Gender equality and subjective well-being: Progress for whom?
MSc Business Administration, Leadership and Management track MASTERS THESIS FINAL VERSION
Thesis supervisor: Dr. Richard Ronay Author: Natalie Sourisseau 11652640
STATEMENT OF ORIGINALITY
This document is written by Natalie Sourisseau, who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
TABLE OF CONTENTS
1. INTRODUCTION ... 1
2. THEORETICAL BACKGROUND ... 4
2.2. Gender equality ... 4
2.3. Subjective well-being ... 5
2.4.1. Gender equality only increases SWB for women at ‘the top’ ... 10
2.5. Low gender equality is not characterized by low SWB ... 10
2.5.1. System justification theory: Women rationalize existing inequalities ... 10
2.5.2. Relativity, low expectations, and a match between aspiration and attainment 12 2.5.3. Social role theory and gender bias ... 13
2.6. What happens as gender equality improves? ... 15
2.6.1. Increased expectations and aspirations ... 15
2.6.2. Barriers: Lagging cultural bias and competing responsibilities ... 16
2.6.3. The resulting gap between aspiration and attainment ... 18
3. METHODS ... 20
3.2. Sample and data collection ... 20
3.3. Variables and measures ... 21
3.3.1. Independent variable ... 21 3.3.2. Dependent variable ... 21 3.3.3. Moderator ... 22 3.3.4. Control variables ... 23 4. RESULTS ... 23 4.2. Descriptive statistics ... 23 4.3. Hypothesis testing ... 24 4.3.1. Multi-level model ... 24
4.3.1.1. Step 1: Null hypothesis ... 24
4.3.1.2. Step 2: Level 1 model ... 26
4.3.1.3. Step 3: Level 2 model ... 28
4.3.2. Moderation ... 30
5. DISCUSSION ... 32
5.2. Findings ... 32
5.2.1. Life satisfaction vs. happiness... 32
5.2.2. Main relationship, gender equality and female SWB ... 33
5.2.3. Moderation ... 35
5.3. Practical implications ... 38
5.4. Strengths and limitations ... 40
5.5. Suggestions for further research ... 42
6. CONCLUSION ... 44
REFERENCES ... 45
ABSTRACT
This study examines the main and interactive effects of gender equality and employment success on two measures of subjective well-being (life satisfaction and happiness) among women living in 30 countries throughout Europe. In doing so, it aims to evaluate the idea that greater gender quality does not raise the average SWB of all women, but is likely to do so for a small minority of highly successful women. Data was sourced from the 2011 European Quality of Life Survey, and the World Economic Forum’s 2011 Global Gender Gap Index. A multi-level (two level) regression analysis was used in order to account for country level variance while testing the predicted effects. Against expectations, results indicated a positive relationship between gender equality and both life satisfaction and happiness, providing support for the idea that women’s subjective well-being is higher in countries with higher levels of gender equality. Employment success was found to negatively moderate this relationship, such that gender equality had a less positive relationship with SWB in the presence of employment success. Results suggest that there is much more to be understood about the relationship between gender equality, employment success, and subjective well-being. The lack of consensus in this area raises questions about the aim of gender equality initiatives as well as how we define and measure gender equality and success in society.
1. INTRODUCTION
Gender equality is an issue at the forefront of today’s humanitarian agenda. The principle has been entrenched in European Union (EU) treaties since 1957 (European Union, 2017), and is one of the main aims of the United Nation’s 2030 sustainable development goals (United Nations, 2015). Although there is still a long way to go, numerous governing bodies, institutions and scholars highlight the positive progress that has been made. More women are participating in the workforce, in politics, and in education at all levels (Blau, 1998; Davies & Joshi, 1998; European Union, 2017; McGann & Steil, 2006). This narrative looks promising, but should we be patting ourselves on the back? Are we actually improving women’s well-being through our efforts to promote gender equality? What are the consequences of such progress, and for whom?
These may seem like odd questions, as it is automatic to assume that gender equality is advantageous for women. Surprisingly, this has not been proven with any certainty. As Duflo (2012, p. 1076) points out, research on the subject suggests that women’s empowerment may not be the “magic bullet it is sometimes made out to be”, as there is little support for the idea that gender equality benefits women’s subjective well-being (SWB). While some studies support a positive relationship under certain conditions (Bjørnskov, Dreher, & Fischer, 2007; Jorm & Ryan, 2014; Ruth & Napier, 2014) more have found that higher female status and greater gender equity are negatively correlated with female SWB (Blanchflower & Oswald, 2004b; Duflo, 2012; Meisenberg & Woodley, 2015; Stevenson & Wolfers, 2009; Tesch-Röemer, Motel-Klingebiel, & Tomasik, 2008; Vieira-Lima, 2011).
Although a widely studied phenomenon, there remain gaps in our understanding of the relationship between gender equality and female SWB. There is clearly a need for greater examination of moderating mechanisms, as well as a better understanding of how processes triggered by gender equality impact SWB (Zuckerman, Li, & Diener, 2017). Furthermore,
while much of the research focuses on barriers, Madsen and Scribner (2017) ask that future studies uncover situations in which women’s status is positive or improving. While it is important to recognize existing barriers in order to eliminate or mitigate them, it is also crucial to expose situations in which the relationship between gender equality and female SWB is positive. If gender equality does not improve female SWB for most, what is unique about the situations in which it does? These positive deviants can help uncover circumstances that lead to better outcomes for women. Lastly, although many causal explanations for a negative relationship have been proposed, few of these have been tested empirically (Batz & Tay, 2017).
Meisenberg and Woodley (2015, p. 1551) propose one possibility for their negative findings, suggesting that although “a higher proportion of women in high-status occupations does not raise the average SWB of all women... it is likely to do so for the minority of highly ambitious women competing for these positions.” Zuckerman et al. (2017, p. 335) ask a similar question, “what might change if [women] are successful (or not) in these career choices?” While studies often account for general demographic variables such as age, health, and employment status (employed versus unemployed), the idea that gender equity may only benefit the most successful women calls for the need to examine more unique subsets of the working population. As such, this study examines the moderating factor of employment success on the relationship between gender equality and SWB. Data on gender equality gathered from the World Economic Forum’s Global Gender Gap Index will be analyzed, alongside 2011 survey data from the European Foundation for the Improvement of Living and Working Conditions. See Figure 1.0 for the proposed research model.
As gender equality increases, social norms become less pronounced, and female rights and opportunities increase. Alongside these advantages come increases in women’s aspirations and society’s expectations. However, while women are increasingly told they can ‘have it all’, they continue to face barriers associated with lingering cultural bias. These barriers limit the
degree to which women can actually achieve their heightened aspirations. While successful women ‘at the top’ may benefit from more lenient social roles, goal attainment, and the fulfilment of higher order needs, the rest may be left feeling let down. As such, it is hypothesized that employment success will positively moderate the relationship between gender equality and SWB, where the few women who do attain top-level employment success will enjoy an increase in SWB as gender equality increases. In line with the majority of prior research, the general relationship between gender equality and SWB is hypothesized to be negative.
