Socioeconomic status and social capital as predictors of happiness: evidence and gender differences

Abstract

Understanding socioeconomic status (SES) and social capital as predictors of life outcomes has remained central to humanitarian, health and social research. This current study explores how socioeconomic status and social capital (community solidarity, locus of control and generalised trust) predict happiness among a sample of South African adults. Possible gender differences were also considered. Cross-sectional data from (n = 1049) South Africans on SES, social capital, happiness and sociodemographic characteristics were collected using an online questionnaire between January 2021 and September 2021. A correlation matrix was calculated to explore bivariate associations between the variables. Hierarchical regression analysis was conducted to examine the predictive influence of SES and social capital indices on happiness and the gender differential in these relationships. SES and all social capital variables were significantly correlated with happiness. The model, including SES and social capital, accounted for 25% of the variance in happiness. Gender differences were found regarding the relationship between SES and all social capital variables (except locus of control) and happiness. The results confirm that efforts to facilitate individual social capital correspond to greater happiness among South Africans. Furthermore, it highlights the need for further empirical exploration of social predictors of life outcomes. The current analysis provides a framework for developing and implementing policies that promote happiness and other life outcomes.

Introduction

Research on socioeconomic status (SES) suggests that individual subjective life evaluation, health behaviour, and treatment may differ based on their economic position (Hoffmann et al., 2018; Proctor et al., 2017). Findings from this research submit that people with lower SES are more likely to report poorer subjective health (Hoffmann et al., 2018; Wang and Geng, 2019) and lower satisfaction with life (Adedeji et al., 2021a; Proctor et al., 2017), and may suffer from more chronic and long-term conditions (Sahni et al., 2017; Zeng et al., 2018). Similarly, social capital as a concept of social interactions and community participation (Ehsan et al., 2019) facilitates health promotion and subjective wellbeing (Wind and Villalonga-Olives, 2019; Zeng et al., 2018; Zhang et al., 2017). Sociological and health studies have classified social capital as resources that exert their effects by providing individuals and communities with mechanisms to address adversities and strengthening network bonds for individual, mutual or communal benefits (Adedeji et al., 2021a; Ward et al., 2021). The observational science of subjective wellbeing is often assessed regarding happiness, quality of life or life satisfaction (Diener et al., 2018). However, unlike happiness, a standard definition of social capital has proven problematic. The underlying dynamic and complex context entangled in this construct makes it almost impossible to agree on what is included as social capital. The current study considers social capital as a social infrastructure facilitating social communication and exchange (Adedeji et al., 2021b; Wind and Villalonga-Olives, 2019).

As a theoretical construct, happiness is described as a positive emotion with multifaceted nature connected to the optimal functioning of behavioural systems (Averill and More, 1993). Over the last decade, understanding individual happiness has become popular among social and humanitarian researchers. This popularity is attributed to the empirical connection between happiness and health (Petrovič et al., 2021; Scorsolini-Comin and Santos, 2010) and the increasing recognition that social and environmental factors are crucial for health and life outcomes.

Various studies have explored the association between happiness, SES, and social capital (Fournier, 2020; Maharlouei et al., 2020; Rodríguez-Pose and von Berlepsch, 2014). Fournier (2020), for example, found that SES might serve as an instrumental resource in people’s strivings for autonomy and relatedness, allowing them to exert more control over their lives, thereby promoting happiness (Fournier, 2020). Similarly, Maharlouei et al. (2020) explored indicators of happiness using data from the General Social Survey (GSS) between 1972 and 2018 in the United States. They found a significant association between SES indicators and happiness. They, however, also found that the interactions between gender and the SES variables were significant. This suggests that even though SES may predict happiness, these effects vary based on gender (Maharlouei et al., 2020).

Furthermore, Rodríguez-Pose and von Berlepsch (2014) tested the hypothesis of whether social capital influences individual happiness across European countries and regions. They found social capital as a predictor of happiness. However, socioeconomic status and demographic features such as age and gender influenced the connection between social capital and happiness. They concluded that the diversities of peoples, cultures, traditions and social institutions would likely affect the interaction between social capital and happiness in different ways (Rodríguez-Pose and von Berlepsch, 2014). These findings have suggested group differences in the significance and predictors of happiness based on study settings and population characteristics such as gender.

