Exploring Health Effects of Terrorism: A Multi-Level Analysis for Turkey

Abstract

Terrorism has the potential to affect population health through various pathways. Since the literature mostly analyzes tangible economic costs, there is dearth of evidence on health effects of terrorism. In an effort to address that gap, this article explores the relationship between terrorism and health satisfaction of Turkish citizens by combining province-level and individual-level data sets. In order to quantify determinants of health satisfaction, a multi-level modeling framework is employed. Empirical analysis suggests that individuals with higher exposure to terrorism are more likely to report lower health satisfaction in Turkey. Health satisfaction of individuals is significantly and positively correlated with individual-level covariates such as education level, marital status, employment status, household income, housing floor area per person, interest in health issues and becoming a parent within the past year. Age, being female and utilization of health services display negative associations with health satisfaction of individuals. Finally, province-level GDP per capita and schooling ratio exhibit significantly positive associations with individual health satisfaction in Turkey.

Headnote

Terrorism has the potential to affect population health through various pathways. Since the literature mostly analyzes tangible economic costs, there is dearth of evidence on health effects of terrorism. In an effort to address that gap, this article explores the relationship between terrorism and health satisfaction of Turkish citizens by combining province-level and individual-level data sets. In order to quantify determinants of health satisfaction, a multi-level modeling framework is employed. Empirical analysis suggests that individuals with higher exposure to terrorism are more likely to report lower health satisfaction in Turkey. Health satisfaction of individuals is significantly and positively correlated with individual-level covariates such as education level, marital status, employment status, household income, housing floor area per person, interest in health issues and becoming a parent within the past year. Age, being female and utilization of health services display negative associations with health satisfaction of individuals. Finally, province-level GDP per capita and schooling ratio exhibit significantly positive associations with individual health satisfaction in Turkey.

Keywords: terrorism, intangible costs, health satisfaction, multi-level model, Turkey

JEL codes: I10, D74, F50, O53, C20

Acknowledgements:

The authors would like to thank Julide Yildirim for her valuable support and suggestions, and Muserref Kucukbayrak for her contribution to the data collection process. The authors are also grateful to Hamza Suffian Abbasi and Esra Erzen for their comments on improving the exposition.

Terror activities pose a significant threat for many countries. Although Middle Eastern and African countries frequently experience terror incidents, European countries including France, Belgium and Germany also suffered occurrences of terrorism. The most prominent costs of terrorism are deemed to be psychological, economic, political and social. In this context, terrorism leads to inevitable costs for terror-targeted countries and their neighbors through spillover effects.

Terrorism has both tangible and intangible costs, which affect a country's economy through various channels. The short-run costs of terrorism consist of immediate losses of human and physical capital; increasing risks and uncertainties in decision-making processes of consumers and investors; and health problems of victims (Schneider et al. 2011). The longterm costs of terrorism include distortions in various types of economic activity and reductions in productivity (Abadie and Gardeazabal 2003; Schneider et al. 2011). The tangible costs of terrorism include declines in economic growth, national investment and national income (Blomberg et al. 2004; Gaibulloev and Sandler 2011; Eckstein and Tsiddon 2004; Fielding 2003; Bilgel and Karahasan 2017; Ocal and Yildirim 2010); increases in unemployment due to firm closures; and rising military and security expenditures (Greenbaum et al. 2006; Goldstein 2005; Omay et al. 2013). A terrortargeted country bears the economic costs not only within its domestic borders but also in its international engagements. Terror incidents reduce tourism revenues (Eisinger 2004; Frey et al. 2007; Drakos and Kutan 2003; Feridun 2016) and foreign direct investment inflows (Enders and Sandler 1993; Omay et al. 2013) and increase transaction costs in bilateral trade (Frey et al. 2007; Mirza and Verdier 2008).

Intangible costs of terrorism include, but are not limited to, reduced standard of living, increased level of stress, and emotional harm. Intangible costs include loss of individual well-being, which is measured by various specifications. Mitchell et al. (2005) define well-being as the difference between all sources of satisfaction and all sources of apprehension. Therefore, well-being may be defined in the context of individual satisfaction, which could be affected by terrorism. One branch of literature discusses the relationship between life satisfaction and terrorism. The findings indicate that incidences of terror directly lead individuals to state lower levels of life satisfaction in terror-targeted countries (Frey et al. 2007; Vorsina et al. 2017; Romanov et al. 2012). Recently, Yildirim et al. (2016) report that terrorism is detrimental to individual happiness of Turkish citizens and that individual happiness levels are lower in the eastern and southeastern provinces of Turkey. Moreover, terrorism has adverse impacts on life quality by worsening health outcomes. For instance, researchers found that mental health outcomes and stress levels of United States citizens were affected by the September 11, 2001, terrorism events (Richman et al. 2008; Galea et al. 2002; Schuster et al. 2001). Similarly, Bleich et al. (2003) found that individuals who were exposed to a terror incident exhibited at least one of the symptoms of depression in Israel.

Although the tangible costs of terrorism are frequently analyzed, analysis of intangible costs would provide fruitful information. This article focuses on the well-being costs of terrorism at the individual level by constructing a link between terrorism and health outcomes of individuals in Turkey. Terrorism has been a persistent issue in Turkey for decades, with both domestic and international-level attacks.1 According to the Institute for Economics and Peace (IEP) (2017), Turkey has the highest number of deaths from terrorism among members of the Organization for Economic Cooperation and Development (OECD) in the past decade. The Syrian civil war caused a turmoil in the Middle East and there was an increase in terrorist attacks in Turkey. These events created major difficulties in daily life and negatively affected the mental health of Turkish citizens. Increasing level of fear and stress, inconsistent sleep and work hours, and problems in accessing education, health and transportation services significantly affect the wellbeing of individuals in the country. This research sets out to explore the association of terrorism with health satisfaction levels of Turkish citizens. Being a crucial component of well-being, health outcomes are highly likely to be affected by terrorism.

In the aftermath of a terror incident, observing the well-being of the society or the region affected by an attack is vital for drawing robust conclusions about its health outcomes. Thus, analysis of "health outcomes of terrorism" at the individual level would provide crucial insights for policy-making processes. Countrylevel impacts comprise a combination of the terrorist attaack's direct effects on primary victims - who profoundly experience the effects of an attack - and its indirect effects on collateral victims in society at large. Public health policies of governments include programs for "preparedness against terrorism" in order to deal with short and long-term health implications. Threat of terror causes significant changes in both policy making and budget allocation decisions of governments. For instance, the United States made significant changes in federal budget allocation after the September 11 World Trade Center terrorist attack (Sidel and Levy 2003). Given that even the possibility of a future terrorist attack leads to changes in resource allocation, establishment of public medical organizations, and budget planning by governments, it should be acknowledged that presence of terrorism affects public health and economic policies of terrortargeted countries.