I begin by giving a brief overview of gender equality and SWB, defining each of the variables and highlighting why they are important to study. Following this, I review existing literature on the relationship between gender equality and SWB. I then introduce the suggestion that gender equality only increases SWB for women at ‘the top’, and use system justification, aspiration-attainment, and social role theories to support the hypothesis that this will hold true. Next, the methods are described, and finally, I present results and discuss the findings.
More women are entering the workplace, and more policies push for equal representation of women in both the work force and in high-status or top-level positions. Despite these trends, women’s participation at the top of the employment chain is one of the slowest moving measures of gender equality (European Union, 2017). It is crucial to consider how these trends impact women’s SWB. This is not only true for women as a whole, but also for unique subpopulations. If current progress towards gender equality merely increases well-being for a small minority of working women, what else must be done to buffer the negative impact on other less privileged groups? More importantly, what can be done to ensure that they too benefit from gender equality? Why do we continue to strive for policy changes in gender equality if what is currently being done does not have a positive effect women’s SWB as a whole? Such a lack of consensus in this area asks important questions about how we evaluate
policy initiatives and their resulting progress. Furthermore, it forces us to consider the criteria we use to define both gender equality and success in today’s society, as well as the implications of these definitions on individual well-being.
2. THEORETICAL BACKGROUND 2.2. Gender equality
Gender equality includes equal access to opportunities, resources, and power among males and females. It is frequently evaluated using objective measures such as earnings equality, human capital ratio (e.g. Lagerlöf, 2003), as well as equal representation in the workforce, education, and high-status positions (e.g. Meisenberg & Woodley, 2015). Many studies also use ratios calculated from combined scores of multiple variables (e.g. Zuckerman et al., 2017), and some have used more subtle measures including cultural attitudes towards gender equality norms (e.g. Tesch-Römer et al., 2008). One of the most well-known measures of gender equality is the Global Gender Gap Index, which is published yearly by the World Economic Forum. The Index measures the relative differences between men and women in four main categories: health, education, economy, and political participation (World Economic Forum, 2011).
Such a wide variety of measurements exist due to the difficulty in pinpointing an exact definition or specific conditions of gender equality. The concept is much more complex than equal representation of men and women in the workforce, or of boys and girls in school. It is more than equality in income distribution or equal access to high-status positions. Many authors criticise the use of such objective measures, often citing the work of Amartya Sen (1992). Sen condemned income-based measures of inequality, suggesting that focus should instead be placed on more intrinsic capabilities and freedoms. These include the agency to lead a rewarding life and to enjoy a state of positive well-being. When examining this topic, we must understand that the mere presence of structural opportunities for women does ensure
gender equality. Sen’s approach therefore highlights the greater need to evaluate SWB. However, despite this need, many policy efforts aimed at gender equality remain focused on objective outcomes. For example, the EU’s 2017 Gender Equality Report refers specifically to targets including increasing female labour force participation, reducing gendered gaps in earnings, and promoting equal representation in positions of decision-making (European Union, 2017).
Gender equality is an important variable to study for numerous reasons. Firstly, ambitious links have been made between gender equality and social and economic progress. Moser (1989) termed this ‘the efficiency approach’ and various studies have promoted gender equality as a means of ‘smart economics’ (Chant & Sweetman, 2012). For example, gender equality has been correlated with macro level improvements in GDP and is frequently linked to economic development (Diener, Diener & Diener, 1995; Lagerlöf, 2003; Meisenberg & Woodley, 2015; Seguino, 2000). However, like its definition and measures, such a straightforward formula of gender equality’s impact has also been criticized for narrow-mindedness. Chant and Sweetman (2012) argue for a more comprehensive picture, perhaps including rights-based development or well-being measures. Oishi and Diener (2014) take this concept one step further. Instead of economic or political outcomes, the authors propose that the ultimate goal of public policy should be to enhance citizen’s happiness. As such, gender equality is not only important for economic development, but is crucial to examine due to its potential impact on livelihood and well-being.
2.3. Subjective well-being
Subjective well-being refers to ones’ perception of their own life. This is influenced by mood, emotional reactions to events or situations, and judgment of life satisfaction (Diener, Oishi & Lucas, 2003). SWB is most commonly measured by positive affect, negative affect, and life satisfaction. These three components are impacted by certain variables to differing degrees,
and may move in independent directions. Without distinguishing them in analyses, studies may fail to detect any differences (Batz & Tay, 2017).
Extensively studied, SWB appears related to demographic (income, marital status, health), individual (personality, temperament) and situational (societal, cultural, referent point) factors. Main findings from the literature suggest income is positively related to SWB, but that this relationship diminishes as income rises (Binswanger, 2006; Diener & Biswas-Diener, 2002; Dolan, Peasgood, & White, 2007). Relative income has also been found to have a stronger influence on SWB than absolute income (Dorn, Fischer, Kirchgassner, & Sousa-Poza, 2007; Easterlin, 1974). Physical and psychological health have been highly correlated with SWB (Dolan et al., 2007), and age is proposed to have a U-shaped relationship to SWB with younger and older individuals reporting the highest levels (Blanchflower & Oswald, 2004a). SWB has also been correlated with marital and employment statuses (Dolan et al., 2007; Treas, van der Lippe, & ChloeTai, 2011; Trzcinski & Holst, 2012), and conditions at work (Moen & Yu, 2000). However, some of these relationships may be the result of reverse causation, as higher SWB could lead to better psychological health, or more success in the labour market (e.g. Lyubomirsky, King, & Diener, 2005). Furthermore, given that there are so many related variables, these likely interact with one another.
Many studies have also examined the influence of gender on SWB. While some studies find that men have higher SWB (Montgomery, 2016; Stevenson & Wolfers, 2009; Tesch-Römer et al., 2008), others show that women are ahead (Graham & Chattopadhyay, 2013; Meisenberg & Woodley, 2015). Biological explanations for these observed gender differences in SWB have been proposed, however these are not well supported (Nydegger, 2004). Much more support can be found for the idea that external factors are at play. As Jorm and Ryan (2014, p. 337) argue, “There is clear evidence from nation-level studies that social, economic and political features of nations are associated with their SWB, and clear evidence from
multilevel analyses that these national features contribute to SWB over and above characteristics of the individuals that make up the nation.”
