Volunteering

Volunteering is defined as any unpaid activity in which "…time is given freely to benefit another person, group or organisation" (Wilson, Reference Wilson2000, p. 215). The scientific literature provides distinct explanations for this behaviour (Musick & Wilson, Reference Musick and Wilson2008; Studer & von Schnurbein, Reference Studer and von Schnurbein2013) but there are no clear conclusions regarding a specific volunteer profile (Norris & Inglehart, Reference Norris, Inglehart, O’Neill and Gidengil2006). In addition to the heterogeneity of the population studied, the terminology [volunteer, volunteering, volunteerism] include a wide range of activities that are carried out by numerous organizations with different causes. Wilson and Musick (Reference Wilson and Musick2012) show three main areas of research on volunteering participation as “the antecedents” …” experiences” …and “consequences”.

In this research, we analyze the antecedents of volunteering, which is referred to what influences an individual to select one type of organisation to volunteer instead of another type of organisation. We only examinate the effect of formal volunteering in different organizations in our study because it defines direct activities and can more accurately be measured.

Individual Factors and Volunteering

Nichols and Shepard (Reference Nichols and Shepard2006) have shown that the relationship between voluntary participation and age corresponds to an inverted U: volunteering peaks in middle age and then begins to decline, whereas Wymer (Reference Wymer2011) notes that volunteers are often older. According to life course theories, in middle age, volunteering has to do with family and work roles such as stable jobs and parenting that increase interest in the community (Flanagan & Levine, Reference Flanagan and Levine2010). The values of predisposition to change can influence older adults to volunteer (Ariza-Montes et al., Reference Ariza-Montes, Giorgi, Leal-Rodríguez and Ramírez-Sobrino2017). Motivations and personality traits, such as extraversion, appear to be relevant in deciding to volunteer (Kwok et al., Reference Kwok, Chui and Wong2013). Other factors such as the sense of belonging to the community are decisive in being a volunteer among the elderly (Pozzi et al., Reference Pozzi, Marta, Marzana, Gozzoli and Ruggieri2014). However, the factors that influence how seniors choose to volunteer are not conclusive, more research is needed to strengthen their participation (Boccacin & Lombi, Reference Boccacin and Lombi2018; Gil-Lacruz et al., Reference Gil-Lacruz, Saz-Gil and Gil-Lacruz2019).

The study on whether men or women are more likely to volunteer has yielded contradictory results (Downward et al., Reference Downward, Hallmann and Rasciute2020; Taniguchi, Reference Taniguchi2006; Wilson & Musick, Reference Wilson and Musick2012). Data from Bureau of Labor Statistics (2009) suggest more women are involved in activities than men in US, while in Sweden, there are more male volunteers than female (Musick & Wilson, Reference Musick and Wilson2008). These results vary depending on the type of activity. Some categories have strong gender norms, while in other areas gender differences would not be relevant (Fyall & Gazley, Reference Fyall and Gazley2015). A recent European study has found a positive relationship in the decision to volunteer in awareness and social justice activities for women, while in professional activities for men (Gil-Lacruz et al., Reference Gil-Lacruz, Marcuello and Saz-Gil2018). This may be due to social expectations about what activities are appropriate for women and men (see West & Zimmerman, Reference West and Zimmerman1987, for a further review of "doing gender").

Gender differences in volunteering could be explained by gender roles (Eagly, Reference Eagly1987). Due to the socially expected role of men as breadwinners, unemployment inhibits men's volunteering as they devote their time to seeking paid employment to fulfil their role as primary breadwinner (Taniguchi, Reference Taniguchi2006). In contrast, unemployed women may find participation in volunteer organizations more acceptable, as long as their family members are well cared for (Taniguchi, Reference Taniguchi2006). However, women devote more hours to household chores than men and provide more informal support (Hook, Reference Hook2004), so these additional responsibilities may hinder women's access to volunteering and their time to volunteer. Albeit that the above results focus on middle-aged people, it is unknown whether this pattern is perpetuated among older people.

Given that socialization in many countries has changed over time in progress toward equality and occurs at a young age, Wemlinger and Berlan (Reference Wemlinger and Berlan2016) report that there are age differences between women and men who volunteer. On the one hand, gender equality has a great impact on young people who feel less constrained by traditional gender roles than precedent generations. On the other hand, progress in equality has less impact on seniors as they have experienced a greater degree of gender socialization than the youngest.