In addition to policy-making and resource allocation costs in the short run, terrorism potentially causes long-term economic costs related to formation of human capital and intergenerational transfers. In this regard, terrorism creates additional economic costs through its adverse impacts on individual-level health and educational outcomes. Researchers have found that the presence of conflict and violence leads to lower birth weights (Camacho 2008); lower fertility rates (Berrebi and Ostwald 2014); worsening of mental health (Bleich et al. 2003); lower educational attainment and resources (Black et al. 2007); and lower educational performance (Kibris 2015; Bruck et al. 2014; Shany 2016). These adverse effects of terrorism on early life outcomes and educational performance measures have repercussions on later life outcomes such as labor market performance at individual level and human capital formation at aggregate level. Even though the literature provides evidence for associations between terrorism and individual-level mental health outcomes and country-level macroeconomic indicators, research on the indirect impacts of terrorism on economic well-being through the link between health and economic outcomes of individuals is sparse. Hence, this study sets out to complement prior research by exploring health effects of terrorism for Turkish case.

A substantial body of literature focuses on determinants of objective and subjective health outcomes for individuals. The findings suggest that socioeconomic conditions and demographics such as income, education, employment status, gender, race, and age are significantly correlated with health outcomes (Gächter et al. 2012; Idler and Benyamini 1997; Denton and Walters 1999; Marmot et al. 1997; Case and Deaton 2005; Soytas and Köse 2014; Williams et al. 1997). Another branch of literature investigates the determinants of life satisfaction in Turkey and finds that income level, employment status, demographics, trust in public institutions, relative income level, relative consumption level, future expectations, and terrorism are significantly associated with life satisfaction of Turkish citizens (Selim 2008; Ekici and Koydemir 2014; Dumludag 2013; Caner 2014; Yildirim et al. 2016). Although the literature offers many determinants for self-rated health and life satisfaction levels, there is a lack of research focusing on the health effects of terrorism. This study combines an individual-level and a province-level data set to conduct an empirical analysis of the association between self-rated health satisfaction and exposure to terror incidents in Turkey.

The individual-level data set for this research is obtained from the 2013 Life Satisfaction Survey (LSS) conducted by the Turkish Statistical Institute (TSI). Province-level data were constructed for eighty-one provinces using the Global Terrorism Database (GTD) for terror measures and TSI regional statistics for province-level measures. Given the nested structure of the data at individual and province levels, a multilevel ordered logit estimation framework is employed on account of the ordered nature of the health satisfaction measure as the dependent variable. Empirical results indicate that exposure to terrorism is associated with lower health satisfaction of Turkish individuals. In addition, demographics, household-level, and province-level variables are significantly correlated with health satisfaction. Females and older people are less likely to be satisfied with their health status. On the other hand, education level, income level and being employed are positively correlated with health satisfaction. Additional factors such as new parenthood in the last year and being interested in healthrelated issues have positive correlations with health satisfaction of Turkish citizens. Any increase in the housing floor area per person is positively associated with level of health satisfaction. Regarding provincelevel variables, individuals living in provinces with higher school attendance and higher per capita GDP are more likely to report higher health satisfaction. The findings of the present research suggest that terrorism has a significant relationship with health outcomes of individuals; which in the longer term would be reflected in economic decision-making processes of households and overall economic activity of the society.

1.Theoretical Background

Terrorism may have an impact on health satisfaction of individuals both through physical damage and psychological effects on the society. Since terrorist attacks use lethal weapons, explosives, chemical substances, and incendiaries, they may have irremediable consequences for individuals' health status such as fatalities, non-fatal injuries and other diseases. Prior research finds that disaster-related physical and mental symptoms (including cancer, embolism, asthma, other respiratory diseases, gastroesophageal reflux, depression, emotional distress, post-traumatic stress disorder [PTSD], and behavior disorder) can last five to ten years after the date of event (North et al. 1999; Pfefferbaum et al. 2001; Stellman et al. 2008; Brackbill et al. 2009).

Economic impacts of terrorism can be also explained through its intergenerational health effects on human capital accumulation. After considering labor market performance measures and health outcomes in later stages of life, previous research finds that birthweight is a significant predictor. Low birth-weight causes not only type II diabetes, heart and lung diseases (Almond et al. 2005) but also results in lower cognitive abilities, lower intelligence quotient and learning disorders (Couzin 2002). Regarding sources of low birth-weight, statistical evidence illustrates that exposure to violence and conflicts during the first trimester of pregnancy diminishes birth-weight due to increased stress hormone in the mother's body (Camacho 2008). Hence, long-run impacts of terrorism are associated with human capital formation and future labor market outcomes. Poor health status, psychological stress and welfare concerns due to exogenous shocks are found to be significant determinants of individuals' decisions to get married, to have children, and to migrate for job opportunities (Blanc 2004; Woldemicael 2008; Guha-Sapir and D'Aoust 2011). Existing research indicates an association between fertility levels and social conflicts. Considering terrorism as an exogenous shock for societies, Berrebi and Ostwald (2014) suggest that terrorism first reduces female labor force participation rates and then constitutes a psychological stress leading to lower levels of fertility. Thus, both current and future formation of human capital level and labor market conditions are affected by terror incidents.

As a crucial component of human capital formation and other individual economic decision-making processes, educational outcomes are also affected by terrorist attacks. For instance, Black et al. (2007) show that in Israel lower levels of schooling attainment, increases in alcohol consumption, cigarette smoking and cannabis use are directly correlated with terror attacks. Kibris (2015) finds that university entrance exam scores of Turkish students living in provinces where the intensity of Turkish-Kurdish conflict is high are negatively affected by those events. On the other hand, evidence on the second Intifada period in Israel shows that terror attacks reduce not only school attendance rates due to fear, but also lower the probability of passing final high school exams for enrolling at a university (Bruck et al. 2014; Shany 2016). These studies emphasize that violence and conflicts lead to impairments in academic performance through two channels: serious deteriorations in psychological wellbeing of students resulting in lower educational attainment, and reduced education quality stemming from difficulties in attracting teaching staff and resources for particular regions.

There is also well-established relationship between health status and education level of individuals. Considering the fact that future economic outcomes depend on early life conditions, researchers argue that increasing levels of mental health disorder, safety concerns, and greater poverty in conflict-affected populations lead to reductions in access to education and affect long-term human capital stock (Lai and Thyne 2007; Gates et al. 2012; Poirier 2012; Shields and Paulson 2015). The presence of terrorism may interact with this relationship in the Turkish case, in the sense that exposure to terrorism can indirectly affect health outcomes and income levels making individuals unable to meet the cost of education or making them feel too anxious to attend educational institutions.

