Motivations for Volunteerism, Satisfaction, and Emotional Exhaustion: The Moderating E ﬀ ect of Volunteers’ Age

: This investigation aims to explore the moderating role of volunteers’ age in the relation between motivations for volunteering and, respectively, satisfaction with volunteerism and emotional exhaustion. A longitudinal study was conducted with a sample of 241 Spanish healthcare volunteers. Results show that volunteers’ age moderates the relations between social motivations and satisfaction, and social motivations and volunteers’ emotional exhaustion, and also between growth motivations and satisfaction, and volunteers’ emotional exhaustion. The relationships between security motivations and satisfaction and emotional exhaustion are not moderated by age. Our ﬁndings underline that, for younger volunteers, satisfaction decreases when social motives are high, rather than low, and, in the opposite, emotional exhaustion increases when growth motives are high, rather than low. For older volunteers, instead, the only signiﬁcant e ﬀ ect concern satisfaction, which is higher when social motives are high, rather than low.


Motivations for Volunteerism and Outcomes: The Influence of Volunteers' Age
Studies on motivation in general and its applications to work are abundant, whereas specific works on motivation for volunteerism are less numerous, and their results indicate a need to expand the empirical research [15]. Different theories have been proposed to explain why people spend time and effort on volunteer tasks, such as the models of Omoto and Snyder [9] and of Clary and colleagues [14]. Omoto and Snyder [9] investigated the volunteer process by taking into account psychological, behavioural as well as social and organizational features of the antecedents, the actual experiences and the consequences of volunteerism. Clary et al. [14], instead, adopted a functionalist approach, according to which people can, and do, perform the same actions in the service of different psychological functions, or to fulfil different motivations. According to the functionalist approach, diverse classifications of motivational work orientations have been proposed and one of them, useful also for volunteerism, distinguishes between motivations for personal growth and motivations for affiliation [16]. Motivations for personal growth assess the importance, or preference, for task characteristics related to achievement and mastery. Social or affiliative motives value the preference for membership and collaboration with others, either co-workers or supervisors. Lastly, various authors added motivations for security, including measures to assess the importance or preference for satisfaction of material and physiological needs related to general well-being, such as material rewards [17]. The functionalist approach has proposed that persuasive messages that match important motivations succeed in motivating individuals to initiate their volunteers' service, and that volunteers whose motivational interests are better served by their participation will continue their activity and will be more satisfied than volunteers whose motivations are not, or only partially, met [18].
In the last decade, some studies have proposed that people's motivations are flexible and vary as a function of age, such that some motivations increase and others decrease along the time [19]. Specifically, even though there is little empirical research in the work setting [17,20] and even less in relation to volunteerism [21], there seems to be a general tendency for social motivations to increase with age whereas growth and security motivations decrease as volunteers grow older. Despite evidence suggesting that older and younger volunteers might prioritize different motivations, the possible Sustainability 2019, 11, 4477 3 of 16 differential effect of age on the relation between motivations and volunteerism outcomes is not yet sufficiently justified.
However, recent advances in aging theories suggest that healthy aging is characterized by an optimization of person-job fit. Following Beier and Kanfer " . . . an age-related shift from achievement-related to emotionally-relevant goals and a shift in priorities to opportunities that will optimize worker fit . . . " can be observed [22] (p. 106). Based on socio-emotional selectivity theory [23] and selective optimization and compensation [24], during aging, people tend to maximize social and emotional gains along with reducing their losses. As empirical evidence indicates, these developmental changes imply improvements in emotion regulation [25], selective attention to positive instead of negative information [26], and increased resilience to stressors [27]. At the same time, adult people try to engage in activities that are adapted to their values and needs. While they move towards activities that better fit their skills and personal traits during the lifespan process, they also increase their positive attitudes about the activities that they usually perform.
In other words, to increase their well-being along the life-span, individuals learn and develop better strategies in order to match their personal characteristics to those of the environment, considered as their organizations' and groups' requirements and practices. The person-environment fit perspective [13] states that individuals' congruence with their work environments leads them to better job performance and to rate that environment as more satisfying. When person-environment fit deteriorates, individuals begin seriously considering alternatives, such as quitting their actual job and searching for another one. Fit has been conceptualized as a function of environmental levels such as occupation, organization, job, and group [28,29] but, despite existing inter-cultural differences, the extended meta-analysis by Oh and colleagues clearly showed "that fit happens and has significant employee and organizational consequences" [30] (p. 135), especially in terms of job performance, job satisfaction or intention to quit.
In this regard, Westerman and Yamamura [31] showed that differences in perceived fit between younger and older workers significantly predicted different work outcomes. Moreover, fit related to growth motives (as autonomy task orientation or expectations for success) influenced job satisfaction and intention to remain with the organization for younger but not for older employees, whereas fit related to social motives (cohesion and good relationships with co-workers and managers) was predictive of job satisfaction for older but not for younger workers. Hence, evidence suggests that age-related differences in the person-environment fit would differentially affect work outcomes, such as job satisfaction and emotional exhaustion, but there is a lack of research on this topic with volunteers. This knowledge is important because age could take a different function, or role, according to the functionalist approach (that would predict that younger and older volunteers may have different motivations for volunteering and, consequently, may experience the same level of satisfaction, or emotional exhaustion, but for different reasons) or a life-span approach (that would predict a better adaptation and an higher performance and satisfaction in older volunteers in comparison to younger ones). In the present study, therefore, it is predicted that volunteers' age will moderate the relations between motivations and the outcomes-satisfaction and emotional exhaustion.

