Korean Journal of Biological Psychiatry

Address for correspondence: Yong-Min Ahn, MD, Department of Neuropsychiatry, Seoul National University Hospital, 101 Daehak-ro, Jongro-gu, Seoul 03080, Korea
Tel: +82-2-2072-2457, Fax: +82-2-744-7241, E-mail: aym@snu.ac.kr

Introduction

ㅔHow a society perceives suicide influences the suicidal behavior of individuals in that society. Previous studies show that the more permissive toward sucide a society is, the more frequent suicide is in both individual and country levels.1)2)3)4)5)6) If this correlation between permissiveness, in terms of admissibility or acceptability, and suicide rate does in fact exist, then South Korea, which has one of the highest suicide rates in the world, must have a highly permissive attitude toward suicide.
ㅔOur study aims to explore and compare the national attitudes towards suicide with a representative sample of the general population. For this purpose, we carried out a survey in three countries: South Korea, Japan, and the United States. South Korea and Japan share many cultural similarities : strong tradition of Confucianism ; highly homogeneous ethnic demographics ; long history of nationalism ; importance of collectivism, etc. The United States, on the other hand, is a cultural mirror image of South Korea and Japan with its individualist tradition and pragmatism deeply rooted in the basis of American society.
ㅔSuicide is such a complex phenomenon that it needs to be approached from multiple dimensions. Various disciplines cast suicide in their own lights : for psychology suicide is mainly an individual behavior ; for sociology or anthropology it is a socio-cultural problem ; for psychiatrics it is a mental disorder that requires proper treatment. Of course, we need to have multi-disciplinary approaches to truly understand the depth and complexity of suicidal behavior. Psychiatric problems that are considered suicide related are depression, substance abuse, and excessive stress, as well as physical illness and interpersonal relationship, occupational, and economic concerns.7)8) In regard to socio-cultural factors, suicide has been considered as a product of social conditions ever since Emile Durkheim proposed a sociological theory of suicide in Le Suicide in 1897. An excess of two main sociological forces, "social integration" and "social regulation," may result in increased suicide rates. Since then, many sociological studies emerged to explain the difference between countries.9)10)
ㅔAmong the various aspects of explanation for suicide between countries, we focus on the effect of the general perception of suicide in a society. For example, South Korea has been under the strong influence of Confucianism since the 14th century, which emphasizes the importance of family value. Such a general culture may influence permissiveness for suicide and affect suicide behaviors. Attitude toward suicide was used as one of methods in order to approach the general thought for suicide.
ㅔPrevious studies showed that attitude toward suicide is affected by a number of factors such as sex, age, religion, prior contact with suicidal individuals, and one's own history of suicidality. But the outcomes of those studies are not consistent.2)11)12)13)14)15)16)17) In terms of the association between permissive attitude toward suicide and suicidality, some studies have shown a significant correlation.1)2)3)5)18) Despite these results, some other studies have argued that countries with more permissive attitudes toward suicide are more likely to have lower suicide rates. The common limitation of these previous research efforts is that they do not have samples that can represent the general population of a society. Usually these studies only have specific and limited samples such as health care workers, medical school students, nurses, or university students. As a result, samples from almost all previous research were not representative and were limited in large scale comparisons, such as difference between countries. The discrepancy among previous literatures may be caused by this problem of partial samples.

Methods

Procedures
ㅔThe data for this study was gathered by a multi-stage stratified sampling method. The target population was 20- to 59-year-old adults from South Korea, Japan, and the United States. The panel data was divided according to gender, age, and residential area of individuals, and an email with a hyperlink to our web survey was sent to the randomly selected participants in each stratums. The surveys were conducted by three survey companies – Macromill Embrain in South Korea, Global Mappping International in the United States, and Micromill in Japan. The web survey questionnaire contained a self-administered questionnaire and information about the purpose of the research. The total number of the respondents was 2247 (790 from South Korea, 730 from the United States, and 727 from Japan). The web-questionnaire was designed to not allow a respondent to skip an item, so that there were no missing values in the dataset. The only exception was an item that measured the level of suicidal ideation of a respondent. But only a small number of respondents chose not to answer this question (10 missing values from South Korea, 15 from Japan, and 9 from the United States).
ㅔThis survey was monitored and approved by the Institutional Review Board of Seoul National University. Any questions that could reveal the identity of individuals were dropped from the questionnaire. Through a filter question that appeared before the main body of the questionnaire, individuals who had sought professional help for their mental problems or attempted suicide were excluded from the survey.