Although researchers have attempted to understand how social capital and happiness interact, these studies have focused on Europe and other economically advanced nations (Doherty and Kelly, 2010), where the formation and utilisation of social capital are different compared to developing countries (Dinda, 2008; Fine, 2004; Godoy et al., 2007; Maluccio et al., 2000). Furthermore, the results of these studies remain largely inconsistent as they strongly depend on the conceptualisation of indicators used to capture social capital (Rodríguez-Pose and von Berlepsch, 2014).

As a result of South Africa’s complex history (Colvin, 2017; Davenport and Saunders, 2000; Feinstein, 2005) and multiculturalism, the issue of discrimination and social segregation continue to play a significant role in social communication and exchange (Spiegel, 2018; Steyn et al., 2019). Components of social capital such as trust (Zhang, 2020), solidarity (Seedat and Lazarus, 2011) and empowerment through the locus of control are assumed to facilitate indigenous welfare practice (Patel et al., 2007) and are considered to be crucial for wellbeing (Pronyk et al., 2008).

As a multidimensional psychosocial domain, trust is long referred to as a key to positive interpersonal relationships in multicultural settings (Fox, 1974) and central to interacting with others (Berscheid, 1994). A study among the elderly in Durban, South Africa, projects trust as an essential determinant of mental wellbeing (Chipps and Jarvis, 2016). Similarly, community solidarity is beneficial behaviour (such as providing support, collaboration, and cooperation) that facilitate mutual habitation (Laitinen and Pessi, 2014) and connect people in a community to foster collaboration (Althammer, 2019). On the other hand, locus of control is a psychological concept capturing “whether or not the person perceives a causal relationship between his own behaviour and the reward” (Rotter, 1966, p. 1). This concept is crucial for determining health, especially among disadvantaged groups (Kesavayuth et al., 2020).

While these social predictors of wellbeing have been extensively studied in western settings, the cultural relevance and the socioeconomic stand in South Africa and other countries in Africa may encourage attributing different meanings or amplifies the importance of these determinants for wellbeing (Gyekye, 1997; Mahlangu et al., 2021; Moyo and Dhliwayo, 2019; Werner, 2003). Understanding how these three components of social capital (trust, community solidarity and locus of control) interact with happiness in South Africa will allow for a more practical application of social determinants of wellbeing in facilitating and promoting better life outcomes. Similarly, re-evaluating the projected SES (Doherty and Kelly, 2010) and gender (Hori and Kamo, 2018) differences might explain why previous researchers have found significant differences in wellbeing in South Africa favouring females (Etinzock and Kollamparambil, 2019).

Research objectives

The current study examines the predictive influence of SES and social capital variables on happiness and how this influence varies by gender among a South African sample. The following specific objectives were set:

• 1. Examine SES as a predictor of happiness

• 2. Investigate the association between social capital variables (community solidarity, locus of controls, and generalised trust) and happiness

• 3. Explore the effect of gender as a moderator of the relationship between SES, social capital variables, and happiness.

Method

Study design and sample characteristics

Data on happiness, social capital, SES and demographic characteristics were collected across the nine provinces of South Africa. A total of 1062 South Africans participated in the cross-sectional survey between January 2021 and September 2021. The survey questionnaire was accessible online using the LimeSurvey Platform. The standardised questionnaires were administered in the English language. Cases with extensive missing data (n = 13) were removed from the dataset. Given the importance of gender as a central variable for the current analysis, only data from participants who identified as male or female were included. Data from participants who identified as “other” or preferred not to disclose their gender was removed from the inferential analysis due to an incomparable sample size (n = 49). The final sample for the analysis was 1000.

The participants in this study were recruited using snowball sampling techniques. Information about the survey was shared on social media (e.g., Facebook, Twitter, and WhatsApp), the project webpage (www.beliv-study.com), and with personal contacts and professional networks. After completing the survey, participants were encouraged to share the link with their friends, family, and contacts. Each participant was offered a 10 Rand (approximately 0.56 €) voucher for a telephone service provider as an incentive.