Satisfaction with Volunteerism
One of the most studied variables in volunteerism is volunteers' satisfaction. In general, it has been considered that volunteers' motivations are good predictors of their satisfaction, and that satisfaction, in turn, explains the intention to remain and the dedication to the task. Finkelstein [7] evaluated the influence that volunteers' motivations and the fulfilment of such motivations have on satisfaction. In her investigation, she found that values, knowledge, improvement of self-esteem, and social aspects, and especially the fulfilment of such functions, are the best predictors of satisfaction. Other studies examined in depth the relation between participation in volunteerism and volunteers' satisfaction according to, for example, the number of hours devoted to the task and future intentions to remain [6,7,9,[32][33][34]. The study of Okun and Schultz [21] found that younger volunteers have lower levels of social motivations, probably because they already invest heavily in their own social, family, and occupational networks, whereas older volunteers, who fear losing these networks when they retire, try to offset that decrease through volunteerism. In addition, as mentioned above, Westerman and Yamamura [31] observed that growth motives influence job satisfaction and intention to remain with the organization for younger but not for older employees, whereas social motives predict job satisfaction for older but not for younger workers. Thus, although prior investigation is scarce, age seems to play a moderating role in the relations between motivation and satisfaction, even if we cannot predict the direction of the difference. Accordingly, the following hypotheses are proposed: Hypothesis 1a. Volunteers' age (T1) moderates the relation among social motivations (T1) and satisfaction with volunteerism (T2). Hypothesis 1b. Volunteers' age (T1) moderates the relation among growth motivations (T1) and satisfaction with volunteerism (T2).

Healthcare Volunteers' Emotional Exhaustion
Although much research has been carried out on volunteer satisfaction, emotional exhaustion of volunteers has attracted much less interest and has been analysed by few studies [35][36][37], although its influence on distress and poor health has been well documented [12,[38][39][40].
A study investigating the relation between motivations for volunteerism and emotional exhaustion examined four main motivations (the desire to offer altruistic help to others, the desire to conform to social norms regarding volunteer activity, the desire for personal development through volunteer activity and the desire to fill free time) and found that only women with altruistic motivations showed low levels of emotional exhaustion [11]. In contrast, in the same study, all the four motivations for volunteerism were positively related to emotional exhaustion in men. However, only few works have examined the differences in volunteers' emotional exhaustion as a function of age, as for instance the one by Okray and Abatay [41] that observed that the emotional exhaustion of volunteer nurses was negatively related to age.
In summary, in the present study, we explore the moderating role of volunteers' age in the relations among motivations for volunteerism and emotional exhaustion, by positing that: Hypothesis 2a. Volunteers' age (T1) moderates the relation among social motivations (T1) and volunteers' emotional exhaustion (T2).

