Analysis of Factors Affecting Individuals' Sources of Happiness with Multinomial Logistic Model

The happiness levels of individuals and their sources of happiness have been wondered a lot and researched from past to present. The aim of this study is to examine the factors that affect individuals' sources of happiness. The data set of the study was obtained from the Life Satisfaction Survey of the Turkish Statistical Institute. 9212 individuals were included in the study. In the study, chi-square independence tests were conducted to examine the relationship between the source of happiness and the independent variables included in the model, and multinomial logistic regression analysis was applied to determine the factors that may have an effect on the sources of happiness of individuals. As a result of the study, it has been determined that the factors of the individual's age, gender, marital status, educational status, satisfaction with income level, welfare level, life satisfaction, satisfaction with a social life are effective on sources of happiness. At such a time when it is clear that the coronavirus epidemic adversely affects many aspects of our lives, especially our psychology, and will leave a mark on our tomorrows, and the activities of decision-makers and policymakers are shed light through the study in order to increase the happiness of individuals and to ensure that the future will be better.


INTRODUCTION
Happiness is a positive emotion that makes an individual's life meaningful and valuable (Muthuri, Senkubuge & Hongoro, 2020). Happiness, life satisfaction, subjective well-being have always been the focus of attention of researchers, especially social sciences. Long-term happiness is possible when we gain acquisitions for our values or goals (Diener, Sapyta & Suh, 1998;Pollock et al., 2015). Values and goals can have different meanings for each individual, and this situation has made it valuable to examine the factors affecting the sources of happiness of individuals and has been a source of motivation for this study. The aim of the study is to examine the factors that affect success, health, love, money, work, and other resources, which are the sources of happiness of individuals and will touch the spirit of individuals, and even societies, for decision-makers and policymakers, and the aim of this study is to be a guide that will contribute to making them happy.
In the body of literature, the concepts of subjective well-being, happiness, and life satisfaction are intertwined. In his study, Diener (2016) defined subjective well-being as a scientific term used for happiness and life satisfaction. There are many studies examining the effect of subjective well-being on different issues. As a result of the examining that Winkelmann (2005) conducted on the factors affecting the subjective well-being of individuals with the ordinal probit regression model; it has been determined that there is a "u" relationship between age and subjective well-being, unemployment negatively affects subjective well-being, and health is an important determinant of subjective well-being. Similarly, Chen and Short (2008), who investigated the effects of households on the subjective well-being of individuals, determined that subjective well-being of lonely individuals is lower, living with a close family (spouse or children) positively affects subjective well-being, health, education, and financial independence positively affects subjective well-being. Likewise, some studies examined subjective well-being with A literature review was conducted for the independent variables in the study. Afterward, chi-square analyzes were made, and independent variables were included in the model. In the study, sociodemographic, economic, and individual factors that may be effective on individuals' sources of happiness were taken as independent variables. Age (18-27,28-37, 38-47, 48-57, 58-67, 68 and +), gender, an education level (not finished school, primary school graduate, secondary-primary school graduate, high school graduate, college-faculty graduate, 5 or 6-year college postgraduate), marital status (married, single, widowed-divorced) variables are sociodemographic factors. Employment status of the individual (working, not working but still related to his job-not working), satisfaction with monthly income level (satisfied (very satisfied-satisfied), medium, not satisfied (not satisfied-not satisfied at all)), welfare level (low (0,1,2,3,4), medium (5), high (6,7,8,9,10)) variables are economic factors. Individual's level of happiness (happy (very happy-happy), moderate, not happy (unhappy-very unhappy)), those who make happy (self, children-spouse, whole family-niece-granddaughter, otherfriends) life satisfaction (not satisfied (0,1,2,3,4), moderate (5), satisfied (6,7,8,9,10)), satisfaction with health (satisfied (very satisfied-satisfied), moderate, dissatisfied (not satisfied) not satisfied at all)), satisfaction with the education he received (satisfied (very satisfied-satisfied) medium, not satisfied (not satisfied-not satisfied at all), not educated)), satisfaction with his social life (satisfied (very satisfiedsatisfied), moderate, dissatisfied (not satisfied at all)), hope (very hopeful-hopeful, hopeless-very hopeless), past comparison (improved, same, regressed, no idea), future comparison (will improve, same, regressed, no idea) variables are individual factors.

Data Analysis
Microsoft Excel was used to make the data suitable for analysis, SPSS 20 for chi-square independence tests and Stata 14.1 for multinomial logistic regression analysis were used.