While merely a person’s subjective opinion, self-reported well-being scales yield acceptable levels of reliability which are found to be consistent over time (Lucas & Donnellan, 2007). Furthermore, I argue that SWB is more relevant to examine than objective measures of health or status. Regardless of personal situation, an individual’s perception of that situation is likely to be what drives their actual happiness. In other words, ones’ own reality is more relevant to their happiness than the reality which any external observer may perceive. Given that happiness has been purported as the ultimate goal of an ideal society (Oishi & Diener, 2014), SWB is of great concern to behavioural scientists.
2.4. Gender equality and subjective well-being
If structural and societal factors are likely to contribute to SWB, gendered SWB should fluctuate in relation to levels of gender equity. This relationship has been examined by numerous studies.
Bjørnskov et al. (2007) analysed human rights practices in 66 countries examining how changes in economic, political, and social rights impact gendered life satisfaction. While more favourable economic rights 20 years ago seemed positively correlated with female life satisfaction, present levels of both economic and social rights had no correlation. Research by Ruth and Napier (2014) found a broad positive relationship, discovering that country-level gender equality led to increases in the overall life satisfaction of a nation (men and women). Jorm and Ryan (2014) conducted a literature review also suggesting positive overall results. The review found that predictors of SWB were highly correlated with relative equality. While these studies suggest a positive relationship, they are broadly focused and examine national-level factors and national-national-level SWB. This may hinder our understanding of how societal conditions impact different groups in different directions, and could mask individual level
trends. For example, results from Bjørnskov et al. (2007) were heterogeneous among different societal groups, highlighting that gender equality impacts different groups in different ways, and validating the need to study unique subpopulations. As such, the authors concluded that no straightforward conclusions could be drawn on the relationship between gender discrimination and SWB, stating that whether women benefit from gender equality remains an open question. Despite some positive and broad-level findings, more of the literature reveals opposing results. Blanchflower and Oswald (2004b) found that despite increased resources, opportunity, and authority, women’s happiness decreased over time between the early 1970s and late 1990s in both the United States and Great Britain. Other studies have yielded similar results, finding that although women hold increasingly high-status and meaningful careers in both America and Europe, their level of happiness has declined in comparison to men’s (Easterlin, 2001; Stevenson & Wolfers 2009).
Numerous cross-national studies have also found a negative relationship (Tesch-Röemer et al., 2008; Vieira-Lima, 2011, Duflo, 2012). Among the most recent, Meisenberg & Woodley (2015) discovered among a sample of 95 countries that higher female status and greater levels of gender equality were associated with lower female (relative to male) life satisfaction. The study suggests that female labour force participation and non-agricultural employment reduced relative female well-being. Zuckerman, Li and Hall (2016) conducted a meta-analysis of 1148 studies from 2009 to 2013 in an effort to examine societal conditions and self-esteem. The authors found that women were further behind men in SWB in countries with better quality of life, superior government services, higher participation in the workforce, and values that promoted gender equality and freedom. Zuckerman et al. (2017) built on these findings and proposed a three-stage model suggesting that women’s SWB is highest under the best and worst societal conditions, and worst in moderately favourable societal conditions. The authors suggest that higher SWB results from an acceptance of the status quo in unfavourable
societal conditions. When gender equality first begins to improve, an initial decrease in women’s SWB occurs as women gain access to opportunity, yet become aware of continued disadvantage. In the third stage, women’s SWB is expected to increase, where true progress is made. However, as I will later highlight, even in countries with the best societal conditions, cultural attitudes and more subtle forms of gender inequality still exist. This raises the question of what else is necessary to make the third stage truly possible in the future.
It is important to note that many studies in this area focus on gender differences and compare female and male SWB to one another. This is problematic because women have been shown to report higher levels of positive and negative affect, having a greater propensity to express emotion (Batz & Tay, 2017). Furthermore, it seems unnecessary to pit the sexes against each other in a zero-sum game. Though an ideal situation would show no difference (equal SWB), focusing on comparisons may hinder our understanding. For example, if gender equity leads to a SWB improvement of an equal magnitude among both men and women, the gender difference in SWB would remain the same. This suggests that no progress has been made, when in fact both sexes have realized greater SWB. From a utilitarian point of view, this situation is better than no improvement in the SWB of either gender (or than an improvement in female SWB only). As such I will avoid comparing female SWB relative to male SWB. In order to gain a deeper understanding and closer focus, this study will look exclusively at women, examining the measures of life satisfaction and happiness. In line with the majority of existing research, I hypothesize that the overall relationship between gender equality and women’s SWB will be negative.
2.4.1. Gender equality only increases SWB for women at ‘the top’
While many causal explanations for a negative relationship between gender equality and female SWB have been proposed, these are understudied and rarely evaluated empirically (Batz & Tay, 2017). One example is the suggestion that while gender equality increases SWB for women at ‘the top’, it is likely to decrease SWB for a large majority of others who happen to be less successful (Meisenberg & Woodley, 2015; Zuckerman et al., 2017). While studies have compared the SWB of employed and unemployed women (e.g. Stevenson & Wolfers, 2009), the proposed explanation highlights the need to examine different types of working women, which is largely absent from the literature. The present study aims to confirm the validity of this statement by examining the moderating factor of top-level employment success on the relationship between gender equality and female SWB.
2.5. Low gender equality is not characterized by low SWB
To support the above explanation, I begin by proposing that low gender equality is not detrimental to female SWB for the average woman, but may be for the highly successful woman. Firstly, under system justification theory, women are likely to rationalize their disadvantaged position, shielding them from the emotional distress of inequality. Furthermore, relatively low societal expectations cause women to hold low aspirations and low referent points, resulting in satisfaction despite low achievements. Lastly, under social role theory, there may be a SWB advantage to following the status quo and avoiding the backlash involved with high achievement.
2.5.1. System justification theory: Women rationalize existing inequalities
In conditions of low gender equality, societal circumstances favour men over women. As such, women are likely to have fewer opportunities, resources, and power. Self-determination and needs fulfilment theories state that SWB levels depend on an individual’s ability to meet physical and psychological needs (Patrick, Knee, Canavello & Lonsbary, 2007). These include
the attainment of basic resources such as food, housing, and medical services. However, needs also include Maslow’s higher order intrinsic needs of autonomy, competence and relatedness (Ryan & Deci, 2000). As such, it may be assumed that gender inequality hinders women’s ability to meet these needs, resulting in low SWB.
However, women’s relative disadvantage may not have a detrimental impact on SWB. Objective conditions are not strongly correlated to individual satisfaction with those conditions, as the disadvantaged often report equal levels of satisfaction when compared to those more privileged (Major, 1994). According to system justification theory, this is because individuals strive to perceive the world as predictable, giving them a sense of control (Rankin, Jost, & Wakslak, 2009). As such, they have motivation to assume that society is just and fair. Major (1994) argued that this amendment to what individuals feel they deserve is one of the most prominent consequences of social inequality. Examples of system justification include the embracement of stereotypes that underprivileged individuals are actually happy, or that wealthy individuals are unhappy (Kay & Jost, 2003). Another example is the widespread belief that success is a result of hard work. This belief is held even among those who are the most disadvantaged, despite the fact that it implies that their own work ethic is to blame for their lack of success (Jost, Pelham, Sheldon, & Sullivan, 2003).