Participants
The present study used a longitudinal design, with two moments of data collection by a paper and pencil survey, with an interval of approximately six months. At Time 1 (T1), data collection took place in September-November of 2013, and at Time 2 (T2), data were collected during the months of March-May of 2014. The participants in this study were 241 healthcare volunteers with a mean age of

Instruments
Motivations for volunteerism. This was assessed with the Volunteer Functions Inventory of Clary and colleagues [14], adapted to Spanish language by Dávila and Chacón [42]. The inventory consists of 30 items that assess the reasons why people are dedicated to volunteerism with a Likert-type response scale, ranging between 1 (strongly disagree) and 5 (strongly agree). The factor structure of the scale did not seem stable, as previous studies provided diverse solutions, with six [42] or fewer factors and with diverse factor distributions of the items [32]. Therefore, we carried out an exploratory factor analysis, which yielded a three-factor solution that explained 57.9% of the variance. The first factor (growth motives) included the items that loaded on the subscales of knowledge and improving one's curriculum in the original factor structure, and it explained 22.9% of the variance. The second factor (social motives) grouped the items of the social relations subscale, explaining 18.8% of the variance; and the third factor (security motives) included items from the scale of ego-defence and mood enhancement, explaining 16.2% of the variance.
Satisfaction with volunteerism. This variable was assessed with the scale of general satisfaction with the organization where the volunteer worked, designed by Vecina, Chacón, and Sueiro [34] from prior studies [8]. The scale consists of seven items, one alluding to general satisfaction and the rest to various aspects of the volunteer's relationship with the institution. The reliability of the original scale was good (α = 0.88) and, in our study, it reached a value of α = 0.87.
Emotional exhaustion. This variable was assessed with the subscale of psychological exhaustion of the Spanish Burnout Inventory [43]. The complete subscale consists of four items in which the term "work" was replaced with "volunteer work", following the methodology of prior studies [44]. Adequate psychometric properties for the subscale have been firmly established [45]. The reliability of the measure employed in the present study was adequate, reaching α = 0.83.
Sociodemographic variables. Volunteers' chronological age was assessed by asking them the date of their birthday.

Procedure
To recruit the participants, the Department of Social Psychology of the National University for Distance Teaching (UNED) proposed a research project to the centres in charge of organizing the Municipal Social Volunteerism of the town halls in the southern outskirts of Madrid. Letters were addressed to the centres explaining the reasons for the study and the characteristics of the data collection procedure. Ten organizations that answered the first written communications were personally contacted; only five of them agreed to collaborate and facilitated access to their members. This recruiting procedure resulted in a convenience sample, based on how easy it was for our research group to get in touch with the potential participants. This non-probability sampling method is used when there are time and cost constraints in collecting feedback.
The participants were assembled at the headquarters of the organizations, and the research team presented the conditions of the study: goals, contents, anonymity, voluntariness, and the possibility of leaving the study at any time. Questionnaires at Time 1 were handed out to attendees who agreed to participate, to be completed in their homes and returned to the research team in a closed urn placed for this purpose in the headquarter of the organization.
Four hundred questionnaires were distributed, trying to collect data from all the potential participants, with a sampling error of +/− 2.06 (95% CI). At Time 1, 316 responses were collected (79% response rate). The participants created an individual code that would have allowed researchers to match the questionnaires of participants responding at T1 and T2. At T2, the questionnaires were sent to the headquarters of the organizations, in closed envelopes with prepaid postage material to return the responses to the University. The collaborating organizations informed their members that the second part of the investigation was available, and they sent them a new massive reminder two weeks later.
The response rate at T1 was high, probably due to the active involvement of the organizations, but at T2, there was no possibility to individually contact the participants, that instead were informed through massive posts in the headquarters of the organizations. This may have contributed to the decrease of the response rate in the second stage of the longitudinal study. We finally received 252 completed questionnaires at T2 (63% response rate), of which 11 were eliminated because more than 30% of the data was incomplete. The final sample of the longitudinal research was made up of 241 participants.