Measurement
ㅔTo measure the perceptual differences on suicide in different cultures, this study adopted the Attitudes Toward Suicide questionnaire (ATTS). The ATTS was originally developed by Swedish researchers.19) Compared to other questionnaires that are being widely used such as Suicide Opinion Questionnaire (SOQ) or Suicide Attitude Questionnaire (SUIATT), the ATTS has the advantage in its relative simplicity and conciseness, so it is better equipped for large-scale public survey.20)21) Out of the forty items from the 2008 version of ATTS, thirty-seven items were employed in this study. The ATTS items were measured with a five-point Likert Scale, which ranged from Strongly Agree to Strongly Disagree. The ATTS had been translated into Japanese by Japanese researchers,20) and this study utilized this version of the Japanese translation. The Korean version of ATTS was translated by Park and Kim.22)

Data analysis
ㅔExploratory factor analysis is performed with the data from the three countries. Parallel analysis method is employed to determine the number of factors in each analysis, which is considered to be the most accurate way of estimating factors.23)24)25) Parallel analysis reveals that the ATTS items converge into five factors in all three countries. After deciding the number of factors, exploratory factor analysis with maximum likelihood extraction method and promax rotation is performed. The items for which the factor loading value are smaller than 0.35 were discarded. Also the items with high factor loading value (> 0.35) for multiple factors are excluded from the analysis. Cronbach's α is calculated for each factor.
ㅔThe exploratory factor analysis of ATTS in each country shows that the permissiveness factor in all three countries is comprised of very similar items (Table 1, 2, 3). Also, among all the factors the permissiveness explains the most variance (11.0% in South Korea, 11.1% in Japan, 16.8% in the United States). To compare permissiveness in the three countries, a summation scale is constructed by adding up five items that commonly are part of the permissiveness factor (e.g., suicide is an acceptable means to terminate an incurable disease ; would consider suicide if I suffer from a severe, incurable disease).
ㅔOne-way ANOVA is conducted is to test the group difference for permissiveness (Table 5). The demographic variables that are employed in ANOVA are gender, age, level of education, employment status, and marital status. In addition, the national characteristics of the items in the permissiveness factor that were not shared by all three countries were examined using one-way ANOVA and Tukey's post hoc test.
ㅔFor significant test results, 5% level of p values is used for two-tailed tests. All the statistical analysis of this research are conducted using STATA version 13.1 (StataCorp., College Station, TX, USA) and SPSS for Windows version 21.0 for PC (SPSS Inc., Chicago, IL, USA).

Results

ㅔTable 4 summarizes the characteristics of sample demographics for the three countries. Table 1, 2, 3 present the result of factor analysis of ATTS in each country. In the case of South Korea, five factors are found which consisted of nineteen items. Likewise, the American ATTS model has five factors with nineteen items, but each factor is constructed with different items in two countries. The Japanese factor model also has five factors, but those factors contain twenty-five items.
ㅔThe permissiveness factor, which is the major concern of this article, is found in all three countries. The Korean permissiveness factor has five items, and these items are also commonly found in the Japanese and American models. Based on these factor analyses, a summation scale for permissiveness of suicide is constructed with the five common items that all three country models shared. Cronbach's alpha for this permissiveness scale is 0.75 in South Korea, 0.79 in Japan 0.79, and 0.84 in the United States.
ㅔTable 5 presents the results of the one-way ANOVA of permissiveness. The national mean value of permissiveness was not significantly different between countries, but permissiveness according to various demographic variables was different in each country. In South Korea, none of the demographic variables has significant association with permissiveness. Basically, for all the Korean groups in each demographic, variables have more or less the same level of permissiveness of suicide. But in the case of Japan, statistically significant group differences are found in gender, education, and marital status : Japanese male respondents are more permissive than female respondents; the level of education has a non-linear relationship with permissiveness-while college graduates have more permissive attitudes than those with lower education levels, the people with graduate level education have the most intolerant attitudes toward suicide. If we devide into two groups, less than high school education and more than college graduation, people with higher education levels were more significantly permissive than people with relatively lower education levels in Korea and the United States [mean 13.94 (SD 0.32) in less than high school education and 14.69 (SD 0.15) in college graduate or higher education in Korea ; mean 14.00 (SD 0.29) in less than high school education and 15.06 (SD 0.25) in college graduate or higher education in the United States]. Finally, the experience of marriage significantly reduces the tendency of permissive attitude toward suicide. But employment status and age do not have significant impact on permissiveness. The ANOVA of American data reveals that gender and education are statistically significant : American males are more permissive than females ; and American respondents with college level education have the most permissive attitude than other education level groups, which is the same as the Japanese model.
ㅔThe last row of Table 5 reports the national mean of permissiveness for each country. The United States has the most permissive attitude toward suicide (14.59), while South Korea and Japan have relatively smaller values of permissiveness (14.55 and 14.52, respectively). T-test showed that the differences of permissiveness among the three countries are not significant. Overall, we can conclude that the three countries of this research, even with their cultural differences, have more or less the same level of permissive attitude toward suicide.
ㅔIn the permissiveness factor, one item, "May be situations where suicide is the only resolution," was found only in Japan, and the other item, "People do have the right for suicide," appeared only in Japan and the United States. For the former, significantly higher agreement was observed in Japan than in both South Korea and the United States (p < 0.001, one-way ANOVA, Tukey's post hoc test), between which there was no significant difference, and, for the latter, significantly higher agreement was shown in the United States than in Japan (p < 0.001, one-way ANOVA, Tukey's post hoc test) and in Japan than in South Korea (p < 0.001, one-way ANOVA, Tukey's post hoc test).