A descriptive analysis of sample characteristics (see Table 1) shows a mean age of 26.36 years (SD = 7.20, range = 18–45). Furthermore, about 63% of the participants are female. The percentage distribution of participants’ racial identity confirms that a Black African majority constituted approximately 78.6% of the sample. About 11.1% identified as Coloured/mixed race, 7.8% as Whites and 2.5% as Indian South Africans. This distribution mirrors the South African racial distribution as reported in the 2016 Census (Statistics South Africa, 2016).

Table 1 Sample characteristics (N = 1000).

Measures

Happiness

Participants’ subjective happiness was measured using a single item scored on a 5-point Likert scale ranging from “very unhappy” to “completely happy”. This single measure of happiness has been tested and found to capture individual subjective evaluations of their life (Abdel-Khalek, 2006; Chae, 2018; Sarracino and O’Connor, 2021).

Social capital

Social Capital was measured using the World Bank Social Capital Scale (Grootaert et al., 2003). For the current study, we examined three aspects of social capital:

• 1. Community Solidarity

  Participant community solidarity was measured using a subscale from the World Bank Social Capital Scale (Grootaert et al., 2003). The two items evaluate participants’ likelihood of cooperating to solve community problems and the willingness of others to help if needed. For example: “If there is a problem in your community (for example, a water problem), how likely is it that people will come together to solve it?” These two items were ranked using a 5-point Likert scale ranging from “very unlikely” to “very likely”, with a sum score between 2 and 10. A higher score suggests stronger community solidarity. Internal consistency (Cronbach’s Alpha) for the two items was 0.33. Although this value is low, research supports that a scale with fewer items may show poor internal consistency (Gliem and Gliem, 2003).

• 2. Generalised Trust

  Data on trust was collected using a measure of an individual’s general level of trust towards others (Arbor, 1971). This has been established as a valid measure of trust in the social context (Körber, 2018; Kulin and Johansson Sevä, 2021). The item was coded into dichotomous choices. One of the two choices is the high trust response, and the other is the low trust response. Generalised trust was measured with the item: “Generally speaking, would you say that most people can be trusted or that you cannot be too careful in dealing with people?” This item was coded as 1 = most people can be trusted; 0 = can’t be too careful. A 0 score suggests low trust, while a score of 1 implies a high level of trust.

• 3. Locus of Control

Participants’ locus of control was measured as their ability to make important life decisions. The item “Do you feel able to make important decisions that will affect the course of your life?” from the World Bank Integrated Questionnaire for measuring Social Capital was adopted (Grootaert et al., 2003). The answers were collected using a five-point Likert scale between (1) “totally unable”, (2) “unable”, (3) “neither able nor unable”, (4) “able”, and (5) “totally able”.

Socioeconomic status (SES)

The SES score was computed as household income, educational attainment and occupation ranking. Household income was reported as the family’s approximate annual income before taxes and other deductions (Maphupha, 2018). Income was measured with the South African currency Rand (R). Following Maphupha (2018), participants with annual income below R 54344 were coded as “poor”. Participants with income between R 54345 and R 151 727 as “low emerging middle class”, R 151 728 to R 363 930 were coded as “emerging middle class”, R 363 931 to R 631 120 as “realised middle class”, R 631 121 to R 863 906 as “upper middle class” and R 863 906 to R 1 329 844 as “emerging affluent”. Education attainment was assessed as the highest educational level completed. Participants were required to choose one of the options ranging from none, meaning no formal education, to doctorate/postdoctorate (Brauns et al., 2003). Consequently, participants ranked the appropriateness of their current primary occupation concerning their educational attainment. Four options were provided, ranging from “I am not exercising an occupation at present” to “I am occupied above my qualification level”. A total SES score ranging from 3 to 18 was computed as the aggregate of participants’ educational attainment, income and occupational ranking. Higher scores suggest better SES. Scores from 3 to 6 were categorised as “very low”, 7–9 as “low”, 10–12 as “moderate”, 13–15 as “high”, and 16–18 as “very high”. The aggregated SES in this study was done following the SES Index of Lampert et al. (2013).