Data Analysis
In the present work, we carried out descriptive and correlational analyses of the variables under study. Next, in order to test the hypotheses, we carried out multiple linear regression analysis with the PROCESS macro for SPSS [46], which allows appraising the moderating effect of age in the relationship among the volunteers' motivations and their outcomes of satisfaction with volunteerism and emotional exhaustion with the task. To determine whether age moderated the relation between predictors and criteria, we used the indicators of the regression model, non-standardized coefficients (b), standard errors, levels of probability associated with the contrast statistics, and, especially the confidence intervals (CI) of the bootstrapping of 5000 samples (with a 95% CI), for the standardized effect of the predictor on the criterion at the different levels of the moderator.

Results
The descriptive statistics and the correlation matrix of the variables of the study are shown in Table 1. The correlation matrix reveals that age was significantly, and negatively, related only to social motivations. With regard to the outcomes, age was negatively related to emotional exhaustion and had no significant relation with satisfaction. However, social motivations was positively related with emotional exhaustion, whereas growth motivations showed a positive relation with satisfaction.  Table 2 reports results concerning Hypotheses 1. Results show that Hypothesis 1a is supported by the data: satisfaction with volunteerism is significantly predicted by age, social motivations and by their interaction term. In addition, the moderation of age shows that the effect of social motives on satisfaction is negative and statistically significant for the younger volunteers (mean age of 30 years), but it is nonsignificant for the older volunteers (mean age of 55 years). As shown in Table 3, in the younger volunteers, the CI did not contain zero and in the older volunteers, it did. The results of the moderation of age in the relation between social motives and satisfaction with volunteerism are depicted in Figure 1.
Results concerning Hypothesis 1b confirm the moderation of age in the relation between growth motives and satisfaction, given that probability associated both with the predictors and with the interaction term were significant, and the increase of variance explained by the interaction, although small in absolute value, was also statistically significant, as shown in the second column of Table 2. The values of the effect of growth motives on satisfaction vary as a function of volunteers' age, being nonsignificant for the young volunteers but positive and statistically significant for the older ones, as can be seen in Table 3, in the lower rows. This effect is depicted in Figure 2.   Results concerning Hypothesis 1b confirm the moderation of age in the relation between growth motives and satisfaction, given that probability associated both with the predictors and with the interaction term were significant, and the increase of variance explained by the interaction, although small in absolute value, was also statistically significant, as shown in the second column of Table 2. The values of the effect of growth motives on satisfaction vary as a function of volunteers' age, being nonsignificant for the young volunteers but positive and statistically significant for the older ones, as can be seen in Table 3, in the lower rows. This effect is depicted in Figure 2.
Results completely disprove Hypothesis 1c, given that neither the interaction nor the increase of variance associated with the interaction were significant, as shown in the third column of Table 2. In relation with the second block of hypotheses of this study, it can be seen in Table 4 that the two first hypotheses were confirmed by the data, whereas the third hypothesis has to be rejected. The probability associated with the interaction of age in the relationship between social motives and emotional exhaustion was statistically significant (Hypothesis 2a), as it was for the moderation of age in the relationship between growth motives and emotional exhaustion (Hypothesis 2b). A nonsignificant relation is observed in the moderation of age in the relationship between volunteers' security motives and emotional exhaustion (Hypothesis 2c). Volunteers' age (T1) moderates the relation between social motives for participation (T1) and their emotional exhaustion (T2) such that, as can be seen in Table 5, the effect of social motives on Results completely disprove Hypothesis 1c, given that neither the interaction nor the increase of variance associated with the interaction were significant, as shown in the third column of Table 2.
In relation with the second block of hypotheses of this study, it can be seen in Table 4 that the two first hypotheses were confirmed by the data, whereas the third hypothesis has to be rejected. The probability associated with the interaction of age in the relationship between social motives and emotional exhaustion was statistically significant (Hypothesis 2a), as it was for the moderation of age in the relationship between growth motives and emotional exhaustion (Hypothesis 2b). A non-significant relation is observed in the moderation of age in the relationship between volunteers' security motives and emotional exhaustion (Hypothesis 2c).
Volunteers' age (T1) moderates the relation between social motives for participation (T1) and their emotional exhaustion (T2) such that, as can be seen in Table 5, the effect of social motives on emotional exhaustion is large and statistically significant in young volunteers but it is not significant in older volunteers. Volunteers' age (T1) moderates the relation between growth motives for participation (T1) and their emotional exhaustion (T2) because, as seen in Table 5, the effect of growth motivation on exhaustion is positive and statistically significant for young participants but it is not significant in older volunteers.   Table 4 also shows that Hypothesis 2c is rejected because neither the main predictors (security motives and age), nor their interaction term reach statistical significance. The statistically significant effects of moderation related to Hypotheses 2a and 2b are represented graphically in Figures 3 and 4.