Demographic variables are associated with health status of individuals. Due to deterioration of health over time, older people are more likely to face health issues. Existing research finds an inverse association between age and health outcomes for several countries (Hu et al., 2016; Deeks et al. 2009). Growing older may thus be considered a natural fact of life that makes individuals less satisfied with their health outcomes. On the other hand, there are significant gender differences in health outcomes (Gächter et al. 2012; Soytas and Köse 2014). Terrorism may also cause women to report higher levels of dissatisfaction with their health outcomes. In addition to reduced fertility levels and lower birth weights stemming from terrorism-related stress hormones (Wadhwa et al. 1993; Rondo et al. 2003; Camacho 2008), some studies suggest that parental desire to protect daughters from social conflict may result in girls dropping out of school due to fear of rape, abduction, and rebellion (Shemyakina 2011; Singh and Shemyakina 2016; Buvini et al. 2014). Hence, health satisfaction of Turkish females may be related to job losses, lower income levels, and lower educational attainment associated with conflict and violence in parts of the country.

Previous studies reveal that married individuals are more likely to be both mentally and physically healthy (Kiecolt-Glaser and Newton 2001; Robles and Kiecolt-Glaser 2003). Employment status is associated with health outcomes (Ross and Mirowsky 1995; Case and Deaton 2005). Employed individuals are more likely to have health insurance and access to health services. Health literacy of individuals is positively related with health outcomes (Berkman et al. 2011; DeWalt and Hink 2009). Moreover, research indicates that parents and non-parents have different health outcomes (Umberson 1987) and that subjective health status is inversely correlated with use of medical services (Miilunpalo et al. 1997). Household resources, such as household income level, have significant associations with health outcomes of individuals (Smith 1999; Ettner 1996). The literature also discusses impacts of changing family size on health outcomes (Lundborg et al. 2014; Solari and Mare 2012). Smaller house size may cause individuals to live with other family members in an overcrowded and unhygienic setting. Crowded housing conditions may lead to spread of infectious diseases and worse health outcomes (Krieger and Higgins 2002).

In addition to quantifying level of terrorism and health satisfaction of individuals, this study therefore considers control variables at individual level, household level, and province level. Variables of interest and details of corresponding data are presented in the next section.

2.Data and Empirical Approach

This article exploits individual-level survey data from the 2013 Life Satisfaction Survey conducted by the Turkish Statistical Institute. Unlike previous surveys, this version of the data set provides province-level information on current location of participants and includes a significantly higher number of participants. Other versions of the data set have fewer participants and do not release location information, thus precluding panel data analysis at province level. The survey was conducted with adults aged 18 or older. It gathered data from 196,203 individuals in 125,720 households throughout Turkey. The questionnaire included 312 questions; 239 at individual level and 73 concerning household characteristics. Face-to-face interviews were conducted with individuals in participating households.

As the main focus of this study is health outcome of individuals, self-reported health satisfaction is the dependent variable for empirical analysis. In the survey, health satisfaction of individuals is measured on a Likert scale of 1 to 5, with 1 indicating "very satisfied" and 5 indicating "not satisfied at all." For ease of interpretation, and following the related literature, coding of the dependent variable is reversed, with 1 indicating "not satisfied at all" and 5 indicating "very satisfied." In order to analyze the relationship between terrorism and self-reported health, terrorism indexes were constructed for 81 provinces in Turkey using the Global Terrorism Database (GTD). The GTD provides detailed information on terrorism events throughout the world. This study considers four indicators of terrorism to establish terrorism indexes by province and year: (1) number of terror incidents, (2) number of fatalities, (3) number of injuries, and (4) occurrence of property damage. Detailed descriptions of these variables are provided in the Appendix, Table A1.

Table 1 shows terrorism indicators for Turkey across three years. There was a rise in the level of terrorism in Turkey in 2012, with 184 incidents in 24 provinces resulting in 248 deaths and 490 injuries. 2011 and 2013 saw relatively lower number of incidents, 51 and 41 respectively. The figures for fatalities, injuries and property damages are also relatively lower in 2011 and 2013.

Province-level terrorism indexes for 2011, 2012, and 2013 were constructed using principal component factor analysis. Table 2 presents corresponding eigenvalues and factor loadings for terrorism indexes. For all years, the factor solutions with the highest eigenvalues were retained for construction of terrorism indexes. These factor solutions account for 64, 84 and 66 percent of variation in terrorism indicators for 2011, 2012 and 2013 respectively. Additionally, all factor loadings are greater than 0.57 and satisfy reliability conditions suggested by the literature (Hair et al. 1998; Stevens 2012; Tabachnick and Fidelį 2007). Finally, Cronbach's Alpha values, given in Table 2, are higher than 0.80 indicating that the constructed measures of terrorism demonstrate high levels of reliability and internal consistency. The calculated terrorism indexes are rescaled between 1 (lowest level of terrorism) and 10 (highest level of terrorism). Thus, a higher terrorism index indicates higher exposure to terrorism in the corresponding province. In order to account for cumulative and longer-lasting effects of terrorism, this paper considers empirical models with terrorism indexes for 2011, 2012 and 2013. Moreover, additional estimates are provided with an average terrorism index which is the simple average of terrorism indexes for the corresponding years. Empirical results indicate that association of terrorism with health satisfaction is robust with respect to choice of the year for terrorism index.

In addition to the main variables of interest, the empirical approach of this study considers control variables at individual level, household level and province level. Individual-level control variables include age, gender, education level, marital status, employment status, interest in health issues, and utilization of health services. Household income and housing floor area per person are household-level measures. Province-level controls for economic conditions, education services and health services include variables such as GDP per capita, schooling ratio, hospital beds per capita, and population per family doctor. Detailed descriptions and summary statistics of all variables used in the analysis are summarized in the Appendix (Table A1 and Table A2).

Table 3 presents the distribution of respondents across all levels of health satisfaction for a selection of explanatory variables. The mean age in the sample is 44.29 with a minimum of 18 and a maximum of 99. The mean health satisfaction score is 3.57 with a standard deviation of 0.89. 62.96 percent of survey participants report that they are "satisfied" with their health while 13.83% are "not satisfied." Females constitute 57.66 percent of the sample. 58.75 percent of females state that they are "satisfied" with their health status whereas 16.63 percent are "not satisfied". Similarly, 68.69 percent of males indicate that they are "satisfied" and 10.01 percent are "not satisfied". Frequency distributions of health satisfaction categories are different across categories. 62.97 percent of married respondents report that they are "satisfied" and 13.97 percent of are "not satisfied". 72.81 percent of never married respondents are "satisfied" and 6.26 percent are "not satisfied". 45.35 percent of widowed respondents are "satisfied" and 26.16 percent are "not satisfied". Moreover, individual health satisfaction and education level are correlated. 72.82 percent of respondents with graduate degree report being "satisfied" and 5.10 percent are "not satisfied". However, 25.54 percent of respondents with no educational qualifications are "not satisfied" with their health status. Employment status is also associated with health satisfaction. 70.78 percent of employed respondents report that they are "satisfied" with their health status while 59.39 percent of unemployed respondents state that they are "satisfied". The distributions of health satis- faction categories are similar across levels of individual's interest in health-related issues. Respondents who became parents in the previous year are more likely to report better health satisfaction levels. Finally, individuals who used health services within the survey year were more likely to report lower health satisfaction levels.