Data analysis

Data analysis was conducted with SPSS version 27. The assumption of normality of data was performed with skewness and kurtosis analysis (Table 2). The dataset had no substantial non-normality, as skewness values were less than two and kurtosis less than seven (Finney and DiStefano, 2013). Tests for outliers were done with z test. Mahalanobis D² statistics based on the recommendation that the z-score should not be >/= 3.29 for a univariate outlier, and the p value should not be less than 0.001 for a multivariate outlier (Werner, 2003). There was no significant concern for outliers in the dataset of this study.

Table 2 Means, standard deviations and correlations between study variables (N = 1000).

Descriptive statistics were computed for SES, social capital variables, and happiness. A correlation matrix was computed to explore the bivariate associations between components of SES, social capital variables, and happiness. Correlation coefficients were interpreted as small (r = 0.10), medium (r = 0.30) or large (r = 0.50) (Cohen, 2013). Furthermore, hierarchical regression statistics were used to test the hypotheses in this study. Multicollinearity between the predictor variables was examined based on the indices of Tolerance and VIF. VIF values between 5 and 10 indicate highly correlated variables (Daoud, 2017), while a tolerance value >0.4 indicates high tolerance (Allison, 1999). A significant F statistic when the predictor variables are entered into the model indicates a model fit. In addition, a significant change in R² when the interaction of the predictor variables and gender is entered into the regression model. The significant effects of the interaction terms would indicate a moderation effect of gender (Aiken and West, 1991).

Result

The means, standard deviations and correlations between study variables are presented in Table 2. As displayed, SES and all social capital indices positively correlated with happiness except for generalised trust, which was in a negative direction. Age had a significant positive relationship with only SES and generalised trust. At the same time, gender was related only to the locus of control and happiness.

Hierarchical regression analysis was performed to examine the relationship between SES and indicators of social capital (i.e., community solidarity, generalised trust, and locus of control) and happiness (Table 3). The correlations between the predictor variables were low, and the Tolerance and VIF values were all within acceptable limits. In the first model of the regression, SES, community solidarity, trust, locus of control and gender were inserted as predictors of happiness. In the next step, the interaction terms of community solidarity and gender, locus of control and gender and generalised trust and gender were entered as predictors of happiness. The first model shows a model fit [F(5, 994) = 67.58, p < 0.001] and accounted for ~25% (d = 0.33) of the variance in happiness. SES (β = 0.09, p < 0.01), community solidarity (β = 0.27, p < 0.001), and locus of control (β = 0.33, p < 0.001) were positively associated with happiness. In addition, the generalised trust had a significant negative relationship with happiness (β = −0.08, p < 0.01), while being a female was associated with greater happiness (β = −0.09, p < 0.01). Locus of control had the most unique contribution to happiness, explaining 9% of the variance in this variable. The second model was also significant [F(5, 994) = 39.65, p < 0.001] and explained an additional 1% of the variance in happiness [ΔR² = 0.01, ΔF(4, 990) =3.79, p =< 0.05]. This confirms the moderating effect of gender on the relationship between SES, social capital, and happiness. The relationship between SES and happiness was stronger for women (β = −0.09, p < 0.05), community solidarity and happiness were stronger for men (β = 0.08, p < 0.05), locus of control and happiness was not significant (β = −0.03, p = 0.40). Generalised trust and happiness were stronger for men (β = 0.08, p < 0.05).

Table 3 Hierarchical regression model of SES and social capital variables as predictors of happiness with gender as a moderator.

Discussion

The current study examined the relationships between SES and three indices of social capital (i.e. community solidarity, locus of control, and generalised trust) and happiness among a South African adult sample. The result shows that SES, community solidarity, locus of control, and generalised trust explain ~25% of the variance in happiness. The results confirm the associations between higher SES and increased happiness. Higher household income, educational attainment, and occupational stance are associated with higher happiness for the South African sample. This finding supports research that has found SES to contribute to happiness and, in general, the wellbeing of individuals. SES contributes to happiness as it is associated with more social networks, social engagements, financial stability, opportunities for life improvement, and generally better physical health (Muhammad et al., 2022; Rahman and Singh, 2019).

Conversely, poverty and scarcity cause psychological distress and deplete the cognitive resources for self-control and making decisions or plans (Spears, 2011). Furthermore, this finding can be understood within the social context of South African society in which SES represents a key determinant of life outcomes (Botha et al., 2018). South Africa experiences grave social inequality in which a large percentage of its population experience poverty, low education and poor access to basic amenities (Bhorat et al., 2015; Home Office, 2020).