Previous literature has highlighted that socio-economic status and educational level are positively correlated with volunteering (Gil-Lacruz, et al., Reference Gil-Lacruz, Saz-Gil and Gil-Lacruz2019). In addition to individual factors, other contextual factors could be influencing how gender differences are reflected in senior volunteering (Fyall & Gazley, Reference Fyall and Gazley2015; Gil-Lacruz et al., Reference Gil-Lacruz, Marcuello and Saz-Gil2018). We are interested in considering how social factors related to gender equality and welfare systems influence volunteering by older people.

Macrosocial Factors and Volunteering

Despite the marked tendency to focus on consequences of volunteering in elderly, several theorists have considered macrosocial factors that could be key to understanding senior volunteering. For instance, some authors have recently shown that contextual factors, such as the government expenditures on health may reduce volunteer rates, whereas government expenditures on other social issues could reinforce them (Gil-Lacruz et al., Reference Gil-Lacruz, Marcuello and Saz-Gil2018, Reference Gil-Lacruz, Saz-Gil and Gil-Lacruz2019). Veal and Nichols (Reference Veal and Nichols2017) showed that higher levels of volunteering in European countries are associated with lower levels of income inequality. Researchers have reported that other macro-social factors, such as the COVID-19 pandemic, could be positively related to senior volunteering (Chan et al., Reference Chan, Chui, Cheung, Lum and Lu2021). However, the results for test of other structural factors in relation to age and gender are less convincing (Phillips, Reference Phillips2015).

Feminist scholars (Rotolo & Wilson, Reference Rotolo and Wilson2007; Wemlinger & Berlan, Reference Wemlinger and Berlan2016) have maintained that the main cause cause of the gender gap in employment are due to macro-social norms related to gender and the resulting organisational segregation (Fontanella, et al., Reference Fontanella, Sarra and Di Zio2020). Few studies have empirically examined the relation between gender-related macrosocial indicators and volunteering (Fyall & Gazley, Reference Fyall and Gazley2015; Wellinger & Berlan, Reference Wemlinger and Berlan2016). Following the volunteering process model (Snyder & Omoto, Reference Snyder, Omoto, Spacapan and Oskamp1992; Wilson & Musick, Reference Wilson and Musick2012), in relation to antecedents, we focused on the social context through two kinds of factors: (a) one ideological factor belonging to the macro system related to norms, beliefs, and values about the gender roles in society; and (b) two structural indicators belonging to be macro system related to gender equality and countries. Specifically, we examined one gender-related ideological macrosocial factor (attitudes toward gender equality) and two structural macrosocial factors (indicator of gender equality and welfare systems).

In this sense, the United Nations (2018) points out the importance of the concept of gender equality referred to equal rights and opportunities (in work, education, access to resources, etc.) to achieve sustainable development. Despite the apparent relevance of attitudes toward gender equality in work highlighted by international organizations, such as the United Nations (2018), we have not found many studies that examine their role in volunteering and from a cross-cultural point of view. Wemlinger and Berlan (Reference Wemlinger and Berlan2016) have found that less acceptance of traditional gender roles facilitates and promotes women’s advancement in volunteering. For instance, increasing women's volunteerism in organisations commonly associated with men such as sports and politics (Boje et al., Reference Boje, Hermansen, Møberg, Henriksen, Strømsnes and Svedberg2019). Therefore, attitudes which reject gender roles could be relevant in the position of men and women in society and in non-profit organizations (Shantz et al., Reference Shantz, Banerjee and Lamb2019).

In the current study, we analyzed the relation between volunteering in different organizations and a structural gender-related macrosocial factor using Gender Inequality Index [GII]. The GII indicator “measured the gender gaps in three aspects of human development—reproductive health, empowerment and the labor market—and exposes differences between men and women in the distribution of achievements” on a scale from 0 (full equality) to 1 (total inequality)” (United Nations Development Program, 2005/14). There is empirical evidence for the relation between structural indicators of gender equality and volunteering (Fyall & Gazley, Reference Fyall and Gazley2015). For instance, the occupational status of women in a given country has been linked to volunteering, specifically, in professional and social justice activities (Gil-Lacruz & Saz-Gil, Reference Gil-Lacruz, Marcuello and Saz-Gil2018).

Previous literature has explored a wide variety of individual characteristics to explain cross-country differences in volunteering ratios (Gil-Lacruz et al., Reference Gil-Lacruz, Marcuello and Saz-Gil2018, Reference Gil-Lacruz, Saz-Gil and Gil-Lacruz2019; Wilson, Reference Wilson2000). Meanwhile a high female participation at age 55 + is found in Latvia, Greece, Poland and Spain, a high male participation at age 55 + is found in Belgium, the Netherlands, France and Germany (Boccacin & Lombi, Reference Boccacin and Lombi2018). In relation to activities, in countries with gender inequality, there is segregation in the type of organizations in which men and women participate (Wemlinger & Berlan, Reference Wemlinger and Berlan2016), with men volunteering more in instrumental organizations, while women volunteer more in expressive organizations. Plagnol and Huppert (Reference Plagnol and Huppert2009) concluded that differences in volunteering ratios between countries could not be fully explained by differences in individual social, psychological or cultural factors associated with volunteering. Macrostructural explanations for cross-country variations in volunteering are therefore necessary to understand national patterns of individual behaviour.