Discussion
The present study had the goal of exploring the moderator role of volunteers' ages in the relationships between their motivations for volunteerism and the outcomes-satisfaction with volunteerism and emotional exhaustion-in a group of healthcare volunteers aged between 26 and

Discussion
The present study had the goal of exploring the moderator role of volunteers' ages in the relationships between their motivations for volunteerism and the outcomes-satisfaction with volunteerism and emotional exhaustion-in a group of healthcare volunteers aged between 26 and 64 years.
The results allow us to affirm, firstly, that the influence of social and growth motivations on volunteers' satisfaction is negative and statistically significant for the younger volunteers, whereas it is positive and loses significance for the older ones. As suggested by Beier and Kanfer [22], one of the terms of the person-environment fit equation are personal characteristics and motivations, which are affected by age-related changes. Our findings would support that changes in motives for participation affect perceptions of fit differently for younger and older volunteers, influencing their satisfaction. Secondly, the influence of social and growth motivations on emotional exhaustion is high and statistically significant for the younger volunteers but it is nonsignificant for the older ones. Thirdly, the relations between security motivations and satisfaction and emotional exhaustion are not moderated by volunteers' age. Accordingly, it is important to note that the hypotheses related to the influence of social and growth motivations are confirmed, whereas those related to security motivations are rejected. These findings contribute to a growing body of evidence about the existence of emotional exhaustion in volunteers [47].
The motivations for volunteerism have been investigated for some time, although many approaches have preferentially focused on altruism [48]. The present findings, which confirm the existence of a variety of motivations for volunteerism, also contribute to the debate related to the classification of these motivations. The most recent investigations with international volunteers propose the existence of two motivational patterns, one oriented inwards and the other outwards, which could be applied to volunteerism in a broad sense, and which seem to correlate with volunteer satisfaction [49]. Other international studies have also supported the existence of two motivational patterns, one implicit and the other explicit, which are linked to desirable behaviours and have been shown to be stable cross-culturally [50].
The finding that different motivations may have different impacts on the outcomes is consistent with other recent studies that show that volunteers who have self-determined motivation, but not control motivations, have higher levels of health and less burnout than non-volunteers [51]. Likewise, Stukas et al. [52] found that the differences in well-being of Australian volunteers were related to motives oriented towards others or towards oneself.
The moderating role of age deserves a more detailed comment, as older volunteers with greater social and growth motivations express more satisfaction, whereas the opposite pattern is observed in younger volunteers. In contrast, when social and growth motivations are higher, the older volunteers show less emotional exhaustion. As recent studies suggest, the relevance of social motives for Generation Zs seems less influencing on their attitudes towards volunteering [53]. These findings seem to support the idea that social and growth motivations both have a more pronounced impact on the outcomes for older than for younger people, accordingly to recent research on the impact of stress on youth volunteers [54].
This evidence has diverse implications. In view of the theory of socio-emotional selectivity and of the motivational approach of Kanfer and Ackerman [19], this adds a piece of novel evidence, as all prior studies focused on paid employees and did not explore such differences in the volunteerism context. These results also enhance the debate on the existence of other variables, as motivations to volunteer, that influence emotional exhaustion and satisfaction, and that influence them differentially as a function of age. On the one hand, it has been suggested that involvement in volunteerism can be a substitute for the loss of roles in retirement [55], and this variable would affect the late adults or older people. On the other hand, it is possible that the balance between life and work may affect outcomes such as satisfaction, and this variable would influence younger people more than late adults and older people [56]. In the same vein, other variables could be included in future studies, due to the preliminary evidence that suggest its relevance, as emotional intelligence [57], communication practices [58], organizational pride and trust [59], volunteer management practices [60], among others.