Similarly, the results confirm that better community solidarity and higher locus of control predict increased happiness. In the social relations and wellbeing literature, community solidarity conflates social cohesion, social integration, and social connectedness, thus espousing it as a variable that denotes a feeling of belongingness. Experiencing community solidarity embodies the psychological sense of belongingness and mutual aid. The feeling of belongingness is a basic human need that has been shown to have positive outcomes for health in general (Arslan, 2018; Arslan and Duru, 2017). Living in an environment with acceptance and social support can improve the resources to deal with life challenges (Arslan, 2018; Arslan and Duru, 2017). Community solidarity is also perhaps a significant variable for happiness among the current sample, given the multi-group characteristic of South African society and the animosity between the different groups. The physical, social, and economic divide between the population groups and the inter-group resentment and togetherness still motivate the choice to remain in neighbourhoods dominated by one’s group (Donaldson et al., 2013). In multi-group settings with experiences of discrimination, people identify with their ethnic group, and group cohesion and bonding emerge as a coping mechanism (Bliuc et al., 2020; Greenaway and Cruwys, 2019). Thus, in South Africa, we argue that community solidarity may function as a support framework that compensates for economic disadvantages or boost subjective security that increases happiness.

The result of this study also shows a positive association between locus of control and happiness. In the current sample, the locus of control shows the strongest direct association with happiness. It explains a unique 9% variance in happiness. This finding complements other studies among different populations where a sense of empowerment predicted participants’ wellbeing (Hussain et al., 2020; Moreira et al., 2020). The contributory role of locus of control to happiness emphasises the importance of the perception that one can determine one’s life’s outcomes. Perception of locus of control motivates optimism, hope, hardiness, proactiveness, and the readiness to take actions to change life outcomes or handle life challenges across different life domains such as savings and investment (Salamanca et al., 2020), job search (Caliendo et al., 2015) as well as physical and psychological health (Kesavayuth et al., 2020).

Furthermore, locus of control may be particularly important in developing countries in which life is generally uncertain and in which outcomes may be largely out of one’s control. South African society experiences poverty, inequality, unemployment, crime, and conflicts. Living with such social realities can affect how people perceive themselves to be able to influence or control their life outcomes and, invariably, their happiness (Martončik, 2019). Thus, the contextual uncertainty associated with South Africa due to the characteristics mentioned above may play a role in the importance of locus of control to happiness.

Lastly, as the third indicator of social capital examined in this study, generalised trust is associated with lower happiness. This finding contradicts the expectation that trust constitutes a social capital resource contributing to an individual’s happiness. This relationship may be attributable to the historical social and economic segregation in South Africa that has entrenched intergroup distrust and the contributory role of increasing social inequality among the population groups to this distrust (Breetzke, 2018; Moosa, 2021; Tewolde, 2020). In addition, the presumed inadequacy of the government to stem the increasing tide of inequality may have further extended the distrust from outgroups to the government and its agencies (Levy et al., 2021). In South Africa, trust is limited mainly to one’s own group members. A previous study has suggested that many South Africans perceive that outgroup members act to disadvantage or stunt their group’s growth (Gordon et al., 2012). Clime, where trust is limited within the social group, encourages public distrust, resulting in the nonsignificance of trust as a predictor of happiness.

Gender-based analysis project a stronger association between SES and happiness for women, community solidarity and happiness for men and trust and happiness for men. There was no significant gender difference in the association between locus of control and happiness. Previous studies have also highlighted significant differences in the social predictor of happiness (Gumà et al., 2019; Hori and Kamo, 2018). For example, Hori and Kamo (2018) found gender differences in the determinants of happiness in East Asian countries. Contrary to the result from the current study, they also found that full-time employment was positively associated with men’s happiness in China, and social support was positively associated with individuals’ happiness, especially for women. The gender difference for SES in the South African sample might result from the existing gender gaps in unemployment. More females in South Africa are unemployed. Therefore, having a job as a female might contribute more to happiness for the female than for male participants (Chitiga et al., 2022; Mosomi, 2019).