The discrete choice models, which are the backbone of empirical analysis for many fields, including economics, psychology, transportation, public policy, are used to estimate the probability of choosing an alternative under the assumption that decision-makers will maximize utility among finite alternatives (Ben-Akiva & Bierlaire, 1999;Garrow, 2016;Newman, Lurkin & Garrow, 2018). The multinomial logistic regression model, which is one of the discrete choice models, is applied when the dependent variable contains three or more categories without being subjected to an order (Koppelman & Wen, 1998).
Since the dependent variable of the study is the sources of happiness of individuals, multinomial logistic regression model, which is one of the discrete choice models, was used in the analysis of the data due to the categorical nature of the dependent variable In the study, firstly, the frequencies and percentages of the individuals participating in the study were calculated according to their sources of happiness. Afterward, chi-square independence tests were conducted to examine the relationship between the source of happiness and the independent variables included in the model, and multinomial logistic regression analysis was applied to determine the factors that may have an effect on the sources of happiness of individuals. According to the probe values of the chi-square independence tests in Table 2, it has been determined that there are statistically significant relationships between individuals' sources of happiness and sociodemographic, economic, and individual indicators.

Model Estimation
In the study, a multinomial logistic regression model was used to determine the factors that affect individuals' sources of happiness. An important assumption of multinomial logistic regression analysis is the assumption of independence of irrelevant alternatives (Vijverberg, 2011). The assumption of independence of irrelevant alternatives means that the relative probabilities of each pair of alternatives are independent of the presence or absence of all other alternatives. Violation of this assumption leads to incorrect estimates (Greene, 2002;Koppelman and Wen, 1998). Small-Hsiao test was used to test this assumption. The results of the independence test of irrelevant alternatives of the multinomial logistic regression model are given in Table 3.

293
With reference to Table 2, it is concluded that the H0 hypothesis cannot be rejected for categories such as success, health, love, work, money, and other categories that are sources of happiness. Thus, the assumption of independence of irrelevant alternatives is provided. Another assumption of the multinomial logistic regression model is that there is no multicollinearity between the independent variables. Because of this, variance inflation factors (vif) were examined. The variance inflation factor being less than 5 indicates that there is no multicollinearity (Alkan & Abar, 2020). All of the variance inflation factors are less than 5 and there are no independent variables with multicollinearity problems in the study.
The estimation results of the multinomial logistic regression model are given in Table 4. In the model, the "health" category of the dependent variable was taken as the reference category. The estimated multinomial logistic regression model was found to be statistically significant (P<0.000).
According to the results of the multinomial logistic model given in Table 4, success for the source of happiness; individual's age (28-37, 38-47, 48-57, 58-67, 68, and more), gender, marital status (never married), educational status (primary, secondary, high school, college-bachelor, postgraduate-5 or 6 year faculty), satisfaction with income level (not satisfied), level of happiness (happy), those who make the individual happy (self, children and spouse, mother and father, other), social life satisfaction (satisfied, not satisfied), hope, future comparison (will develop, no idea) variables were found to be statistically significant.
Love for the source of happiness; individual's age (38-47, 48-57, 58-67, 68 and more), educational status (college-bachelor, postgraduate-5 or 6 year faculty), those who make the individual happy (children and spouse), life satisfaction (satisfied), social life satisfaction (satisfied, not satisfied), past comparison (regressed, no idea) future comparison (will improve) variables were found to be statistically significant.
For job money and other sources of happiness; individual's age (38-47, 48-57, 68 and more), gender, marital status (never married), satisfaction with income level (satisfied, dissatisfied), welfare level (high), happiness level (happy, not happy), happy (self, children and spouse, mother and father, other), life satisfaction (satisfied), future comparison (no idea) variables were found to be statistically significant.
As a result of the model estimation, the independent variables will be interpreted with the help of marginal effects. Table 5 shows the marginal effects and standard errors of factors affecting individuals' sources of happiness. According to the multinomial logistic regression model given in Table 5, for the source of success and happiness: being 68 years old or older reduces the probability of being happy with success by 112.1% compared to the reference group. Female individuals are 46.7% less likely to be happy with success than male individuals. Individuals who have never been married are 84.7% more likely to be happy with success than married individuals. The fact that individuals are postgraduates of 5 or 6 years of faculty increases the probability of being happy with success by 120.5% compared to the reference group. Individuals who are not satisfied with their income level are 24.9% less likely to be happy with success than the reference group. Individuals who are happy with their lives as a whole are 31.1% less likely to be happy with success than the reference group. Individuals who are made happy in their lives by their mothers and fathers are 56.1% more likely to be happy with success than the reference group. Individuals who are not satisfied with their social life are 28.5% more likely to be happy with success than the reference group. Individuals who are hopeful about their own future are 16.5% less likely to be happy with success than the reference group. Individuals who think that their situation will improve in the next 5 years are 32.3% more likely to be happy with success than the reference group.