These assumptions serve a palliative function, providing relief from the emotional distress of disadvantage without actually easing inequality (Jost, Wakslak, & Tyler, 2008). As a result, people feel better despite their relatively disadvantaged position. Research has demonstrated that beliefs which rationalize existing inequalities can increase positive affect, decrease negative affect, and boost life satisfaction (Jost & Hunyady, 2002; Jost et al., 2008; Napier & Jost, 2008; O’Brien & Major, 2005). Called ‘the joy of sexism’ (Napier, Thorisdottir, & Jost, 2010), women’s disadvantaged status may therefore have no impact on SWB, perhaps even resulting in higher levels of SWB.
2.5.2. Relativity, low expectations, and a match between aspiration and attainment Evidence suggests SWB is relative, dependent on the referent group with which the respondent compares their own situation to (Diener & Lucas, 2000; Miles & Rossi, 2009; Suls, Martin, & Wheeler, 2002). The impacts of success, societal conditions or resources on SWB therefore depend on variable, individual standards. Support for the relativity argument comes from work by Richard Easterlin (1974) who found that differences in SWB between rich and poor countries were small and inconsistent. He concluded that this was because people rely on those around them as a standard of comparison. If they have more or are better off than those around them, they feel more satisfied, while if they have less or are worse off, they feel less satisfied. This idea may suggest that women have low SWB given that they are worse off than men. However, individuals tend to compare themselves to similar, intragroup others whose class and role statuses are alike their own (Suls et al., 2002). When the average person is asked to rate their intelligence, they do not compare themselves to Einstein when formulating a score. Similarly, female professional athletes do not decide that they have failed if they cannot achieve the athletic standards of the best men. As such, instead of comparing herself to men (or to women in other countries), the average woman likely compares herself to other average women around her. As low gender equality is typically characterized by low levels of female employment (European Union, 2017; World Economic Forum, 2011), women’s employment success is not a visible or common occurrence. For example, in 1890, less than 3% of married women worked outside of the home (Domenico & Jones, 2006). When gender equality is low, women do not see their mothers, sisters, or peers in high-status positions, or possibly in the workforce at all. As such, intragroup comparison suggests women may not be disappointed with their lives despite being worse off than men.
Furthermore, women may hold standards based on how they did in the past, or their own personal goals and expectations. In fact, individuals may be more likely to use personal
goals as a reference compared to any other standard (Diener & Lucas, 2000). In conditions of low gender equality, women are not expected to strive for (or succeed) in the labour market. There is little societal pressure to enter the workforce, to achieve big things, or to excel in the professional arena, and on top of this females typically have low levels of education. As such, is seems unlikely that women hold lofty career aspirations. Studies on gender and career aspirations in the 1970s found this to be true, reporting that girls had more restricted career aspirations than boys (Domenico & Jones, 2006). Heins et al. (1982) also found that while families encouraged the educational and professional aspirations of male children, the same was not true of female children.
Due to low societal expectations and a relatively intimate reference group, women are unlikely to strive for big goals, or to feel badly about themselves if they do not achieve great things. There is likely no decrease in SWB if they do not achieve employment success, as low achievement levels are matched by low aspirations. This is termed the aspiration-attainment hypothesis, which is suggested to be a crucial step in recent research on life cycle happiness trends among men and women (Matteucci & Vieria-Lima, 2014). The hypothesis proposes that a match between aspirations and attainment will promote SWB (Carr, 1997), and suggests women a gender unequal context may be happy despite few opportunities and little power.
2.5.3. Social role theory and gender bias
Social role theory suggests that women may actually enjoy a SWB advantage by sticking to the status quo and achieving very little professionally. This theory posits that men and women are categorized into distinct gender roles representing socially shared, consensual beliefs about appropriate attributes and behaviours (Eagly & Karau, 2002). Expectations and biases inherent in social roles are often unconscious and automatic, and while they could be seen as mere stereotypes, gender roles have significant impact. This is because people not only view and react to others based on their own gendered experiences, but also internalize their own gender
roles (Eagly, Wood & Diekman, 2000). When one violates these roles, others perceive this as jarring. As such, along with role incongruence comes backlash which is likely to decrease SWB (Eagly & Karau, 2002). Given that female workforce participation goes against traditional gender roles where women stay at home and men function as breadwinners, women’s employment success may result in role conflict and related SWB consequences.
It seems logical that gender roles and their associated bias should be most pronounced in societies characterized by very low gender equality. This has been empirically proven in studies which found that cross-cultural attitudes towards women and men were associated with levels of gender equality (Bjørnskov et al., 2007; Glick et al., 2000; Glick et al., 2004). Research also suggests that these cultural attitudes have different SWB impacts among different groups of women. Tesch-Röemer et al. (2008), discovered that in countries where cultural attitudes increasingly accepted gender inequality, female participation in the labour market was associated with lower female SWB. This idea is further supported by research that homemaking women report greater happiness than their working counterparts (Moen & Yu, 2000; Treas et al., 2011).
In summary, low gender inequality may not lead to low female SWB. Conditions typical in conditions of low gender equality include women’s disadvantaged position, low societal expectations, and distinct social norms. Although women are underprivileged in this setting, system justification theory provides evidence that women’s disadvantaged position may not lead to low SWB. This suggests that women may be happy despite limited opportunity and low employment success. Furthermore, low societal expectations of women cause them to hold relatively low aspirations and a low referent point. While they may not achieve much, this may not impact SWB as their achievement levels are congruent with those of women around them, and with their own low aspiration levels. On top of this, employment success may actually decrease women’s SWB under social role theory. In a setting with strict gender norms,
successful women are likely to face discrimination and backlash resulting from role conflict. This suggests a SWB advantage to not achieving high-level employment success. For these reasons, I hypothesize that under conditions of low gender equality, women at ‘the top’ will exhibit relatively low SWB, and women not at ‘the top’ will exhibit relatively high SWB.
H2: In conditions of low gender equality, women who experience top-level employment success will have lower SWB relative to those who do not.
2.6. What happens as gender equality improves?
In line with Meisenberg and Woodley’s (2015) proposition, I argue that conditions of increased gender equality will lead to a decrease in SWB for the average woman, but an increase in SWB for the few who are highly successful. This is due to increased expectations and aspirations, in combination with a lagging cultural bias which hinders the achievement of heightened aspirations. The result of this is low SWB for those women who experience a gap between aspirations and attainment, and higher SWB for those who achieve their high aspirations and enjoy the fulfilment of higher order needs.