Furthermore, generally, women disproportionately receive lower wages. While they represent a significant proportion of the workforce, their opportunity for upward mobility is usually incomparable with their skills and training (Livingston et al., 2022; Shook et al., 2020). We argue that South African women may bear the brunt of economic inequality, low education, and unemployment relative to men. As such, money and other indicators of SES may be more critical for women’s happiness. More than men, statistical reports support that women experience low SES status in South Africa. For example, according to Statistics South Africa’s Quarterly Labour Force Survey report of the second quarter of 2022, women (15–24 years) constitute a higher percentage of youth not in employment, education or training (Stats, 2022). Therefore, given women’s socio-economically disadvantaged position in society, achieving higher SES may represent overcoming social barriers to their personal or professional development and, thus, a source of increased happiness. This argument is further supported by the concept that “the more difficult the obstacle, the greater the reward” (Cheng and Zhang, 2018; Palan et al., 2019).

On the other hand, the stronger association of community solidarity with happiness for men may equally be an outcome of gender role differences. The assigned gender role for men is masculinity, characterised by the display of independence, competence, self-reliance, and strength (Julius Ogunleye, 2015). Such a prescription may preclude men from indicating a desire or need for social support to emphasise masculinity (Barbee et al., 1993). Women, compared to men, have a greater inclination for other-centeredness and to receive and give care as part of their gender role socialisation (Julius Ogunleye, 2015) and, as such, may belong to several social support groups or have greater friendship networks (Łapniewska, 2022). Conversely, men are not oriented to ask for support and are less likely to receive support than women (Barbee et al., 1993). The need for a psychological sense of belongingness with a group is a need general to humans irrespective of their biological sex. As such, men having community solidarity, a sense of belongingness, and mutual aid in their communities could positively impact their happiness.

Similarly, the negative relationship between generalised trust and happiness was stronger for men than women. What is deducible from this finding is that given the less other-centeredness for men and gender-role prescription that emphasises competitiveness, autonomy, self-reliance, and emotional control, men may also be less trusting than women. Generalised trust was associated with less happiness in this study, which was attributed to the socio-cultural reality of South African society in terms of diversity, group resentment, and public distrust. Women generally receive trust and trust more and are also more like to restore trust in others even when such trust has been violated compared to men (Haselhuhn et al., 2015). Thus, the gender role difference may also explain why generalised trust is associated with even less happiness for men.

The lack of significant difference between men and women in the contribution that locus of control makes towards happiness further emphasises its importance in achieving happiness among South Africans. People feel more in control of their lives where the resources to mitigate negative life circumstances exist (Martončik, 2019). Living with economic difficulties and other harsh realities of South African society, coupled with the fact that life outcomes may not be entirely within one’s control, may elevate the need to control one’s life. This perhaps explains the more significant effect of this variable on happiness and its equal importance for men and women.

Limitation

Despite the unique contribution of the current analysis to understanding the interaction between SES, social capital, and happiness, the findings should be interpreted within certain conceptual and methodological limitations. The cross-sectional nature of the data limits the generalisation of the result. Similarly, the definition of what is included within social capital remains contested (Kay, 2020). The measure of social capital adapted in this analysis might exclude other components that ultimately impact individual subjective happiness. It is also important to note that the sampling employed for data capturing does not guarantee a representative sample.

Furthermore, the survey questionnaire was available only in English. Considering the multilanguage status of South Africa, this unavailability of the questionnaire in other languages could lead to the exclusion of people with insufficient English language proficiency. Similarly, the current analysis consciously addresses South Africans holistically, ignoring group diversity. A longitudinal study with more inclusive components of social capital addressing the racial and cultural differences in South Africa will provide a more complete picture of how SES and social capital predict happiness.

Conclusion

The current study provides a unique insight into the linkage between SES, social capital, and happiness among South Africans. It further highlights the gender difference in the relationships. It provides policymakers, researchers, and social and health workers with models that facilitate better life outcomes. The findings underscored the need to encourage community projects that promote social participation, enable cooperation, and empower community members to take more control of their lives. Such a project and improved SES would likely contribute to happiness. Furthermore, gender-tailored actions to facilitate social capital and SES components might be more effective in enhancing happiness. In addition, findings contribute to the science aimed at understanding the social determinants of wellbeing and life outcomes.