When joining is not enough: emergency services volunteers and the intention to remain

Abstract

Volunteer turnover is a continuous concern for emergency services organisations. Using a psychological contract perspective, this paper examines how meeting the initial expectations of volunteers when they join an emergency service is related to the volunteer’s intention to remain with that service. A survey was undertaken by 539 emergency services volunteers in Western Australia after volunteering for about a year. The survey included questions about reasons why volunteers joined the service, activities they expected to undertake as a volunteer, their evaluation of the volunteering experiences in the first year and their intent to stay. Using Latent Class Analysis of their expectations, three types of volunteers were identified. These were ‘focused’ volunteers (having well-defined expectations), ‘lost’ volunteers (lacking clear expectations except for serving the community) and ‘overenthusiastic’ volunteers (expecting status, reputation and career progression among other things). The results showed that the focused volunteers had participated in more activities and had a higher intent to remain a volunteer. Having too few or too many expectations seems to have negative consequences. Therefore, emergency services organisations could strive to understand and shape volunteer expectations to match a new volunteer’s experience by using better aligned recruitment practices.

Introduction

Many communities in Australia rely on the crucial services provided by emergency services volunteers, especially in the states and territories that are relatively less densely populated. In Western Australia, the Department of Fire and Emergency Services (DFES) currently supports over 26,000 volunteers across six services, including the Bush Fire Service and State Emergency Service (DFES 2018). However, these volunteer emergency services organisations can be exposed to high levels of volunteer turnover, with annual rates as high as 23 per cent. This turnover rate is problematic because it potentially puts the delivery of crucial emergency services in jeopardy. This leads to a loss of expended finances, time and human resources. McLennan (2004) estimated at that time it cost $710 to recruit and train a new emergency services volunteer in Australia.

The reasons for the high rates of volunteer turnover are not well understood. Locke, Ellis and Smith (2003) conducted a comprehensive review of the literature on volunteer retention and found many areas of disagreement among researchers about the factors that cause volunteers to stay. Unfortunately, despite additional research conducted since that review was published, the literature on volunteer retention remains limited (Newton, Becker & Bell 2014) especially with respect to the retention of emergency services volunteers.

One study found that intentions to remain a volunteer firefighter were positively associated with better leadership and higher inclusiveness within the volunteer brigade, group or unit (McLennan et al. 2009). Studies of State Emergency Service volunteers found aspects of support and recognition to be strong predictors of volunteer job satisfaction (Fallon & Rice 2015). Moreover, supervisor support, recognition, interactional justice and group cohesion were positively related to ongoing commitment to volunteer (Rice & Fallon 2011). A comprehensive report on emergency services volunteers in Australia identified five main challenges to volunteer retention. These were:

• limited time available for volunteering
• excessive amount of training required
• personal costs involved in volunteering
• lack of recognition for volunteering
• conflicts and tensions with other volunteers (Esmond 2009).

Studies into emergency services volunteer turnover are often limited in the reliance on data collected from exit surveys. Opinions expressed in exit surveys are prone to being plagued by hindsight bias and may provide an overly negative account. In addition, exit surveys do not provide insights into volunteers’ decisions to stay with the service. Existing research provides little theoretical guidance as to how to identify volunteers who are at risk of leaving the service, nor the initiatives to improve volunteer experiences with the service.

This paper builds on the theoretical framework of the ‘psychological contract’ (Rousseau 1995) to explore how emergency services volunteer expectations shape first-year experiences of volunteering and influence intentions to remain with the service. An innovative data analytic technique—Latent Class Analysis (LCA)—was used to identify distinct profiles or ‘types’ of volunteers based on their reasons to join and their expectations of the volunteering role. Grounded in the psychological contract perspective, an explanation is provided of how mismatched volunteer expectations may lead to higher turnover. Practical guidance for managers to identify volunteers at risk of leaving and ways to retain these volunteers in the service is offered.

Literature review

Extant research conducted with emergency services volunteers in Australia offers suggestions as to why volunteers leave. A study of volunteer firefighters in South Australia and Victoria suggests that dissatisfaction with the volunteer role accounts for about 25 per cent of those withdrawing from volunteering (McLennan et al. 2009). But why do volunteers become dissatisfied with the role? If volunteer initial expectations of the role and the purpose of the organisation do not match their actual experiences, they may become disenchanted and, thus leave the service.