2.6.1. Increased expectations and aspirations
As gender equality rises, women are increasingly expected to participate in the labour market. In fact, in many societies today there is little question of whether women will contribute to the workforce. As such, instead of deviations from the norm, working women are actually becoming the norm (Rainey & Borders, 1997). For example, in 2016, women’s employment rate in the EU was 65.5%, while men’s was 77.4% (European Union, 2017). Women are also equal (or more likely) participants in higher-level education (Peter & Horn, 2005), which has been shown to predict future employment (Domenico & Jones, 2006) and highlights a growing interest in professional careers among women. We not only see more women in the workforce
and in education, but also in other supplementary contexts. Successful women are depicted in films, in the media, and on the news, celebrating women’s success stories and promoting the ideal that women can accomplish all that men can.
These trends alter the way in which women imagine and evaluate their futures. Stevenson and Wolfers (2009) explain how women’s lives have become more complex as a result, and as such their SWB likely reflects satisfaction with a larger number of life factors and perhaps even a wider reference group which now includes men. It is therefore likely that due to women’s increased access to opportunity and society’s greater expectations, women possess greater aspirations, and hold higher expectations of themselves (Matteuci & Vieira-Lima, 2014). Support for rising aspirations comes from Gino, Wilmuth and Brooks (2015) who surveyed executives, MBA graduates, undergraduates, and working adults in the United States. The authors found that compared to men, women viewed high-level professional positions as equally attainable, despite the fact that women remain highly underrepresented at the top of the employment chain.
2.6.2. Barriers: Lagging cultural bias and competing responsibilities
We can also use social role theory to examine how gender roles impact women in more gender equal societies. Evidence suggests that women continue to face barriers despite improved conditions, perhaps preventing them from reaching aspirations. Nussbaum (2003) supports this concept, stating that in theory women experience greater opportunities and rights, but in practice social stigma, beliefs, and family commitments interfere with their achievement. Stevenson and Wolfers (2009) propose a similar idea, suggesting that the women's movement may have raised women's expectations faster than society was able to meet them. In fact, research does suggest that society has not yet caught up, as lagging cultural bias appears to exist despite increases in women’s participation in the labour market, even in developed countries. Duflo (2012) cites numerous implicit association tests which prove that wide-spread
implicit biases still exist (even in more gender equal societies), causing men and women to associate females with family and males with career. The author states that “as long as these biases persist, gender equality will be hindered even if the technological conditions for an even playing field are met” (Duflo, 2012, p. 1062). This has been labelled ‘second generation gender bias’ (Ibarra & Petriglieri, 2016, p. 1), and is related to the subtle forms of bias accounting for women’s initial decrease in SWB in Zuckerman et al.’s (2017) second stage.
While they may be less pronounced in more gender equal societies, implicit biases are not absent altogether. Examples include the idea that women are not cut out for leadership. For example, Schein (2007) found that both male managers and male management students held similar gender biases to those held in the 1970s, viewing males as more likely to possess the characteristics required for management success. Another example is the deep-rooted idea that women are responsible for home and family care. Even among women holding Harvard MBAs, over two thirds have been found to take primary responsibility for childcare (Ely, Stone, Ammerman, 2014). Furthermore, stereotype associations suggest women of the future are expected to be both high-earning professionals and primary caregivers, while men’s roles are not expected to change (Diekman & Eagly, 2000; Diekman, Goodfriend, & Goodwin, 2004). Many studies highlight the difficulties women face in balancing career and family, and this barrier as well as other gender stereotypes are extensively emphasized by the literature. As such I will not elaborate on them. Their presence is endorsed by the fact that women tend to hold lower-status, lower-paying jobs, and still remain grouped in a limited number of career types (Tinklin, Croxford, Ducklin, & Frame, 2005).
Although they have real impact and are underlined by numerous examples, subtle gender biases are not often taken into account when measuring gender equality (Matteuci & Vieira-Lima, 2014). Furthermore, even the use of subjective measures may not adequately uncover implicit bias. For example, in Tesch-Röemer et al. (2008), cultural attitudes towards
gender equity were measured using the degree of agreement with the statement: When jobs are scarce, men should have more right to a job than women. This is a bold declaration, yet attitudes occur along a spectrum. There is a difference between agreeing with gender fairness and equal access, and holding no gender preconceptions whatsoever. Using self-reports to measure individual prejudice levels is also problematic, as ones’ actions may not be consistent with the views they claim to hold. Greenwald and Banaji (1995) highlight that the very essence of implicit bias is that it exerts influence on judgement in a way that is unrecognized by the individual. The authors give examples of numerous studies revealing that individuals who explicitly disapprove of prejudice are often found to discriminate themselves.
2.6.3. The resulting gap between aspiration and attainment
Conditions of gender equality are marked by increased aspirations and expectations, but also by persistent gender bias. This bias inhibits women from being able to realize their increased aspirations. As highlighted earlier, the resulting gap between aspiration and attainment is likely to impact SWB. The larger the gap, the bigger the let-down, and the lower the SWB. As a result, while the high-achiever is satisfied, the average woman may be increasingly disappointed.
Support for this idea comes from Carr (1997), who examined women’s health in comparison to the fulfilment of career goals at midlife. Results indicate that women who have fallen short of their early career goals exhibited lower levels of life purpose and increased levels of depression. This was consistent even after controlling for factors including social background, human capital, family, and health characteristics. Further support for the a SWB disadvantage comes from Ely et al. (2014), who conducted a longitudinal study of MBA students at Harvard Business School. The authors found that aspirations about what highly ambitious women and men hoped for in their lives and careers did not differ much. Women believed they could get to the top just as much as men did, and planned to do so just as much
as men did. However, despite equal aspirations and capabilities, women ended up less satisfied with their careers, in unfulfilling positions. Conversely, men ended up more successful, emphasizing the difference between women’s expectations and reality as many women experienced the out-shadowing of their own career by their partner’s.
While those who do not meet their aspirations are expected to report lower SWB, those who achieve high success despite the barriers are expected to report comparatively high SWB. As we have seen, both goal attainment and congruence between aspirations and achievement promote personal satisfaction and increased SWB. Top level career success is also likely to fulfil higher order needs such as autonomy and competence, which have been shown to promote SWB (Ryan & Deci, 2000). Furthermore, if we return to the relativity argument, highly successful women are achieving more than most women around them, which may make them feel increasingly satisfied as they compare their own success to others’. Lastly, while cultural bias still exists, gender roles and their associated bias are less pronounced in more gender equal societies (Bjørnskov et al., 2007; Glick et al., 2000; Glick et al., 2004). High gender equality should therefore allow professional women to evade the much harsher forms of backlash that lead to decreased SWB in conditions of very low gender equality.
