcomes during a disaster.Reference Paton, McClure and Bürgelt 147 It is important to note that increasing individual awareness alone may not necessarily lead to a higher level of disaster preparedness. However, there is evidence of a significant relationship between disaster experience and disaster preparedness.Reference Bradford, O’Sullivan and Van Der Craats 148 People’s high level of preparedness for disasters can be explained by their desire to avoid unexpected negative emotions such as insecurity, fear, and helplessness during a disaster.Reference Terpstra 149 Furthermore, it is well-established that higher disaster risk perception is associated with increased disaster preparedness.Reference Akbar, Suryaratri and Tri 150 , Reference Yong, Lemyre and Pinsent 151
Consistent with hypothesis H4, there was a significant and positive effect of the worry/fear dimension on disaster preparedness. This finding is consistent with similar results in the literature. One of the most crucial components of disaster risk perception is the disaster fears that hold significant importance in the risk assessment process.Reference Cvetković, Öcal and Ivanov 152 It is known that there is a significant and positive relationship between disaster risk perception and fear, with fear being the major factor influencing disaster risk perception.Reference Khan, Rana and Nawaz 111 , Reference Peng, Tan and Lin 153 As a result, these fears can play a significant role in motivating individuals to take preventive measures to protect their lives and property.Reference Lerner and Keltner 154
Inconsistent with hypothesis H4, there was no significant and positive effect of the uncontrollable dimension on disaster preparedness. This finding is in line with different results reported in the literature. Unforeseeable and unpredictable events, often perceived as uncontrollable, can be easily reversed in people’s minds when making decisions.Reference Tversky and Kahneman 155 , Reference Tversky and Kahneman 156 When individuals perceive a disaster as being out of their control, they tend to attribute the occurrence of the disaster to fate. On the contrary, if they believe that their efforts can mitigate the impact of disasters, they will not see disasters as solely dependent on fate.Reference Slovic 54 , Reference Peng, Lin and Liu 107 Moreover, the belief that disasters can be controlled or managed also reduces the fear factor.Reference Slovic and Peters 43 Additionally, individuals with a high sense of controllability may underestimate the effects of disasters, leading them to be less likely to take precautionary measures to cope with risks.Reference Xu, Peng and Liu 99
Limitations
Our study has some limitations. The research was limited to the provinces of Giresun and Elazig only. Although an appropriate sampling method was chosen, the results should be interpreted as a preliminary preparation in terms of generalization. Additionally, the time and cost limitations associated with the random convenience sampling method are among its disadvantages. The survey was conducted both online and face-to-face due to the constraints imposed by the COVID-19 pandemic. When discussing the results of the study, it should be noted that a small portion of the data was collected online. Future studies in the context of strengthening disaster risk management in Turkey can continue to be built upon these results. Lastly, the quantitative method used for assessing disaster risk perception may not be sufficient. Qualitative methods can reveal a more comprehensive dimension of the findings in this study. Therefore, mixed or qualitative methods would be beneficial in further elaborating the findings of this study.
Conclusion and Recommendations
This study provided a valid and reliable scale for determining citizens’ disaster risk perception. The disaster risk perception scale, consisting of 25 items and 5 dimensions (Exposure/Impact, Probability, Uncontrollable, Worry/Fear, Vulnerability), were found to be a reliable and valid tool (See Appendix).
The female participants had significantly higher vulnerability scores compared to the male participants. The participants with previous disaster experience had higher averages in exposure/impact and worry/fear dimensions compared to those without disaster experience. The participants aged 35 and above had higher averages in exposure/impact and disaster preparedness compared to those aged 34 and below. The participants who expressed willingness to volunteer in any disaster had higher averages in exposure/impact, worry/fear, and disaster preparedness dimensions compared to those who did not express willingness to volunteer.
It is recommended to conduct comprehensive studies to explore citizens’ disaster risk perception and risk cultures to increase disaster preparedness levels by involving the public in the disaster risk management process and adopting a community-based disaster management system. Additionally, vulnerable groups such as women and the elderly, who have higher levels of vulnerability during disasters, should not be overlooked. In disaster preparedness and disaster risk perception studies, the values and beliefs of the community should be taken into consideration.
Disaster managers, policymakers, and scientists can utilize the Disaster Risk Perception Scale (DRPS) to measure citizens’ disaster risk perception. While the scale’s items were not specifically tailored for any particular type of disaster and did not include risk profiles of specific regions or provinces, it can be adapted and expanded for other types of disasters (such as earthquakes, floods, landslides, etc.).
Disaster risk perception plays a significant role in disaster risk management and governance. People living under the risk of disasters need to have high levels of disaster risk perception. The Disaster Risk Perception Scale, developed to determine people’s reactions to disasters, their behaviors before, during, and after disasters, the type of risk cultures present, and the factors influencing risk cultures, is of vital importance. Moreover, the developed scale for disaster risk perception will contribute to strengthening disaster risk management.
For future research, it is recommended to apply the disaster risk perception scale to different types of disasters and evaluate the results. In the field of disaster risk perception, one of the significant component of strengthening disaster risk management, it is suggested to apply the scale to samples with different cultural values and beliefs, taking cultural values into consideration.
Acknowledgements
This study in based on the PhD dissertation of the first author.
Author contribution
Both authors read and approved the final manuscript.
Funding statement
The authors report no funding.
Competing interest
The authors have no conflicts of interest to declare that are relevant to the content of this article.
Ethical standards
The participants in this study volunteered to participate. Additionally, the informed consent form provided on the first page of the survey informed the participants that the collected data would be kept confidential, and that they had the right to withdraw from the study at any stage. The Scientific Research and Publication Ethics Committee of Gumushane University approved this research from a scientific and ethical perspective. Permission was also obtained from the Elazig and Giresun Governorships to conduct the survey in the respective provinces.
Appendix. Disaster Risk Perception Scale (DRPS)
The table was reported in English (Turkish) style.