In a volunteering context, the psychological contract refers to the obligations, rights and rewards that a volunteer believes he or she is owed in return for continued work and loyalty to a manager, group, team or organisation (Rousseau 1995). In other words, a psychological contract is what a volunteer expects the volunteering experience to be like, based on what was promised by a manager, group, team or organisation.

In contrast to a traditional written contract, the psychological contract:

• is implicit rather than explicit
• can be shaped by experiences the volunteer has before and during the volunteering experience
• can be unique for every volunteer rather than standardised for all volunteers (Stirling, Kilpatrick & Orpin 2011).

When a volunteer perceives a psychological contract breach, they may feel that trust has been violated and may respond with aggressiveness, negative behaviour and/or leave the service (Vantilborgh 2015). For example, Vantilborgh and colleagues (2012) found that helping other volunteers was perceived as an integral part of the volunteering role, although this task was not expected when joining an organisation. Consequently, some volunteers did not want to undertake this additional task, presumably as a way of reconciling a psychological contract breach.

It is important for organisations managing volunteers to understand that the ‘terms’ of a volunteer’s psychological contract can be formed even before the organisation has interacted with that volunteer. For example, strong stereotypes about a volunteering role (e.g. firefighters rescuing people from burning buildings) might create expectations, accurate or otherwise, about the volunteering experience and what needs or goals might be satisfied by that experience. Recruitment messages and activities will also impact on the expectations volunteers have about a volunteering role.

After a volunteer is recruited, their psychological contract is influenced during ‘socialisation’ experiences. Effective socialisation transforms a volunteer from being an ‘outsider’ to an ‘insider’. Socialisation is the process through which ‘an individual acquires the social knowledge and skills necessary to assume an organisational role’ (Van Maanen & Schein 1979, p. 211). Socialisation can include induction, onboarding (i.e. the process of integrating volunteers into the team) and training. During these activities, psychological contracts can be solidified or transformed. There is also a risk of psychological contract breach if these early activities do not meet the expectations a volunteer had before or during recruitment.

A psychological contract will continue to evolve as a volunteer becomes familiar with a manager, group or organisation. However, it is perhaps the pre-recruitment, recruitment and socialisation experiences that will determine most strongly a volunteer’s intent to remain with an organisation. Therefore, to reduce premature turnover, it is firstly important to understand what volunteer experiences are in his or her early tenure and how these experiences match (or fail to match) the volunteer’s expectations. This study addresses this issue by identifying the expectation profiles of first-year emergency services volunteers and examines how these relate to intentions to remain with a service.

Method

In partnership with the DFES, data from 539 volunteers who completed a survey of their first-year experiences was analysed. The survey was distributed in 2015-2017 to all new volunteers with approximately 12 months of service. A total of 4535 volunteers were invited to complete the survey and 555 valid responses were received (response rate of 12 per cent). However, 16 responses were missing data on all key variables, thus the final sample used in the analyses was 539.

The survey was designed by DFES staff and requested demographic and service-related information. Respondents were asked to select reasons why they had joined the service (e.g. ‘Help the community’) and the activities they expected to be involved in (e.g. training). Figure 1 shows a list of reasons and activities. Respondents further indicated whether they have undertaken the training and activities expected, how many hours a month they volunteered with the service and how long they intended to stay. Finally, volunteers were asked to agree or disagree with a number of statements about their experiences with that service and its leadership (e.g. ‘I feel included and part of the team’ or ‘Experienced volunteers and leadership keep to themselves’).

The LCA statistical technique was used to analyse the data. In general terms, LCA identifies different subgroups (profiles) of people within a larger population (Muthén & Muthén 1998-2012). In concrete terms, the LCA was used to determine whether different types or profiles of emergency services volunteers existed based on the reported reasons for joining and the activities they expected to be involved in.

Three statistical criteria were used to determine how many sub-groups of emergency services volunteers exist. These were:

• the Bayesian Information Criterion (BIC)
• the measure of latent classification accuracy (entropy)
• the Adjusted Lo-Mendell-Rubin likelihood-ratio Test (Adjusted LRT).

The preferred solution should have the lowest BIC value, entropy values closer to one and significant Adjusted LRT statistic (Wang & Wang 2012). Finally, nonparametric Kruskal-Wallis tests were used to compare the identified volunteer profiles on the status of other variables collected in the survey, such as demographics, experiences and intentions.¹

Results

Selected demographic characteristics of the sample are presented in Table 1. Respondents represented all Western Australia regions with the majority (42 per cent) from within the Perth metropolitan area, 20 per cent in the South West, 11 per cent in the Great Southern.² Interestingly, 12 per cent reported volunteering with two services. Overall, these distributions are representative of the volunteer population in terms of demographic characteristics, location and service.