Considering household-level variables, the proportion of individuals reporting higher levels of health satisfaction increases with household income. The average housing floor area per person for the sample is 35.17 m2 with a minimum of 1.11 m2 and a maximum of 500 m2.

Figure 1 illustrates spatial distribution of average health satisfaction levels for provinces in Turkey. The higher satisfaction scores are observed in provinces located in the middle of Black Sea territory, the East of Aegean area, and the West of Marmara region. At province level, Sinop has the highest average health satisfaction with a score of 3.752. The following best scores belong to Cankırı, Tekirdağ, Denizli and Kastamonu with average health satisfaction scores of 3.750, 3.715, 3.706, and 3.705, respectively. Lower health satisfaction scores are observed in the Eastern Anatolia, Southeastern Anatolia, and East Mediterranean regions. Tunceli has the lowest average health satisfaction with a score of 3.276. Mardin, Diyarbakır, Ardahan, and Iğdır follow with average scores of 3.293, 3.315, 3.339 and 3.376, respectively.

As Figure 2 shows terror attacks causing injuries and fatalities are more frequent in the Eastern and SouthEastern regions. The terrorism index at province level is the highest for Hakkari with a score of 10, followed by Şırnak (6.39), İstanbul (4.55), and Diyarbakır (3.21). Terror attacks target not only the Eastern and SouthEastern provinces but also the bigger provinces. For instance, İstanbul is the largest metropolitan province, and ranks third with respect to terrorism index.

Within an ordinal dependent variable, model estimation using Ordinary Least Square (OLS) would result in biased results. OLS requires a continuous dependent variable. The presence of an ordinal response variable violates the fundamental assumption. By ignoring ordered nature of the dependent variable, OLS estimation treats the difference between "not satisfied at all" and "not satisfied" in the same way as the difference between "satisfied" and "very satisfied". Hence, ordered choice frameworks need to be used for estimation, where the response variable is both categorical and ordinal. Moreover, since this paper combines both individual-level and province-level variables, the data exhibit a nested structure. Thus, multilevel treatment of the data with hierarchical models would produce more robust empirical results. In this research, a multi-level ordered logistic regression model is used for empirical analysis. Accounting for nested nature of individual-level and province-level variables, the multi-level ordered logistic framework is more appropriate than a conventional ordered logistic model. Stata 15 was used for data analysis and to estimate random intercept multi-level ordered logit models for health satisfaction in Turkey (StataCorp 2017). Empirical results are presented and discussed in the next section.

3.Empirical Results

Empirical findings are presented in Table 4, in the form of odds ratios for different models. The corresponding coefficients and multi-level ordered logit model statistics are given in the Appendix, Table A3. First, results of likelihood ratio (LR) tests imply that multi-level modeling should be preferred to standard ordered choice models. Wald tests indicate that all models are overall significant. Individual significance levels of estimated coefficients indicate that all individual-level explanatory variables have significant correlations with health satisfaction. At the province level, terrorism index, GDP per capita, and schooling ratio demonstrate significant associations with health satisfaction. Indicators of variance inflation factors (from a corresponding linear model) reveal that there are no serious multicollinearity issues in the empirical models. Furthermore, additional models which account for household-level hierarchy are also estimated for a robustness check of the empirical findings. Estimation results for these models (see Appendix Table A7) are very similar to empirical results provided by Table 4.

Odds ratios from multi-level ordered logistic models are presented in Table 4. All empirical models exhibit a significantly negative relationship between terrorism and health satisfaction in Turkey. Model 1 employs the terrorism index for 2011 and reveals that an instantaneous change in the terrorism index is associated with a 6 percent decline in the odds of reporting a higher level of health satisfaction. Similarly, a unit increase in terrorism index for 2012 is related to a 3.9 percent reduction in the odds of reporting a higher level of health status. A unit rise in the terrorism index for 2013 is correlated with 5.7 percent decline in the odds of reporting higher levels of health satisfaction. Additionally, a change in the average terrorism index over the three-year period is associated with a decrease of 9.99 percent in the odds of reporting better health levels. These findings imply that terrorism has long-lasting impacts on well-being of the society. Consistent with earlier findings (Bleich et al. 2003; Galea et al. 2002; Richman et al. 2008; Schuster et al. 2001), the empirical results of this study show exposure to terrorism to be correlated with health status at the individual level. The findings of this study thus complement the prior research by demonstrating that individual satisfaction with health status is negatively correlated with exposure to terrorism.

Considering demographics, all the multi-level models provide similar results for individual-level determinants of health satisfaction. Older people are more likely to report lower health status, with a unit increase in age associated with a 3 percent decline in the odds of reporting a higher level of health satisfaction. Consistent with the literature (Gächter et al. 2012; Soytas and Köse 2014), this study finds that Turkish females are 39 percent less likely than Turkish males to report higher levels of health satisfaction. The gender gap in self-rated health status thus extends to health satisfaction at individual level.

In line with the literature (DeWalt et al. 2004), empirical findings of the study show that an increase in years of education is positively correlated with the odds of reporting higher levels of health satisfaction: possessing a higher level of education is associated with a 5.8 percent increase in likelihood of reporting a higher level of health satisfaction. Moreover, consistent with previous studies (Ross and Mirowsky 1995; Case and Deaton 2005), employed respondents are 6.7 percent more likely to report higher levels of health status. The empirical findings indicate that married individuals report lower levels of health satisfaction than "never married" individuals. However, married respondents report higher levels of health satisfaction than "divorced" and "widowed" individuals.

Individual-level control variables such as becoming a parent and using health services also demonstrate significant associations with health satisfaction. Becoming a parent within the past year correlates positively with health satisfaction whereas using health services in the survey year shows a significant negative relationship with health level. One should note that both relationships are highly likely to suffer from reverse causality. Individuals with less health satisfaction are more likely to employ health services, while healthy individuals are more likely to become parents. Respondents who became parents within the past year, are 52 percent more likely to report higher levels of health status. Respondents who used health services within the survey year are 50.1 percent less likely to have higher levels of health satisfaction. These findings are consistent with the literature (Umberson 1987; Miilunpalo et al. 1997). Moreover, individual interest in health issues is shown to be significantly and positively related to health satisfaction. The greater the individual's interest in health issues, the more likely they are to report higher levels of health satisfaction. This finding supports the findings of prior research that health literacy of individuals is significantly related to health outcomes (Berkman et al. 2011; DeWalt and Hink 2009).