In this sense, Salamon and Anheier's (Reference Salamon and Anheier1998) theory of social origins points out that welfare systems could represent a relevant factor in volunteering. Different institutional regimes are characterized by different patterns of redistribution and social security, qualities of civil liberties, pluralism and rule of law (Enjolras, Reference Enjolras2021). Therefore, welfare systems cannot be reduced only to the product of the unilateral extension of a single factor, such as income inequality, size of the non-profit sector, diversity, education or industrialization (Esping-Andersen, Reference Esping-Andersen1990), but they represent broad social structures comprising complex interrelationships between social institutions and cultural, social and economic factors.

In relation, a data analysis conducted in the Volonteurope report in 2012 compared senior volunteering ratios among European countries. Meanwhile in Sweden, with a strong welfare system, elderly are more likely to volunteer in expressive activities (culture and advocacy groups), in the UK, with a government with a weaker role in the provision of welfare-related services, the participation of the elderly in care associations is more prevalent. Consequently, each welfare system generates different sets of capabilities that, in turn, encourage or inhibit the participation in volunteer activities (Enjolras, Reference Enjolras2021).

The Current Study

We explore the role of individual and macrosocial factors related to gender and welfare systems in senior volunteering. Specifically, we had three objectives: (a) to evaluate whether the ideological and structural factors related to gender inequality are significantly associated with senior volunteering, (b) to explore the extent to which these macrosocial factors interact with gender in predicting volunteering and (c) to analyze whether welfare systems influence volunteering. Given only a few researchers have analyzed the impact of these variables on different types of volunteering, we conducted separate tests of models for general volunteering, social awareness, professional and political, education and leisure, and religion volunteering.

We used two data sources: (1) a database from World Values Survey (WVS, waves 2005/09 and 2010/14), which was used to obtain the dependent variables (volunteering and categories), the ideological variable (attitudes toward gender equality), and individual predictors (age, gender, income level and education level); (2) the Gender Inequality Index (GII; United Nations Development Program; 2005/14), which we used to obtain data at the country level.

We hypothesized (Hypothesis 1) that the selected individual variables would be significantly related to volunteering: Income level and education level would have a positive influence on volunteering. Moreover, we hypothesized (Hypothesis 2) that at the macro level, the frequency of volunteering would be higher in countries with more attitudes toward gender equality (Hypothesis 2a), in countries with lower GII (Hypothesis 2b) and in countries with a Nordic welfare system (Hypothesis 2c).

We explored whether older women volunteer more than men in general and whether there are differences between organisations. Given that the evidence on age and gender is mixed, we decided not to form specific hypotheses on this. Also, we explored whether women would volunteer more in countries with higher attitudes towards equality and GII. In relation, we explored the differences between the two measures of equality between welfare systems. Finally, because research examining multilevel models specific to different types of volunteering is not extensive, no clear evidence exists on which to form separate hypothesis for each type of volunteering (social awareness, professional and political, education and leisure, and religion), and we considered these aspects of the study exploratory.

Data

The empirical estimation utilised data from the World Values Survey (WVS); and from the United Nations Development Program (UNDP). The WVS is a cross-national survey conducted since 1990. The fourth and fifth waves have been utilized due to the availability of the variables referring to volunteering behaviour, attitudes towards gender equality and sociodemographic variables. It is a robust instrument for the analysis of volunteering. The study concentrated on European countries as they are homogeneous enough to make inferences from the results. The macrodata on Human Development from UNDP provided a gender equality index for the countries studied in the time period.

The sample comprised 8234 people aged 50–80, from 8 countries (Cyprus, Germany, the Netherlands, Poland, Romania, Slovenia, Spain and Sweden), for two-time waves (2005–2009 and 2010–2014).

The WVS looks at participation in 8 voluntary organisations. This paper analyzes volunteering at two aggregation levels: volunteering and type or category of activity. About type of categories of volunteering, we selected Sardinha (Reference Sardinha2011) model: (1) Social Awareness (environmental, humanitarian or charitable organisations); (2) Professional and Political (professional organisations, political parties and trade unions); (3) Education and Leisure (sports and recreation clubs, arts, music and educational organisations); (4) Religion (religious organisations or churches).