Limitations and Future Lines of Research
Some limitations of the study must be considered. Even though we considered a broad array of motivations for volunteerism, such a set of motivations still represents a selection. Other motivations could be significant and should be included in future studies. Especially, ego-defensive motivations and escaping from loneliness might have a strong influence on individual outcomes and, at the same time, be moderated by age.
Moreover, although we think that our measures of volunteers' satisfaction and emotional exhaustion adequately assessed volunteers' attitudes, such measures give a global representation of the emotional experience of volunteers regardless of the activities in which they are involved, more or less regularly. Thus, future studies could use other methodologies, such as the Happiness Test or the Short-Day Reconstruction [61], in order to better consider the complexities of the positive and negative experiences of volunteers associated with the performance of volunteerism tasks.
In this study, we considered only healthcare volunteers, however our results should be tested in volunteers' associations that operate in different fields, such as the environmental or political ones.
Another limitation of the study is that data were obtained from a non-random sample. We cannot exclude that our respondents belong to a self-selected group of volunteers that has different characteristics from volunteers in different centres. Although our findings confirm previous studies, our results should be treated with caution.
On another hand, although our results show that the temporal prospects associated with age could influence the relation between motivations and outcomes, we have no measures of the relevance of the interviewees' age, so this conclusion seems to be based on indirect reasoning rather than on direct evidence. In future studies, volunteers could be asked directly if their age and the associated changes are important for them in relation to their volunteer activities, and to include this information as moderator variables in the analyses [62].
Despite these limitations and the unresolved issues, the present study contributes to the growing body of literature on volunteers' motivations and attitudes. It also extends this literature, adding the evidence of a longitudinal study that relates volunteers' motivations and attitudes with their age. This study has strengths and weaknesses. Among its main weaknesses, we must point out the self-reported nature of the data and the measurement of all the variables by means of questionnaires. A novel aspect is the confirmation of the hypotheses on the different impact of motivation as a function of volunteers' age. Another strength is the longitudinal design which, separating along time the independent and the dependent variables, allows stronger conclusions about the potential effects of motivations to participate on satisfaction for volunteering and the emotional exhaustion related to volunteer activities.

Implications for Practitioners
Motivation of volunteers and consequences resulting from their volunteerism are important for the effective management of volunteers [63]. In this sense, educational programs for non-profit administrators on recruitment, selection and training of volunteers should be beneficial. In particular, it seems that communication concerning social and growth motivations should be present but not be too much emphasized for young people, because high levels of such motivations seems related to the decrease of satisfaction and the increase of emotional exhaustion; thus, unfortunately, our results suggest what not to do with younger people than what to do. More clear results concern the late adults and older people, which seems to be mainly motivated by growth needs, aiming to increase skills or facing challenging activities that have to be mastered. However, inquiry into the strategies which would be more effective to incorporate volunteers into the daily life may help to design successful entry programs. In the same vein, education programs at the universities could help to introduce volunteer students, as suggested [64].
Regarding the differences in motivations of volunteers as a function of their age, it must be considered that most non-volunteers are of working age, and this information should be considered in planning volunteers' activities in all contexts. More project-type activities could be offered that might be more motivating for a wide range of citizens [65].
This line of investigation serves to inform future interventions aimed at increasing volunteers' satisfaction and reducing their emotional exhaustion. Specifically, it may enhance the design of volunteer management strategies, differentiated as a function of volunteers' ages and their most relevant motivations. This will contribute to better performances of volunteers, and to more effective achievements.