Household conditions are significantly associated with health outcomes. As in the literature (Smith 1999; Ettner 1996), the empirical findings of this study indicate that the odds of reporting higher levels of health status increase by 10.9 percent if there is an increase in income level of the household. Ultimately, higher housing floor area per person correlates positively with the odds of reporting better levels of health satisfaction. Thus, the findings of the current paper support the notion that better housing conditions improve health outcomes.

Province-level conditions are not significantly associated with health satisfaction of individuals, with the exception of terrorism, GDP per capita, and education services. In light of previous literature (Ross and Wu 1995; Cutler and Lleras-Muney 2006; Borgonovi and Pokropek 2016), which provides evidence of a positive relationship between health outcomes and education, the present empirical analysis includes a proxy variable to control for the variation in education at province level. A high schooling ratio is significantly related to health satisfaction of individuals. If an individual lives in a province with a higher schooling rate, they are 0.4 percent more likely to report higher categories of health satisfaction on average. Results for the relationship between health satisfaction and province-level GDP per capita are consistent with the literature (Subramanian et al. 2002) and indicate that lower GDP per capita is associated with lower health satisfaction for 2011. Prior research indicates that higher government health expenditures and services are significantly related to better health outcomes (Bokhari et al. 2007; Bradley et al. 2011). The present empirical results, however, indicate that province-level indicators for health services and government expenditures such as hospital beds per thousand capita and availability of government financed family doctors show no significant relationship with health satisfaction of individuals.

Sensitivity of sub-groups to terrorism is analyzed using a regression model that includes interactions of average terrorism index with individual-level variables. Estimation results of the corresponding model are presented in the Appendix, Table A4. Empirical analysis indicates that older people are more sensitive to terrorism. The level of health satisfaction declines with interaction of age and terrorism index. The findings indicate that the relationship between terrorism and health satisfaction is different for females and males. As exposure to terrorism rises, females tend to report lower health satisfaction compared to males. Health satisfaction of widowed individuals is influenced by exposure to terrorism. As the level of terrorism rises, widowed individuals are more likely to report lower levels of health satisfaction compared to married individuals. Lastly, results show that the poor are more sensitive to terrorism than the rich.

Finally, impacts of continuous exposure to terrorism on health satisfaction of individuals are estimated by additional empirical models. Initially, a sub-sample of provinces which experienced terrorism incidents for three consecutive years (2011, 2012 and 2013) is considered. Empirical findings, reported in the Appendix, Table A5, indicate that only the terrorism index of 2013 is significantly and negatively correlated with health satisfaction of individuals. Health satisfaction of individuals who repeatedly experience terrorism respond only to the latest incidents. Moreover, sub-samples of individuals who experienced terrorism incidents only in one year are considered for analysis of rare events on health satisfaction. The estimation results (see Appendix Table A6) show that exposure to terrorism only in 2011 is not significantly correlated with health satisfaction in 2013. Additionally, individuals who experienced terrorism incidents only in 2012 report lower health satisfaction in 2013 than those who did not experience terrorism in 2012. These findings are in line with earlier results (Yechiam et al. 2005) and imply that behavioral responses to rare terrorism incidents are different than responses to repeated incidents.

Combined with findings of Yildirim et al. (2016), this study implies that terrorism has significantly negative correlations with both life satisfaction and health satisfaction. It thus confirms the findings of earlier research by providing additional evidence for a significantly negative relationship between terrorism and measures of welfare at the individual level (Frey et al. 2007; Vorsina et al. 2017; Romanov et al. 2012). Overall, the findings of this study expand knowledge on the indirect costs of terrorism for society both in the shorter and the longer term. Researchers and policymakers should therefore not only account for the indirect costs of terrorism but also focus on its consequences over time.

9. Conclusion

Terrorism has devastating effects on tangible and intangible values. These human-caused incidents spread harmfully and their impacts are not limited to victims and survivors who experience the event directly. Given that terror attacks are unforeseen events, individuals are unconsciously exposed to anxiety, fear and anger in their daily lives. Terrorism may have substantial effects on different dimensions of life. Prior empirical research agrees that terrorism has a variety of inevitable costs, including economic, political and psychological impacts. Turkey, as a country targeted by terror for many years, has been enduring the costs of terrorism.

This paper explores health effects of terrorism in Turkey by combining micro-data at individual level and macro-data at province level. The empirical strategy focuses on quantifying the relationship between health satisfaction of individuals and a province-level terrorism index. Although macro-level effects of terrorism are well-documented, individual-level analysis of health outcomes is limited. This article adds to the literature on measuring the costs of terrorism by providing evidence on intangible costs and complements studies seeking to establish a link between terrorism and life satisfaction as a measure of welfare.

Due to the hierarchical structure of the data and the ordered nature of the self-reported health satisfaction measure, multi-level ordered choice models were employed for empirical estimation. The empirical results of a multi-level ordered logistic model show that terrorism is negatively associated with health satisfaction. Furthermore, consistent with the literature, demographics and household-level variables demonstrate significant correlations with health status. Women and older people are less likely to be satisfied with their health status. Employed respondents are more likely to be satisfied, and marital status significantly correlates with health satisfaction. Education level, being a new parent, and level of interest in health issues are positively associated with health satisfaction. Individuals who used health services within the survey year are more likely to have lower satisfaction levels. Household income and housing floor area per person exhibit positive correlations with individual health satisfaction. Moreover, province-level indicators for education and GDP per capita are significantly and positively associated with individual-level health satisfaction. However, number of hospital beds capita and number of family doctors per capita are not related to health satisfaction.

Overall, this study demonstrates that terrorism has indirect effects on economic well-being through the link between health and economic outcomes of individuals. Since exposure to terrorism has repercussions on health outcomes, these effects would be reflected in human capital formation through intergenerational transfers and educational outcomes of current and future generations. This research suggests that in addition to its economic costs, terrorism has health-related costs for society. Thus, policy-makers should pay greater attention to organizing individual-based policies to tackle adverse effects of terrorism both in the short run and in the long run.

Finally, this study is not without limitations and its findings should be interpreted accordingly. First, health outcomes of individuals are measured by selfreported health satisfaction levels rather than objective measures of health. By the nature of survey data, this study may be subject to reporting biases of individuals. Moreover, the survey data used in this study is not able to account for lives lost, and thus underestimates the health effects. Based on a cross-sectional data set, this study is not able to account for changes in health satisfaction over time. Findings of the current study display correlations among variables of interest. Exploration of causal mechanisms for associations of health satisfaction with terrorism and other variables is beyond the scope of this study. Further research is needed for identification of causality frameworks. Future studies may also focus on panel data and analyze the time dimension of changes in health satisfaction with respect to terrorism.