The respondents’ selected reasons to join the service and the expected activities (both coded as zeroes and ones) were used to identify the profiles of first-year emergency services volunteers. Based on these criteria, a three-profile solution was the best fit with the data³ (BIC=11,721, entropy=0.84, Adjusted LRT=314.3, p<0.05).

Figure 1 shows the profiles along with the estimated probability that a volunteer from a given profile will select the corresponding reason to join the service and the activity expected once in the service. The three identified profiles of volunteers are labelled ‘focused’, ‘overenthusiastic’ and ‘lost’. There was no difference in the gender composition for each profile.

The focused volunteers (57 per cent of all respondents, orange bars in Figure 1) reported a fairly clear and focused idea of why they joined the service; to help the community, learn new skills and seek camaraderie. The focused group also had a well-defined set of expectations regarding what their volunteering activities entail, especially with respect to training.

Table 1: Demographic characteristics of respondents.

The overenthusiastic volunteers (13 per cent, maroon bars in Figure 1) mentioned a variety of factors to join the service and were the only group to report status, reputation and career path as reasons to volunteer. This group also expected to be involved in many different activities. A larger proportion of overenthusiastic volunteers were under 36 years of age (54 per cent compared to other profiles (i.e. focused, 41 per cent; lost, 24 per cent; p=.001).

The lost volunteers (30 per cent, blue bars in Figure 1) did not have a clear idea of what influenced them to join, aside from helping the community. Volunteers in this profile also did not have strong expectations regarding what they would do in the role, aside from providing emergency response and general help. A larger proportion of lost volunteers were over 55 years of age (34 per cent) compared to other profiles (i.e. focused, 18 per cent; overenthusiastic, 21 per cent). Further, a significantly larger proportion of lost volunteers were located in regional areas (39 per cent) as compared with metropolitan areas (17 per cent, p=.001).

Figure 1: Estimated probability for citing the different reasons to join the service and activity expected by volunteer profile.

Figure 2 shows the volunteers responses when asked if they had undertaken the expected activities. 10 per cent of lost volunteers reported no activities in their first year. A larger proportion of overenthusiastic volunteers reported being involved in only some expected activities (31 per cent, p=.064).

A larger proportion of lost volunteers’ reported volunteering for ten hours or fewer a month (46 per cent), compared to focused (37 per cent) and overenthusiastic volunteers (22 per cent, p=.059). A similar proportion of focused and overenthusiastic volunteers reported greater than 20 hours a month, on average, dedicated to the service. Only 64 per cent of lost volunteers reported attending a call-out in their first year, compared to 76 per cent of focused and 78 per cent of overenthusiastic volunteers.

On average, focused and overenthusiastic volunteers reported more positive experiences (mean of 3.2 and 4.0, respectively), compared to lost volunteers (mean of 2.1, p=.001). Finally, when asked how long they intended to remain with the service, a smaller proportion of the lost volunteers intended to volunteer in the long term (48 per cent), compared to overenthusiastic volunteers (66 per cent) and focused (59 per cent, p=.055). Figure 3 shows the intention to remain with the service by volunteer profile.

An interesting result was the larger proportion of overenthusiastic volunteers considering leaving the service in the near future or unsure about their length of volunteering (13 per cent), which was comparable to the lost volunteers (12 per cent).

Discussion

The aim of this study was to identify profiles of first-year emergency services volunteers based on their reasons to join and expected activities. The study sought to identify how these expectations influence volunteers’ experiences with the service and their intention to remain.

The three distinct profiles of first-year emergency services volunteers—focused, overenthusiastic and lost—primarily differed in the number of activities volunteers expected to be involved in and the number of motives they expected to satisfy by joining. The focused volunteers had a well-defined set of expectations about the activities they would undertake and the reasons for joining the service. Whereas the lost volunteers had barely any expectations and the enthusiastic volunteers were expecting many things. Having a well-defined set of expectations meant that focused volunteers had better experiences in their first year with the service and were intending to continue volunteering for longer. This is consistent with the psychological contract perspective where realistic initial volunteer expectations (relative to the actual experiences of volunteering) lead to reduced turnover. However, there is a potential danger in that the focused volunteers might be too rigid in their expectations and lack the flexibility required to deal with the unexpected demands of the volunteering role (e.g. Vantilborgh et al. 2012).