Discussion

ㅔWe aimed two goals : the exploration of the difference of structure of attitude in three different countries and the relationship between permissive attitude toward suicide and national suicide rate. First, the results of our study show that there exist differences in the composition of factors for the attitudes toward suicide in the three countries. This is the first report to compare the attitude toward suicide between countries with a large number of representative sample and a full 37 items of the ATTS Scale. Next, with regard to permissive attitude toward suicide, the mean value of permissiveness was not significantly different between countries, but permissiveness according to education level, gender, and marital status was different in each country. This result seems to be a little different than we expected, in part, regarding higher permissiveness toward suicide in countries with high suicide rates. However, it is affecting the culture of each country by the difference between the configuration of each demographic factors of attitude rather than simply the overall attitude. This result makes it possible to understand a little more in-depth about the concept of suicide.
ㅔFirst, we focus on the structures of items, especially the number of factors and consistent concept of suicide. Items of ATTS were divided with five factors in all three countries ; however, the composition of each item was different in Korea, Japan, and the United States. It is difficult to compare the results of previous factor analysis studies directly because they used different versions of ATTS. For instance, there was a factor analysis study with a 20 item Korean version of ATTS26) and another analysis performed in Sweden, Norway, and Russia with a 40 item of ATTS.20) In a 37 item version study performed in Japan, it was analyzed with six factors and there was little difference of composition in the included items. This is due to the fact that the subjects enrolled in the survey were different : The subjects of the previous Japanese study were specialists that participated in one seminar; on the other hand, our sample is representative of the general population. Therefore, our results may be representative of the structure of ATTS in each country.
ㅔSecond, with regard to permissive attitude, it is understandable on the basis of contradictory results from previous researches. There were several researches about association with more permissive attitudes toward suicide and suicide as an outcome in individual aspects1)2) and international comparison between countries.3)4)5)6) Overall, the more permissive attitude towards suicide was interconnected with higher prevalence of suicide ; however, some results have shown more permissive attitudes in countries with lower suicide rates. Our results were also not correlated with suicide rate and permissive attitude toward suicide, uniformly. According to the Organization for Economic Cooperation and Development (OECD) Health Statistics report of 2016,27) the South Korean suicide rate was 28.7 per 100000 population in 2013, and in Japan in 2013 it was 18.7. But the American suicide rate in 2013 was 13.1, which is distinctively lower than the other two countries and more or less the same with the OECD average suicide rate (12.1 in 2014 or nearest year). This result suggests that the suicide rate in a country could not be understood simply by comparing one aspect and should be considered comprehensively with a number of related factors.
ㅔAdditionaly, we analyzed the permissiveness toward suicide with several demographic variables. Males were more permissive about suicide in Japan and the United States ; however, there was no significant difference in Korea. In previous researches, women were reported to be more likely to disapprove of one's right to suicide14) and tended to believe that rational suicide should be allowed under certain conditions.15) The results about gender difference of Japan and the United States in our study are consistent with previous studies, but only the result from Korea was inconsistent. Likewise, our study further showed that this gender difference of permissive attitude also differ in countries.
ㅔIn regard to age in our study, age did not affect the attitudes about suicide in all three countries. About permissive attitude by aging, there were different previous results. It was reported that older subjects tended to be more empathetic toward suicidal patients11) ; on the contrary, older subjects were reported to be more condemnatory toward the right to suicide12) or to be more comprehensible than younger subjects in another researches.13) Our results were not inconsistent with previous researches.