Footnote

1 Ocal and Yildirim (2010) provide an excellent review of the history of terrorism in Turkey.

References

References

Abadie, Alberto, and Javier Gardeazabal. 2003. The Economic Cost of Conflict: A Case Study of the Basque Country. American Economic Review 93:113-32.

Almond, Douglas, Kenneth Y. Chay, and David S. Lee. 2005. The Cost of Low Birth Weight. QuaarterlyJournal of Economics 120 (3): 1031-83.

Berkman, Nancy D., Stacey L. Sheridan, Katrina E. Donahue, David J. Halpern, and Karen Crottay. 2011. Low Health Literacy and Health Outcomes: An Updated Systematic Review. Annals of Internal Medicine 155 (2): 97-107.

Berrebi, Claude, and Jordan Ostwald. 2014. Terrorism and Fertility: Evidence for a Causal Influence of Terrorism on Fertility. Oxford Economic Papers 67 (1): 63-82.

Bilgel, Firat, and Burhan Can Karahasan. 2017. Thirty Years of Conflict and Economic Growth in Turkey: A Synthetic Control Approach. Defence and Peace Economics 2694 (112): 1-23.

Black, Sandra E., Paul J. Devereux, and Kjell G. Salvanes. 2007. From the Cradle to the Market? The Effect of Birth Weight on Adult Outcomes. QuaarterlyJournal of Economics 122 (1): 409-39.

Blanc, Ann. K. 2004. The Role of Conflict in the Rapid Fertility Decline in Eritrea and Prospects for the Future. Studies in Family Planning 35:236-45.

Bleich, Avraham, Marc Gelkopf, and Zahava Solomon. 2003. Exposure to Terrorism, Stress-Related Mental Health Symptoms, and Coping Behaviors among a Nationally Representative Sample in Israel. Journal of the American Medical Association 290 (5): 612-20.

Blomberg, S. Brock, Gregory D. Hess, and Athanasios Orphanides. 2004. The Macroeconomic Consequences of Terrorism. Journal of Monetary Economics 51 (5): 1007-32.

Bokhari, Farasat A.S., Yunwei Gai, and Pablo Gottaret. 2007. Government Health Expenditures and Health Outcomes. Health Economics 16 (3): 257-73.

Borgonovi, Francesca, and Artur Pokropek. 2016. Education and Self-Reported Health: Evidence from 23 Countries on the Role of Years of Schooling, Cognitive Skills and Social Capital. Plos One 11 (2): e0149716.

Brackbill, Robert M., James L. Hadler, Laura DiGrande, Christine C. Ekenga, Mark R. Farfel, Stephen Friedman, Sharon E. Perlman, Steven D. Stellman, Deborah J. Walker, David Wu, Shengchao Yu, and Lorna E. Thorpe. 2009. Asthma and Posttaraumatic Stress Symptoms 5 to 6 Years Following Exposure to the World Trade Center Terrorist Attaack. Journal of the American Medical Association 302 (5): 502-16.

Bradley, Elizabeth H., Benjamin R. Elkins, Jeph Herrin, and Brian Elbel. 2011. Health and Social Services Expenditures: Associations with Health Outcomes. BMJ Quaality and Safety 20:826-31.

Bruck, Tilman, Michele Di Maio, and Sami Miaari. 2014. Learning The Hard Way: The Effect of Violent Conflict on Student Academic Achievement, HICN Working Paper 185. N. p.: Households in Conflict Network.

Buvinic, Mayra, Monica Das Gupta, and Olga N. Shemyakina. 2014. Armed Conflict, Gender, and Schooling. World Bank Economic Review 28 (2): 331-19.

Caner, Asena. 2014. Happiness, Comparison Effects, and Expectations in Turkey. Journal of Happiness Studies 16 (5): 1323-45.

Case, Anne, and Angus S. Deaton. 2005. Broken Down by Work and Sex: How Our Health Declines. In Analyses in the Economics of Aging, ed. David A. Wise, 185 - 212. Chicago: University of Chicago Press.

Camacho, Adriana. 2008. Stress and Birth Weight: Evidence from Terrorist Attaacks. American Economic Review 98 (2): 511-515.

Couzin, Jennifer. 2002. Queirks of Fetal Environment Felt Decades Later. Science 296 (5576): 2167-69.

Cutler, David M., and Adriana Lleras-Muney. 2006. Education and Health: Evaluating Theories and Evidence, Working Paper 12352. Cambridge, MA: National Bureau of Economic Research.

Deeks, Amanda, Catherine Lombard, Janet Michelmore, and Helena Teede. 2009. The Effects of Gender and Age on Health Related Behaviors. BMC Public Health 9 (1): 213.

Denton, Margaret, and Vivienne Walters. 1999. Gender Differences in Structural and Behavioral Determinants of Health: An Analysis of the Social Production of Health. Social Science and Medicine 48 (9): 1221-35.

DeWalt, Darren A., Nancy D. Berkman, Stacey Sheridan, Kathleen N. Lohr, and Michael P. Pignone. 2004. Literacy and Health Outcomes: A Systematic Review of the Literature. Journal of General Internal Medicine 19 (12): 1228-39.

DeWalt, Darren A., and Ashley Hink. 2009. Health Literacy and Child Health Outcomes: A Systematic Review of the Literature. Pediatrics 124 (Supplement 3): S265-S274.

Drakos, Konstantinos, and Ali M. Kutan. 2003. Regional Effects of Terrorism on Tourism in Three Mediterranean Countries. Journal of Conflict Resolution 47 (5): 621-41.

Dumludag, Devrim. 2013. Life Satisfaction and Income Comparison Effects in Turkey. Social Indicators Research 114 (3): 1199-1210.

Eckstein, Zvi, and Daniel Tsiddon. 2004. Macroeconomic Consequences of Terror: Theory and the Case of Israel. Journal of Monetary Economics 51 (5): 971-1002.

Eisinger, Peter. 2004. The American City in the Age of Terror: A Preliminary Assessment of the Effects of September 11. Urban Affairs Review 40 (1): 115-30.

Ekici, Tufan, and Selda Koydemir. 2014. Social Capital, Government and Democracy Satisfaction, and Happiness in Turkey: A Comparison of Surveys in 1999 and 2008. Social Indicators Research. 118 (3): 1031-53.

Enders, Walter, and Todd Sandler. 1993. The Effectiveness of Antiterrorism Policies: A Vector AutoregressionIntervention Analysis. American Political Science Review 87 (4): 829-44.

Ettner, Susan L. 1996. New Evidence on the Relationship Between Income and Health. Journal of Health Economics 15 (1): 67-85.