The explanatory demographic variables were gender and ageFootnote ¹ (between 50–64 and 65–80); the socioeconomic variables included educational level and income.

Two explanatory variables analyzed gender equality: one is subjective ideological and the other is objective structural. As subjective ideological variable, attitudes towards gender equality were obtained through items from the WVS (Men make better political leaders than women do; University is more important for a boy than for a girl; Men make better business executives than women do; Being a housewife just as fulfilling). As objective structural variable, the Gender Inequality Index (GII) was obtained from the United Nations Development Program (0–1). The higher the value, the greater is the gender inequality. Since the subjective measure was an equality index, the objective index was made positive, in order to facilitate the interpretation of the results (ObjectiveGenderEquality = 1 − ObjectiveGenderInequality).

Countries were grouped by welfare systems (Gil-Lacruz & Marcuello, Reference Gil-Lacruz and Marcuello2013; Sardinha, Reference Sardinha2011): (1) Nordic—Sweden; (2) Continental—Germany, the Netherlands; (3) Mediterranean—Cyprus, Spain; (4) East—Poland, Romania and Slovenia. There were 8 geographical dummy variables for each country and 4 for each welfare system. There were also 2 dummy variables for each wave (2005–2009 and 2010–2014). The set of dummy variables allows the calculation of the geographic and temporal effects.

Empirical Framework

Due to the binary nature of the volunteer dependent variables (1 active; 0 inactive) and the hierarchical organisation of data levels (individuals, national and welfare systems), the empirical study was based on multilevel Logit models (STATA: xtmelogit). Multilevel regression models serve a single dependent variable at the lowest level of disaggregation (individual) and incorporate a set of explanatory variables for each level (micro and macro). The study focused on volunteering as a dependent variable, and the micro variables (age, sex, educational level, and income) and macro variables (attitudes toward gender equality, GII and welfare system) which might explain volunteering. The results measured the sensitivity of the dependent variable when there is change in an explanatory variable. The parameters shown in the tables indicate the statistical significance, the sense and intensity of the effects of the independent variables.

The estimation analysis takes into account a non-linear response model; the data were structured for 8234 people ( $i=1,\dots ,8234$ ) corresponding to 8 European countries ( $j=1,\dots ,8$ ). The probability of volunteering ( ${\text{Unpaidwork}}_{\mathrm{ij}}$ ) was estimated as

(1) ${\text{Unpaidwork}}_{\mathrm{ij}}=X_{\mathrm{ij}}^{{}^{\prime }}{\beta }_j+u_j+e_{\mathrm{ij}}$

where $X$ includes $K$ regressors ( $K-1$ variables and a constant) and the term error is characterized by $e_{\mathrm{ij}}\approx N\left(0,{\sigma }^2\right)$ . The model incorporates as many fixed effects ( $\beta$ ) as there are random effects ( $u$ ). The fixed effects indicate how the explanatory variables affect the decision to volunteer; the random effects include the variability between countries.

The estimates were repeated for each of the four categories ( $f=1$ social awareness; $f=2$ professional and political; $f=3$ education and leisure; $f=4$ religion). The estimates were made independently: the same person can participate in different volunteering activities, so the categories are not exclusive:

(2) ${\text{Unpaidwork}}_{\mathrm{fij}}=X_{\mathrm{ij}}^{{}^{\prime }}{\beta }_{\mathrm{fj}}+u_{\mathrm{fj}}+e_{\mathrm{fij}}$

The objective was to achieve the most accurate estimate of β. Estimates used 3 models: Model 1 included sociodemographic data (age, gender, educational level, and income) and temporal variables (2005–2014); Model 2 incorporated information on gender equality with a subjective variable (Attitude towards gender equality) and an objective variable (GII). The incorporation of the subjective and objective allows an examination of the egalitarian attitude of a population and the country's objective level of gender equality, something that could be decisive in explaining a greater or lesser occupational segregation in volunteering decisions. As explanatory determinants, two further interactions were considered: gender and subjective and objective equality measures. The interactions between the variables allow conclusions to be drawn on the relationship between the gender equality measures and the volunteering decisions of men and women. Model 3 added information on welfare systems (Nordic, Continental, Mediterranean and Eastern). The fact that the estimates were stable (in estimated value and significance) was a guarantee of robustness. Model 3 was justified with the results obtained by the ANOVA, which analyzes the dispersion of random effects according to groups of countries classified by welfare systems. The dispersion of random effects was greater between groups than intra-group, so the welfare system represents an adequate country classification model.

Estimates were carried out for men and women to consider differences between two subsamples. If the β estimation of an explanatory variable was different (in sense and/or intensity), there was evidence of differential gender effects.