In summary, alongside increases in gender equality come parallel increases in women’s own aspirations as well as societal expectations of them. While women are increasingly told that they can ‘have it all’, lagging cultural bias and the associated barriers limit the degree to which women can actually achieve their heightened aspirations. The resulting gap between aspirations and attainment decreases SWB. While successful women ‘at the top’ may benefit from more lenient social roles and the fulfilment of higher order needs, the rest are left unsatisfied. As such, it is hypothesized that in conditions of gender equality, women who experience top-level employment success will report relatively high SWB in comparison to those who do not.
Top-level employment success is therefore hypothesized to positively moderate the relationship between gender equality and women’s SWB. While women at ‘the top’ are expected to benefit from increased gender equity and the relaxation of social roles, the rest are expected to experience a decrease in SWB as gender equality increases. As women who experience top-level employment success make up only a small percentage of women in general, this further supports H1 that the overall relationship between gender equality and women’s SWB is negative.
H3: In conditions of high gender equality, women who experience top-level employment success will have higher SWB relative to women who do not experience top-level employment success.
H4: Employment success will positively moderate the relationship between gender equality and women’s SWB.
3. METHODS
3.2. Sample and data collection
Data on SWB and employment success are sourced from the European Quality of Life Survey (EQLS). Data for the present study come from the 2011 edition, which was conducted in 33 countries across Europe (N = 43,636). Samples of the adult (age 18 and older) population were selected randomly for face to face interviews (Eurofound, 2017). The survey took roughly 40 minutes to complete, and participation was voluntary (Eurofound, 2017). The subset targeted for the present study includes exclusively women. While men are also impacted by gender equality, the most immediate aim of gender equality initiatives is to improve circumstances for women. As such they are of primarily interest. To achieve this focus, all male cases were
excluded, leaving N = 24,848 cases out of the original N = 43,636. Missing values were dealt with in a listwise manner. The average age of the women in the sample was 50 years old.
3.3. Variables and measures 3.3.1. Independent variable
Gender equality is measured using the 2011 Global Gender Gap Index, which has been published annually since 2006. Scores from the 2011 index were used in alignment with the EQLS. The Index measures gender-based disparities in health, education, economy, and political participation (World Economic Forum, 2011). It is formulated using a four-step procedure. See Table 1.0 for an overview of the calculation processes involved.
The Index was chosen as it one of the most well-known measures of gender equality (Mæland, 2015), combining four different components. Access to opportunities in the four areas likely influence the intrinsic freedoms that Sen (1992) highlighted as central to true gender equality. Lastly, the Index measures gaps in access to resources and opportunities instead of overall levels (World Economic Forum, 2011). This prevents measures of gender equality from rewarding countries that are simply more developed. Three countries included in the EQLS dataset were not found in the 2011 Index. While scores for these countries in more recent years were found, there is no way to predict gender equality in 2011 based on previous years. As such, respondents from these countries were excluded from analysis.
3.3.2. Dependent variable
Data on SWB was obtained from the EQLS survey. SWB is measured using two variables, each measured on a 10-point scale. Happiness is measured by response to: Taking all things together on a scale of 1 to 10, how happy would you say you are? Life satisfaction is measured by response to: All things considered, how satisfied would you say you are with your life these days? Please tell me on a scale of 1 to 10. SWB includes positive affect, negative affect, and
life satisfaction. While the EQLS measured affect, this was done using statements that asked how respondents had been feeling over the past two weeks. These responses could be influenced by short-term fluctuations in SWB. This study is interested in long-term well-being and the ability to lead a fulfilling life. The questions related to happiness and life satisfaction were more overarching and included the verbiage taking all this into consideration. As such, they were chosen for analysis. While happiness and life satisfaction were found to have a reliable Cronbach’s Alpha of .780, these will be analyzed separately as combining them may limit observation (Batz & Tay, 2017).
3.3.3. Moderator
Data on employment success was obtained from the EQLS survey. This study refers to success at a top-level, classified by holding a management or professional occupation. While success could be associated with different occupations in different cultures, it is assumed that high-status, non-entry level positions requiring advancement, experience, and education are associated with success in the majority of Europe. Furthermore, they are prestigious career-type jobs, where individuals mark their achievements through advancement (Wrzesniewski, McCauley, Rozin, & Schwartz, 1997). The jobs included in these categories are also in line with existing literature on top-level employment (e.g. Lippa, Preston, & Penner, 2014). See Table 2.0 for a list of the occupations included.
Survey responses indicating occupation were used to create a new binary variable reflecting employment success. Management and professional occupations were given a score of 1, and all others were given a score of 0. Because those not working did not answer the occupation question, there was a significant number of respondents excluded. As H4 aims to compare women at the top with ‘the rest’, the group of women with a score of 0 needed to also include non-working women. To achieve this, a survey question asking about respondents’
current situation (e.g. employed, unemployed, full-time homemaker, retired…) was used. Respondents who did not answer as employed were given a score of 0 for employment success, in order to include them in the analysis.
3.3.4. Control variables
Variables consistently shown to influence SWB include income (Binswanger, 2006; Diener & Biswas-Diener, 2002; Dolan et al., 2007), age (Blanchflower & Oswald, 2004a; Trzcinski & Holst, 2012), and health (Dolan et al., 2007). Due to their correlations with SWB, these factors may influence results and will be controlled for. Data for each of these are obtained from the EQLS. Health was reverse coded so that a high numerical score was related to high level of health.
4. RESULTS
To test the main effect, the linear relationship between gender equality and each of the dependent variables (life satisfaction, happiness) was first examined. In simple hierarchical regression, it is assumed that intercepts and slopes are fixed and do not vary across groups. However, country-level variation may exist and this could be of substantial interest. Country clustering can also cause problems when using simple regression as it can inflate type 1 error probability (Wears, 2002). As such, a multilevel (two-level) model was used for analysis (Heck, Thomas, & Tabata, 2013). Following examination of the main effect, interactions between the variables were probed to test for moderation.
4.2. Descriptive statistics
All variables of interest were pulled from the dataset and recoded where necessary. Frequencies were checked to examine whether there were errors in the data, however none were found. Aside from income (25.1%), the frequency of missing values for all variables was < 0.5%.
Missing values were dealt with in a listwise manner. Descriptive statistics and normality tests were then run. Happiness showed a skewness of -.809 and kurtosis of .490. Life satisfaction showed a skewness of -.755 and kurtosis of .172. These values are all within the acceptable +/-1 range. Furthermore, with such a large sample size, skewness and kurtosis are unlikely to make a substantial difference in data analysis (Tabachnick & Fidell, 2001).