Feridun, Mete. 2016. Impact of Education and Poverty on Terrorism in Turkey: An Empirical Investigation. Applied Research in Quaality of Life 11 (1): 41-48.

Fielding, David. 2003. Counting the Cost of the Intifada: Consumption, Saving and Political Instability in Israel. Public Choice 116 (3-4): 297-312.

Frey, Bruno S., Simon Luechinger, and Alois Stutzer. 2007. Calculating Tragedy: Assessing the Costs of Terrorism. Journal of Economic Surveys 21 (1): 1-24.

Gächter, Martin, Peter Schwazer, and Engelbert Theurl. 2012. Stronger Sex but Earlier Death: A Multi-Level Socioeconomic Analysis of Gender Differences in Mortality in Austria. DANUBE: Law and Economics Review 1:1-23.

Gaibulloev, Khusrav, and Todd Sandler. 2011. The Adverse Effect of Transnational and Domestic Terrorism on Growth in Africa. Journal of Peace Research 48 (3): 35571.

Galea, Sandro, Jennifer Ahern, Heidi Resnick, Dean Kilpatrick, Michael Bucuvalas, Joel Gold, and David Vlahov. 2002. Psychological Sequelae of the September 11 Terrorist Attacks in New York City. New England Journal of Medicine 346 (13): 982-87.

Gates, Scott, Havard Hegre, Havard Mokleiv Nygard, and Havard Strand. 2012. Development Consequences of Armed Conflict. World Development 40 (9): 1713-22.

Global Administrative Areas Database (GADM). 2018. GADM Data Version 3.6. Retrieved from: https://gadm.org/download_country_v3.html

Global Terrorism Database (GTD). 2016. National Consortium for the Study of Terrorism and Responses to Terrorism (START). Retrieved from https://www.start.umd.edu/gtd

Goldstein, Kevin B. 2005. Unemployment, Inequality and Terrorism: Another Look at the Relationship between Economics and Terrorism. Undergraduate Economic Review 1 (1): Article 6.

Greenbaum, Robert T., Laura Dugan, and Gary LaFree. 2006. The Impact of Terrorism on Italian Employment and Business Activity. Urban Studies 44 (5-6): 1093-1108.

Guha-Sapir, Debarati, and Olivia D'Aoust. 2011. Demographic and Health Consequences of Civil Conflict, World Development Report 2011, Background Paper 62030. Washington, D.C.: World Bank.

Hair, Joseph F., Rolph E. Anderson, Ronald L. Tatham, and William C. Black. 1998. Multivariate Data Analysis. Upper Saddle River, NJ: Prentice Hall.

Hu, Yannan, Frank J. van Lenthe, Gerard J. Borsboom, Caspar W.N. Looman, Matthias Bopp, Bo Burström, Dagmar Dzurova, Ola Ekholm, Jurate Klumbiene, Eero Lahelma, Mall Leinsalu8, Enrique Regidor, Paula Santana, Rianne de Gelder, and Johan P. Mackenbach. 2016. Trends in Socioeconomic Inequalities in SelfAssessed Health in 17 European Countries between 1990 and 2010. Journal of Epidemiology and Community Health 70:644-52.

Idler, Ellen L., and Yael Benyamini. 1997. Self-Rated Health and Mortality: A Review of Twenty-Seven Community Studies. Journal of Health and Social Behavior 38 (1): 2137.

Institute for Economics and Peace (IEP). 2017. Global Terrorism Index 2017. N. p.: Institute for Economics and Peace. httap://visionofhumanity.org/app/uploads/2017/11/ Global-Terrorism-Index-2017.pdf

Kibris, Arzu. 2015. The Conflict Trap Revisited: Civil Conflict and Educational Achievement. Journal of Conflict Resolution 59 (4): 645-70.

Kiecolt-Glaser, Janice K., and Tamara L. Newton. 2001. Marriage and Health: His and Hers. Psychological Bulletin 127 (4): 472-503.

Krieger, James, and Donna L. Higgins. 2002. Housing and Health: Time Again for Public Health Action. American Journal of Public Health 92 (5): 758-68.

Lai, Brian, and Clayton Thyne. 2007. The Effect of Civil War on Education, 1980-97. Journal of Peace Research 44 (3): 277-92.

Lundborg, Petter, Hilda Ralsmark, and Dan-Olof Rooth. 2014. The More the Sicker? Health, Family Size, and Birth Order, Lund University Working Paper.

Marmot, Michael, Carol D. Ryff, Larry L. Bumpass, Martin Shipley, and Nadine F. Marks. 1997. Social Inequalities in Health: Next Queestions and Converging Evidence. Social Science and Medicine 44 (6): 901-10.

Miilunpalo, Seppo, Ilkka Vuori, Pekka Oja, Matti Pasanen, and Helka Urponen. 1997. Self-rated Health Status as a Health Measure: The Predictive Value of Self-Reported Health Status on the Use of Physician Services and on Mortality in the Working-age Population. Journal of Clinical Epidemiology 50 (5): 517-28.

Mirza, Daniel, and Thierry Verdier. 2008. International Trade, Security and Transnational Terrorism: Theory and Empirics. Journal of Comparative Economics 36 (2): 17994.

Mitchell, Alex J., Julian Benito-Leon, Jose-Manuel Morales Gonzalez, and Jesus Rivera-Navarro. 2005. Queality of Life and Its Assessment in Multiple Sclerosis: Integrating Physical and Psychological Components of Wellbeing. Lancet Neurology 4 (9): 556-66.

North, Carol S., Sara Jo Nixon, Sheryll Shariat, Sue Mallonee, J. Curtis McMillen, Edward L. Spitznagel, and Elizabeth M. Smith. 1999. Psychiatric Disorders among Survivors of the Oklahoma City Bombing. Journal of the American Medical Association 282 (8): 755-62.

Ocal, Nadir, and Julide Yildirim. 2010. Regional Effects of Terrorism on Economic Growth in Turkey: A Geographically Weighted Regression Approach. Journal of Peace Research 47 (4): 477-89.

Omay, Tolga, Bahar Araz-Takay, Ayşegül Eruygur, and İlker Kilic. 2013. The Effects of Terrorist Activities on Foreign Direct Investment: Nonlinear Evidence from Turkey. Review of Economics 64 (2): 139-58.

Pfefferbaum, Betty, Carol S. North, Brian W. Flynn, Robert J. Ursano, Gretchen McCoy, Robert DeMartino, Wayne E. Julian, Cedric E. Dumont, Harry C. Holloway, and Ann E. Norwood. 2001. The Emotional Impact of Injury Following an International Terrorist Incident. Public Health Review 29 (2-4): 271-80.