Health for the source of happiness: Individuals aged 68 and above increase the probability of being happy with health by 23.4% compared to the reference group. Female individuals are 7.3% more likely to be happy with health than male individuals. Individuals who have never been married are 13.3% less likely to be happy with their health than married individuals. Being a postgraduate-5 or 6 year faculty for individuals reduces the probability of being happy with health by 21.1% compared to the reference group. Individuals who are made happy in their lives by their mothers and fathers are 11.6% less likely to be happy with health than the reference group. Individuals who are not satisfied with their social life are 6.7% less likely to be happy with their health than the reference group. Individuals who think that their general condition will improve in the next 5 years are 7.1% less likely to be happy with their health than the reference group.
Love for the source of happiness: Individuals aged 68 and over decrease the probability of being happy with love by 37% compared to the reference group. Female individuals are 13.8% more likely to be happy with love than male individuals. Individuals who have never been married are 20.1% less likely to be happy with love than married individuals. The fact that individuals are postgraduates of 5 or 6 years of faculty increases the probability of being happy with love by 45.5% compared to the reference group. Individuals who are satisfied with their lives are 22.2% more likely to be happy with love than the reference group.
For job, money, and other sources of happiness: Individuals aged 68 and above reduce the probability of being happy with a job, money, and other sources of happiness by 59% compared to the reference group. Female individuals are 53.7% less likely to be happy with a job, money, and other sources of happiness than male individuals. Individuals who have never been married are 53.5% more likely to be happy with a job, money, and other sources of happiness than married individuals. Being a postgraduate of college-bachelor for the individuals decreases the probability of being happy with job, money, and other sources of happiness by 48.7% compared to the reference group. Individuals who are satisfied with their income level are 44.1% more likely to be happy with a job, money, and other sources of happiness than the reference group. Individuals with a high level of well-being are 23.2% more likely to be happy with a job, money, and other sources of happiness than the reference group. Individuals who are happy with their lives as a whole are 23.1% less likely to be happy with a job, money, and other sources of happiness than the reference group. Individuals who are made happy in their lives by their mothers and fathers are 59.2% more likely to be happy with a job, money, and other sources of happiness than the reference group. Individuals who are satisfied with their lives are 26% less likely to be happy with a job, money, and other sources of happiness than the reference group.

DISCUSSION and CONCLUSION
The happiness of individuals brings together happy societies and as a natural result, a peaceful environment occurs. In such a system, it may be possible to achieve more effective outputs with less effort for decision-makers on many vital issues from the economy to health and from education to defense. For this reason, happiness should be considered multidimensional and perhaps more emphasis should be placed on interdisciplinary studies in this regard. The happiness of individuals is affected by many factors, especially demographic and economic factors. In this study, demographic, economic, and individual factors that are effective on individuals' sources of happiness were first investigated with chisquare independence tests and then multinomial logistic regression model, which is the discrete choice model.
As a result of the study, while the happiest individuals with success are young, those who are least happy are over 68 years of age. It is possible to say that the probability of being happy because of success decreases as age increases. Parallel to this result, while the probability of being happy with money and other sources of happiness is higher in young people, it decreases after the middle-ages. In the literature, Selim (2008) determined in his study that compared to individuals in all age groups, individuals in the 18-30 age group believe more that power, job, success, money, and love bring happiness. Success is a more important source of happiness for young individuals who have a dynamic career plan compared to older individuals who have completed their career plans. In addition, this may be related to the fact that younger individuals are less satisfied with their lives compared to older individuals. Likewise, Fernández-Ballesteros, Zamarrón, and Ruiz (2001) and Peterson, Park, and Seligman (2005) determined in their studies that young individuals are less satisfied with their lives compared to the elderly. In addition to this, there are also studies in the literature that found that age affects happiness negatively (Atay, 2012;Chen & Short, 2008;Ekici & Koydemir, 2013). Young people are the most likely to be happy with love, and this probability decreases as age increases. This may be related to the fact that young individuals experience emotions such as love more intensely.
Individuals most likely to be happy with health are 68 years and older, and as the age increases, the probability of being happy with health increases. As age increases, the probability of facing health problems is higher. Thus, older individuals care more about health compared to young individuals, and they know the value of health more. Likewise, Bussière et al. (2021) found that the value given to health differs with age, and that aging increases the effect of health on subjective well-being for individuals and strengthens the relationship between them. In addition to this, when it is looked at from another point of view, health has a very important share in the happiness of individuals whether old or young without making discrimination. There are studies supporting this argument in the literature (Akın & Şentürk, 2012;Bussière et al., 2021;Carandang et al., 2020;Fernández-Ballesteros et al., 2001;Çebi-Karaaslan, Çalmaşur, & Emre-Aysin, 2021;Larson, 1978;Selim, 2008).