On average, respondents tended to have relatively high SWB, reporting mean scores of 7.0 for life satisfaction and 7.3 for happiness. Respondents were on average older, and had fairly good levels of health (an average rating of 3.63 out of 5). Only 11.8% rated their health as 4 (bad) or 5 (very bad), while 57.4% rated their health as 2 (good) or 1 (very good). Only 10.3% of respondents fell under the classification of employment success. Means, standard deviations and correlations of the variables are provided in Table 3.0. Table 4.0 reveals the means for life satisfaction and happiness when grouped by category of employment success. Both mean happiness and mean life satisfaction were higher in the presence of employment success than in the absence.
4.3. Hypothesis testing 4.3.1. Multi-level model 4.3.1.1. Step 1: Null hypothesis
First, the null hypothesis is tested to determine whether significant variation in the Level 1 (L1) residuals and Level 2 (L2) means exists. In other words, whether SWB varies across countries. Under the null hypothesis model, we can use the following equations to represent variation at each level:
Level 1 Yij = β0j + εij
εij = β0j – Yij
The first equation states that the SWB score for Person I in country J is a function of the intercept for group J, plus the prediction error at L1. The second equation represents that
random prediction error at L1 is equal to the group intercept (mean) of SWB minus each individuals’ score (within that group) on SWB.
Level 2 β0j = γ00 + µ0j
µ0j = γ00 + β0j
Here, the first equation states that the intercept for country J is equal to the average of the intercepts across all countries (at L2), plus random error. The second equation represents random error, where random error at L2 is equal to the difference between the country intercept and the average of intercepts across L2 units. Using substitution, we get a mixed model that includes a fixed component (γ00), an L1 component representing the residual at L1 (εij), and an
L2 component capturing the random variation of the intercept around the average of those intercepts:
Mixed model Yij = γ00 + µ0j + εij
Tables 5.0 and 5.1 display the results using the dependent variable of life satisfaction. The fixed effect for intercepts across countries (γ00) is 6.93 (SE = .13, p < .001). The variance
in L1 residuals (εij) within countries is 4.46 (SE = .05, p < .001), and the variance in intercepts
across or between countries (µ0j) is 0.55 (SE = .14, p < .001). Tables 6.0 and 6.1 display the
results using the dependent variable of happiness. The fixed effect for intercepts across countries (γ00) is 7.236 (SE = .103, p < .001). The variance in L1 residuals (εij) within countries
is 3.74 (SE = .039, p < .001), and the variance in intercepts across or between countries (µ0j) is
0.35 (SE = .09, p < .001).
For both life satisfaction and happiness, there is significant variance in the residuals at both L1 (within countries) and L2 (between countries), suggesting possible clustering. Given that the null hypothesis would reflect a situation in which each of the variance components was
zero, we can reject the null hypothesis in both cases. We can also calculate the intra-class correlation component (ICC) to determine whether there is significant clustering within higher level units. This is computed by dividing the variance between countries by the sum of the variation between countries and the variation within countries:
ICC= variation between countries (variation of εij)
variation between countries (variation of εij ) + variation within countries (variation of µ0j)
Using this equation, we get ICC values of 0.110 for life satisfaction and 0.085 for happiness. These numbers represent the proportion of the total variance in SWB (life satisfaction or happiness) that is explained by country clustering. Given that both of these values are above 0.05, substantial clustering of observations within L2 units is indicated (Heck et al., 2013). The proportion of variation in life satisfaction that lies between countries is approximately 11.0%, and the proportion of variation in happiness that lies between countries is approximately 8.6%.
4.3.1.2.Step 2: Level 1 model
As significant clustering was found, the next two steps were used to investigate whether the variances could be explained by predictor variables. This will help answer the question of whether a relationship exists between gender equality and SWB at L1. To do so, we first add a L1(fixed) predictor, while allowing intercepts to vary across groups.
Level 1 Yij = β0j + β1jGEij + εij
This equation represents person I in country J’s SWB score, which is equal to the school level mean that is also adjusted for the predictor variable of gender equality. β0j represents the
country level mean or intercept, β1j represents the slope for GE within country J, GEij represents
person I’s value for GE, and εij represents error.
Level 2 β0j = γ00 + µ0j
β1j = γ10
The first equation represents the intercept, which is equal to the average of the intercepts across countries, plus the variation in country intercepts around the grand mean, while the second equation represents the fixed slope which is averaged across all countries for GE. In the second equation, the ‘1’ represents predictor 1 in the model, where the ‘J’ represents a given country. Here we are fixing the slope of GE to be constant across countries.
Mixed model Yij = γ00 + γ10GEij + εij + µ0j
Using substitution, we get the mixed model equation. Here, each person’s (in country J) score on Y is equal to the average intercept across countries (γ00), plus the product of the average
slope across countries (γ10) and the level of gender equality (GEij), plus that person’s L1
residual (εij), and L2 residual (the difference between country mean and average mean, µ0j).
Looking at the estimates of fixed effects for life satisfaction in the L1 model (Table 7.0) we can see that γ00 = 6.94 (SE = 0.10, p < .001), which represents the average of the intercepts
across the countries, adjusted for gender equality. The regression coefficient is estimated by γ10= 0.53 (SE = 0.10, p < .001). This suggests that gender equality is a positive predictor of life
satisfaction, as for every 1 unit increase in gender equality there is a predicted 0.53 unit increase in life satisfaction. Looking at the estimates of covariance parameters for life satisfaction (Table 7.1), we can see that the variation in predictor errors (within countries) at L1 is equal to 4.41 (SE = .05, p < .001), and the variation in intercepts (between countries) is equal to 0.28 (SE =.07, p < .001). Using these values, we can calculate an ICC of 0.060 representing the
clustering effect after taking gender equality into account. This value is above 0.05 and indicates significant variation at the country level.
Looking at the estimates of fixed effects for happiness in the L1 model (Table 8.0), we can see that γ00 = 7.23 (SE = .07, p < .001), which represents the average of the intercepts across
the countries, adjusted for gender equality. The regression coefficient is estimated by γ10= 0.42
(SE.=.07, p < .001). This suggests that gender equality is a positive predictor of happiness, as for every 1 unit increase in gender equality there is a predicted 0.42 unit increase in happiness. Looking at the estimates of covariance parameters for happiness (Table 8.1), we can see that the variation in predictor errors (within countries) at L1 is equal to 3.70 (SE.= .04, p < .001), and the variation in intercepts (between countries) is equal to 0.16(SE = .041, p < .001). Using these values, we can calculate an ICC of 0.041. This value is below 0.05, suggesting that not too much clustering exists within countries after gender equality is taken into account. However, for the sake of results and comparison, the final step of the multi-level model will be run on both life satisfaction and happiness.
4.3.1.3.Step 3: Level 2 model
Here, the role of contextual factors plays in variation in the intercepts and/or slopes is examined. This will address the question of whether the effect of gender equality on SWB compounds at the country level. To address any compounding effect, country mean levels of GE must be computed, which will serve as a L2 predictor of variation in country SWB averages. As gender equality does not vary among individuals within the same country, the mean of a certain country is simply equal to the level of gender equality in that country.