Poirier, Thomas. 2012. The Effects of Armed Conflict on Schooling in Sub-Saharan Africa. International Journal of Educational Development 32 (2): 341-51.

Richman, Judith A., Lea Cloninger, and Kathleen M. Rospenda. 2008. Macrolevel Stressors, Terrorism, and Mental Health Outcomes: Broadening the Stress Paradigm. American Journal of Public Health 98 (2): 32329.

Robles, Theodore F., and Janice K. Kiecolt-Glaser. 2003. The Physiology of Marriage: Pathways to Health. Physiology & Behavior 79 (3): 409-16.

Romanov, Dmitri, Asaf Zussman, and Noam Zussman. 2012. Does Terrorism Demoralize? Evidence from Israel. Economica 79 (313): 183-98.

Rondó, Patricia H. C., Rodrigo Fernando Ferreira, Fabio Nogueira, Maria C. N. Ribeiro, H. Lobert, and Rinaldo Artes. 2003. Maternal Psychological Stress and Distress as Predictors of Low Birth Weight, Prematurity and Intrauterine Growth Retardation. European Journal of Clinical Nutrition 57 (2): 266-72.

Ross, Catherine E., and John Mirowsky. 1995. Does Employment Affect Health? Journal of Health and Social Behavior 36 (3): 230-43.

Ross, Catherine E., and Chia-ling Wu. 1995. The Links Between Education and Health. American Sociological Review 60 (5): 719-45.

Schneider, Friedrich, Tilman Brück, and Daniel Meierrieks. 2011. The Economics of Terrorism and Counter-Terrorism: A Survey (Part I), Economics of Security Working Paper 44. Berlin: Deutsches Institut für Wirtschaftsforschung (DIW).

Schuster, Mark A., Bradley D. Stein, Lisa H. Jaycox, Rebecca L. Collins, Grant N. Marshall, Marc N. Elliott, Annie J. Zhou, David E. Kanouse, Janina L. Morrison, and Sandra H. Berry. 2001. A National Survey of Stress Reactions after the September 11, 2001, Terrorist Attacks. New England Journal of Medicine 345 (20): 1507-12.

Selim, Sibel. 2008. Life Satisfaction and Happiness in Turkey. Social Indicators Research 88 (3): 531-62.

Shany, Adi. 2016. Too Scared for School? The Effects of Terrorism on Israeli Student Achievement, Working Paper. Hebrew University of Jerusalem.

Shemyakina, Olga. 2011. The Effect of Armed Conflict on Accumulation of Schooling: Results from Tajikistan. Journal of Development Economics 95 (2): 186-200.

Shields, Robin, and Julia Paulson. 2015. "Development in Reverse"? A Longitudinal Analysis of Armed Conflict, Fragility and School Enrolment. Comparative Education 51 (2): 212-30.

Sidel, Victor W., and Barry S. Levy. 2003. War, Terrorism, and Public Health. Journal of Law, Medicine and Ethics 31 (4): 516-23.

Singh, Prakarsh, and Olga N. Shemyakina. 2016. GenderDifferential Effects of Terrorism on Education: The Case of the 1981-1993 Punjab Insurgency. Economics of Education Review 54:185-210.

Smith, James P. 1999. Healthy Bodies and Thick Wallets: The Dual Relation between Health and Economic Status. Journal of Economic Perspectives 13 (2): 145-66.

Solari, Claudia D., and Robert D. Mare. 2012. Housing Crowding Effects on Children's Wellbeing. Social Science Research 41 (2): 464-76.

Soytas, Mehmet A., and Tekin Köse. 2014. Gender Differences in Self-Rated Health Status: Cross-Country Evidence from Turkey and the United States, Working Paper. University of Pittsburgh.

StataCorp. 2017. Stata Multilevel Mixed-effects Reference Manual, Release 15. College Station, TX: StataCorp.

Stellman, Jeanne Mager, Rebecca P. Smith, Craig L. Katz, Vansh Sharma, Dennis S. Charney, Robin Herbert, Jacqueline Moline Benjamin J. Luft, Steven Markowitz, Iris Udasin, Denise Harrison, Sherry Baron, Philip J. Landrigan, Stephen M. Levin, and Steven Southwick. 2008. Enduring Mental Health Morbidity and Social Function Impairment in World Trade Center Rescue, Recovery, and Cleanup Workers: The Psychological Dimension of an Environmental Health Disaster. Environmental Health Perspectives 116 (9): 1248-53.

Stevens, James P. 2012. Applied Multivariate Statistics for the Social Sciences, 5th ed. New York: Routledge.

Subramanian, Sankaran Venkata, Paolo Belli, and Ichiro Kawachi. 2002. The Macroeconomic Determinants of Health. Annual Review of Public Health 23 (1): 287-302.

Tabachnick, Barbara G., and Linda S. Fidell. 2007. Using Multivariate Statistics, 5th ed. Boston, MA: Allyn & Bacon and Pearson Education.

Turkish Statistical Institute (TSI). 2014. Life Satisfaction Survey of 2013. http://www.tuik.gov.tr/MicroVeri/YMA_2013/index.html

Turkish Statistical Institute (TSI). 2017. Regional Statistics. https://biruni.tuik.gov.tr/bolgeselistatistik/sorguSayfa.do?target=degisken

Umberson, Debra. 1987. Family Status and Health Behaviors: Social Control as a Dimension of Social Integration. Journal of Health and Social Behavior 28 (3): 306-19.

Vorsina, Margarita, Matthew Manning, Christopher M. Fleming, Christopher L. Ambrey, and Christine Smith. 2017. The Welfare Cost of Terrorism. Terrorism and Political Violence 29 (6): 1066-86.

Wadhwa, Pathik D., Curt A. Sandman, Manuel Porto, Christine Dunkel-Schetter, and Thomas J. Garite. 1993. The Association between Prenatal Stress and Infant Birth Weight and Gestational Age at Birth: A Prospective Investigation. American Journal of Obstetrics and Gynecology 169 (4): 858-65.

Williams, David R., Yan Yu, James S. Jackson, and Norman B. Anderson. 1997. Racial Differences in Physical and Mental Health: Socio-economic Status, Stress and Discrimination. Journal of Health Psychology 2 (3): 33551.

Woldemicael, Gebremariam. 2008. Recent Fertility Decline in Eritrea: Is It a Conflict-Led Transition? Demographic Research 18:27-58.

Yechiam, Eldad, Greg Barron, and Ido Erev. 2005. The Role of Personal Experience in Contributing to Different Patterns of Response to Rare Terrorist Attacks. Journal of Conflict Resolution 49 (3): 430-39.

Yildirim, Julide, Tekin Köse, and Gizem Tanrivere. 2016. Terrorism and Happiness in Turkey. Paper presented at the 20th International Conference on Economics and Security, 15-17 June 2016, Ankara, Turkey.