Compared to men, women are less likely to be happy with success, job, money, and other sources of happiness, but more likely to be happy with health and love. This may be related to the fact that women are more emotional than men. There are also studies in the literature that found that women are happier than men (Duffrin & Larsen, 2014;Ekici & Koydemir, 2013;Greenstein, 2016;Mookherjee, 1997;Lu, 2000;Wood, Rhodes, & Whelan, 1989). While individuals who have never been married are more likely to be happy with success, job, money, and other sources of happiness than married individuals, they are less likely to be happy with health and love. This may be related to the fact that married individuals' motivation sources and priorities are their spouses or children. Thus, married individuals can care more about health and love. There are many studies in the literature stating that married individuals have a higher tendency to be happy (Akın & Şentürk, 2012;Atay, 2012;Bülbül & Giray, 2011;Ekici & Koydemir, 2013;Fernández-Ballesteros et al., 2001;Kangal, 2013;Çebi-Karaaslan et al., 2021;Lee, Seccombe, & Shehan, 1991;Myers 2000;Shinan-Altman, Levkovich, & Dror, 2020;Veenhoven & Dumludağ, 2015). On the contrary, there are studies that state that unmarried individuals have a higher tendency to be happy (Alexandre, Cordeiro, & Ramos, 2009;Kırcı-Çevik & Korkmaz, 2014;Peterson et al., 2005).
As the education level of the individual increases, the probability of being happy with success increases. In the literature, Selim (2008) found that education has an important role in being happy with a job and money. This can be explained by the fact that educated individuals' achievements are more satisfying, especially when they do work related to their field. In addition, there are also studies that found the positive effects of the level of education on happiness (Atay, 2012;Bülbül & Giray, 2011;Chen & Short, 2008;Eren & Aşıcı, 2017;Kangal, 2013;Shinan-Altman et al., 2020) and the negative effects in the literature (Akın & Şentürk, 2012;Öndes, 2019;Servet, 2017).
An individual who is satisfied with his income level is more likely to be happy with his job, money, and other sources of happiness in his life. While an individual who is dissatisfied with his income level is less likely to be happy with success in life, the probability of being happy is higher with a job, money, and other sources of happiness. This situation may be related to the fact that success brings an improvement in the income level with it and that the individual who is not satisfied with the income level attaches importance to money and therefore to his job in order to improve it. In the literature, it is clear that income is one of the most basic factors affecting the happiness of individuals. There are many studies that found that individuals with financial independence are happier (Chen & Short) and that income has a positive effect on the happiness of individuals (Akın & Şentürk, 2012;Atay, 2012;Blanchflower & Oswald, 2004;Di Tella, MacCulloch, & Oswald, 2003;Diener & Diener, 2009;Ekici & Koydemir, 2013;Fernández-Ballesteros et al., 2001;Kırcı-Çevik & Korkmaz, 2014;Veenhoven & Dumludağ, 2015).
Individuals who are satisfied with their lives are more likely to be happy with love than those who are less satisfied, and less likely to be happy with jobs, money, and other sources of happiness. In parallel with this result, individuals who are happy are less likely to be happy with success, job, money, and other sources of happiness, as well. This may be related to the achievement of spiritual satisfaction of these individuals. Likewise, an individual who is not satisfied with his social life is more likely to be happy with success. This situation may be related to the fact that individuals who are not satisfied with their social life keep their motivation areas in this direction by dedicating themselves to success in order to cover their deficiencies in that area of their lives. Social life is important for the happiness of individuals. In many studies in the literature, it has been determined that individuals who are satisfied with their social life and social relations are happier (Elliot, Cullen, & Calitz, 2018;Fernández-Ballesteros et al., 2001;Çebi-Karaaslan et al., 2021;Myers, 2000;Öndes, 2019;Sirgy & Cornwell, 2001). In addition, Chen & Short (2008) found that individuals living with their families were happier than those living alone.
The factors affecting the happiness and sources of happiness of individuals have had great importance from past to present. Being happy is among the most basic needs of individuals. Likewise, Maslow's hierarchy of needs states that the more an individual's needs are met, the happier the individual will be (as cited in Elliot et al., 2018).
In this study, important deductions were made about the factors affecting the happiness of individuals and their sources of happiness. The outputs obtained are presented in comparison with the literature, and attention is drawn to parallel and opposite situations. It has been hoped that the results of the study will shed light on the activities of policymakers and decision-makers who have an impact on individuals, or societies, experts working in this field.