Figure 2: Participant responses to ‘Have you undertaken activities that you expected?’

Figure 3: Intention to stay with the service by volunteer profile.

The overenthusiastic volunteers were somewhat scattered in their expectations of the activities and their reasons for joining the service. Although having volunteers who are passionate and enthusiastic can be very beneficial for a service, there is a danger of a psychological contract breach occurring. For example, overenthusiastic volunteers might expect more activities than is actually possible, hence they report lower involvement with volunteering activities in the first year. Ultimately, this might lead to a lower intention to remain with the service for these volunteers, which is consistent with the psychological contract perspective.

Lost volunteers had minimal clarity about their reasons for joining the service and the expected activities. In line with the psychological contract perspective, these initial volunteer expectations influenced their experiences in that lost volunteers were involved in fewer activities, possibly because they did not know what they were expected to do as a volunteer. The findings suggest that the breach of psychological contract may occur when volunteers have too little expectations. This is evident as lost volunteers reported a significantly lower number of positive experiences in their first year and had a lower intention to stay with the service. Of concern was the finding that a large proportion of lost volunteers reside in regional areas, as these areas are potentially more reliant on volunteer services. Research suggests that emergency services volunteers are most frequently recruited informally, via existing social connections (Baxter-Tomkins & Wallace 2009). It is speculated that informal recruitment is especially widespread in regional areas and this could partially explain why a higher prevalence of the lost volunteers are in these areas.

The results of this profiling are consistent with previous research in the general volunteer population in Australia, which identified six psychographic segments of volunteers based on their motivations (Dolnicar & Randle 2007). The overenthusiastic volunteer profile is similar to ‘dedicated volunteers’ who perceive each of the motives for volunteering as relevant.

Practical implications

The findings demonstrate the importance of applying psychological contract perspective in the recruitment and retention of volunteers. Active recruitment is often the first point of contact between a service and a prospective volunteer and interactions during the recruitment process can shape expectations. Emergency services organisations must be mindful that they provide positive, accurate and realistic information about the volunteering opportunity, lest new volunteers be disappointed by a mismatch between their expectations and experiences. For example, anecdotal evidence suggests that emergency services volunteers are often unprepared for the amount of training they need to undertake before becoming operational and do not realise how much idle time there might be between calls to assist. Such misconceptions of the volunteering role could be clarified by designing accurate recruitment materials and educating recruiters and managers in setting the right expectations.

Research suggests that volunteer managers and volunteers often have different expectations and obligations of the volunteering role (Taylor et al. 2006). It is possible that the mismatch between volunteer expectations and actual experiences occurs because managers communicate organisational expectations of volunteers at the recruitment stage rather than the expectations of volunteers of the organisation. Therefore, it is important that managers involve existing volunteers in the recruitment and socialisation stages of new volunteers. Existing volunteers can clarify actual roles and activities and facilitate a better match of volunteer expectations.

Finally, there are significantly more lost volunteers in regional areas. Regional communities depend heavily on the service rendered by emergency services volunteers. Therefore, it is important to employ realistic recruitment messages as well as existing volunteers to elicit and clarify expectations.

Conclusion

This study considered and addressed an important issue of retention of emergency services volunteers through the lens of the psychological contract perspective. Using survey data from 539 first-year volunteers in Western Australia, profiles of volunteers based on their reasons to join the service and expected activities were identified. The three profiles—focused, overenthusiastic and lost— primarily differ in the quantity of expectations. The study showed that having too few or too many expectations may lead to higher volunteer turnover. While the study is somewhat limited due to the small response rate and the use of cross-sectional data, the findings emphasise the importance of employing realistic recruitment messages and clarifying expectations for new volunteers to reduce turnover.

Footnotes 

1. With binary variables of location and attendance at the call-out, the standard t-test was used to establish significance.
2. In the analyses, all respondents from the areas other than Perth metropolitan are referred to as ‘regional’.
3. Fit statistics for the two-profile solution (BIC=11,880, entropy=0.76, Adjusted LRT=870.1, p<0.01). Fit statistics for the four-profile solution (BIC=11,721, entropy=0.79, Adjusted LRT=156.2, p<0.05).