Level 1 Yij = β0j + β1jGEij + εij
Level 2 β0j = γ00 + γ01GEmeanj + γ02EmpSucc + µ0j
Variation in L2 intercepts β0j are predicted by average gender equality within country J, as well
as employment success. Using substitution, we get the mixed model equation, which contains both the fixed coefficients (γ00, γ01, γ02, γ10) as well as the random components (µ0j, εij) that
aren’t explained by the predictors:
Yij = γ00 + γ01GEmeanj + γ02EmpSucc + µ0j + γ10GEmeanj + εij
Table 9.0 shows that the grand mean of the intercepts for life satisfaction across countries, after accounting for controls, gender equality, and employment success is γ00 = 6.87
(SE =.10, p < .001). As scores increased on GE there was a predicted increase in life satisfaction, γ01 = 0.51 (SE =.10, p < .001), and therefore females from countries with higher
gender equality scores tended to report higher levels of life satisfaction. This rejects H1. The predictor of employment success resulted in a regression coefficient of γ02 = 0.57 (SE = .05, p
< .001). Table 9.1 shows that εij = 4.38 andµ0j = 0.28. Both values are statistically significant
and yield a CCI of 0.060.
Looking at happiness, Table 10.0 shows that γ00 = 7.16 (SE = .07, p < .001). As scores
increased on GE there was a predicted increase in happiness, γ01 = 0.40 (SE = .07, p < .001),
and therefore females from countries with higher gender equality scores tended to report higher levels of happiness. This rejects H1. The predictor of employment success resulted in a regression coefficient of γ02 = 0.55 (SE = .05, p < .001). Table 10.1 shows that εij = 3.68 andµ0j
= 0.15. Both values are statistically significant and result in a CCI of 0.039.
For both life satisfaction and happiness, some variation remains unexplained, despite the addition of employment success at L2. While the CCI suggests other predictors (not included in the model) result in the variation of life satisfaction, happiness had a low value of .040 which suggest little country level variance exists.
4.3.2. Moderation
Both gender equality and employment success were mean-centred, and the product of the two predictors was then used to create a new variable reflecting the interaction term. A multi-level analysis was run including both predictors (gender equality, employment success), the controls, and the interaction term.
Table 11.0 shows the results of the analysis using life satisfaction as the dependent variable. The regression coefficient of the interaction term is equal to -0.124. This value was significant at p < .05, suggesting that there is a moderating effect of employment success on the relationship between gender equality and life satisfaction. Table 11.1 shows that εij = 4.38
andµ0j = 0.28. Both values are statistically significant and result in a CCI of 0.060.
Table 12.0 shows the results of the analysis using happiness as the dependent variable. The regression coefficient of the interaction term is equal to -0.152. This value was significant at p < .001, suggesting that there is also a moderating effect of employment success on the relationship between gender equality and happiness. Table 12.1 shows that εij = 3.67 andµ0j =
0.16. Both values are statistically significant and yield a CCI of 0.041.
After a significant result was found, the interaction was unpacked using simple slopes. Employment success was re-centred twice, at plus one and minus one standard deviation. Two new interaction terms were created, one representing high employment success, and one representing low employment success. The new interaction terms were then re-run separately through the model to determine how high and low levels of employment success impact the relationship between gender equality and SWB.
For life satisfaction, Tables 13.0 and 13.1 reveal that the regression coefficients for the main effect of gender equality are different at high and low levels of employment success. At low levels of employment success (Table 13.0), the coefficient for gender equality was 0.55 (SE = .10, p < .001). At high levels of employment success (Table 13.1), the coefficient for
gender equality was 0.47 (SE = .01, p < .001). This suggests that gender equality has less of an impact on life satisfaction (a flatter slope) when employment success is high, and a more positive relationship with life satisfaction (a steeper slope) when employment success is low. While gender equality has positive coefficients (and a positive overall relationship) in both cases, gender equality appears more beneficial to life satisfaction for women experiencing low levels of employment success. This reject H4. See Figure 2.0 for a visual representation of the moderating effect.
For happiness, Tables 14.0 and 14.1 reveal that the regression coefficients for the main effect of gender equality are different at high and low levels of employment success. At low levels of employment success (Table 14.0), the coefficient for gender equality was 0.45 (SE = .07, p < .001). At high levels of employment success (Table 14.1), the coefficient for gender equality was 0.36 (SE = .07, p < .001). Like in the case of life satisfaction, this suggests that gender equality has less of an impact on happiness (a flatter slope) when employment success is high, and a more positive relationship with happiness (a steeper slope) when employment success is low. As such, gender equality appears to be more beneficial to happiness for women experiencing low employment success. This rejects H4. See Figure 3.0 for a visual representation of the moderating effect.
Looking at Figures 2.0 and 3.0 we can also assess H2 and H3. H2 expected that when gender equality was low, women who experienced employment success would have lower SWB relative to those who did not experience employment success. This hypothesis is rejected, as support was found for a positive relationship between gender equality and SWB under both conditions of employment success. As such, results provide support for H3, that in conditions of high gender equality, women who experience employment success will have higher SWB relative to those who do not experience employment success.
5. DISCUSSION 5.2. Findings
5.2.1. Life satisfaction vs. happiness
As life satisfaction and happiness were analyzed separately, we can consider how results for the two SWB variables differed. Mean happiness was higher than life satisfaction both in the overall sample, as well as within both categories of employment success. This provides support for the idea that the two are not identical constructs, and may suggest that individuals have a tendency to rate happiness higher than life satisfaction. Many researchers conclude that the two variables reflect different dimensions of SWB (e.g. Diener, Ng, Harter, & Arora, 2010). While happiness may be more emotionally driven (Lyubomirsky et al., 2005), life satisfaction seems to be a more calculated evaluation of one’s life and situation as a whole. For example, life satisfaction is thought to be more closely related to economic factors or socio-economic changes in society, whereas happiness is more closely related to personal factors (Saris & Andreenkova, 2001). This distinction should be considered when weaving through a discussion of the results.
One possible explanation for higher happiness ratings could be connected to referent points and the visibility levels of the variables. For example, differences in the life or situational circumstances of others are more visible than differences in others’ happiness. It is easy to observe and acknowledge a wide range of differences in situational or economic circumstances among individuals and across the globe. As such, we may be more likely to make comparisons when evaluating this measure, causing us to feel that our situation could be better. In comparison, the happiness of others is not as easy to determine from the outside. Furthermore, when generally thinking about happiness we are less likely to consider objective circumstances. These factors may therefore make us less likely to make comparisons at all. As such, happiness may be less relative, and evaluations of how we feel internally could be higher than calculated