ㅔAdditionally, permissive attitude was related with education level and marital status. In Korea and the United States, people with higher education levels had more permissive attitudes, but there was no significant difference in Japan. Influence of marital status on permissive attitude was different for each country : Those who never married in the United States were more permissive about suicide than married people, and people who were previously married had lower permissive attitudes. On the contrary, married people in Japan were less permissive than people without marriage experience. This complex demographic difference of attitude toward suicide in each country may come from the cultural differences between countries, so it needs to combine other various cultural factors in order to explain attitude toward suicide. On examples of other cultural factors, it has revealed the association between attitude toward suicide and religion, such as Christianity and Judaism.16) There were reports about another factors relating individual experience or psychiatric problems ; subjects with a history of suicide crisis were more accepting of the idea of suicide than subjects without such history.2) Likewise, it is possible to explain about attitude after considering a number of factors, social, cultural, and individual.
ㅔThe results from the study can also be understood by the notion of the suicidal process, which begins with suicidal ideation, proceeds with suicidal plan, and ends with completed suicide.28) Permissive attitude can be considered to belong to the stage of suicidal ideation whereas the suicide rate marks the terminal point of the process ; therefore, permissive attitude and suicide rate do not necessarily have to correlate. The result that the national mean of permissiveness reveals no significant difference may imply that, in South Korea, suicidal ideation including permissive attitude toward suicide can evolve through suicidal plan or attempt without less difficulty. Moreover other factors such as predictors and protectors of suicide might intervene in and affect the suicidal process resulting in the difference in the suicide rate. To clarify such element, we further investigated two items found in the permissiveness factor of only one to two countries-one item, "May be situations where suicide is the only resolution," found only in Japan, was more strongly agreed in Japan than in both South Korea and the United States. This item asks about "the resolution." It may show that the Japanese are accustomed to pondering about an adverse situation and devising a means to overcome it with, and thus they are likely to find an alternative resolution rather than suicide. The other item, "People do have the right for suicide," found in Japan and the United States, was more stongly agreed in the United States than in Japan and in Japan than in South Korea. This item inquires about "the right." It may imply that, whereas, in Japan and the United States, people do have the right to make a decision, in South Korea, they have relative difficulty in standing on one's right possibly due to its interdependent culture29) and thus are likely to use suicide as a remedy in despair.
ㅔIn addition, individual factors, such as psychiatric problem and experience contacting suicidal people, should also be considered. Our follow up study plan is to build an integrated model combined with those elements based on this study. Our study has several important implications and strengths. First, we selected subjects that is representative of the general population. To overcome the limitations of subjects in previously performed researches, we selected a large number and distributed evenly by age and sex of the general population. Second, our study is an international comparative study between three countries. Suicide problem in Korea can be understood through this comparison study with Japan and the United States, and at the same time, the influence of attitude toward suicide on suicide rate in each country can be also understood.
ㅔOur study also has several limitations. First, because of our web-survey method, the adult population over 60 was excluded from the sampling, due to their lack of internet access and skills. This is an unavoidable shortcoming of web-surveys. However, we distributed sex and age evenly in the three countries and used a large number of subjects in order to minimize these limitations. Moreover, the level of education of subjects was higher than the general population since people accessing the web-based survey may have relatively higher education levels. Lastly, there might have been some limitations in explaining the related attitude toward suicide and suicide rate in different countries because it should be explained through an integrated model including factors influencing general thought or attitude, such as sex, age, income, religion, and relative's or familial experience of suicide. In this study, we performed a cross-sectional analysis and plan further research for an integrated model.
ㅔIn conclusion, this study explored the comparison of the national attitudes toward suicide with a representative sample of the general population in three countries, Korea, Japan and the United States. The structure of factors and composed items in ATTS were different in each country. Overall mean score of permissiveness was not correlated with the suicide rate of each of the three countries. However, permissive attitude by gender, education level, and marital status was different in each country. Our findings suggest that attitude toward suicide is able to influence the suicide phenomenon in each country ; however, it correlates many different aspects, cultural and social factors, and need to explore a combined integrated model in order to understand the suicide phenomenon. Our results which can represent the three countries may be the basis for that.