Many retired people engage in volunteer opportunities as a means to give back to their communities, stay physically and intellectually active, and build and expand their social networks. However, our semi-structured interviews of six retirees found that they typically avoid searching for volunteer opportunities through websites and social media due to a lack of trust in those tools and a concern for privacy. Instead, they rely on word-of-mouth communication facilitated through emails with individuals and organizations they trust. To support this type of communication, we designed an adaptive interaction mechanism in the form of a newsletter with volunteer opportunities that are personalized using recommender system technology. The newsletter mechanism leverages personal connections through user-defined preference-based communities that allow users to share volunteer opportunities with their peers.
portunities, relying instead on word-of-mouth
communication, internal email newsletters, and their knowledge
Retirees are a valuable asset for volunteer organizations, of their community’s needs. Secondly, while retirees do
which in turn provide retirees with a means to stay en- select volunteer activities based on personal motivations
gaged and intellectually active [
Our research team conducted semi-structured inter- mechanism that augments existing volunteer opportunity views with six retirees representing four diferent lifestyles email newsletters with automated recommender system to understand their decisions about volunteering. Our capabilities. Our Personal Community Volunteer System interview focused on understanding the key factors that (PCVS) incorporates social recommender system princimotivate retirees to volunteer, how they assimilate into ples to assimilate the word-of-mouth recruitment stratthe roles they choose, and how they entice others in their egy retiree volunteers are familiar with, while providing social circles to join them in volunteering. Our expecta- a balance between community-building capabilities and tions were that the volunteering roles each retiree takes privacy. on would be a highly personal decision, often reflecting their past professional careers. Futhermore, we expected that they would rely on web sources such as Volunteer- 2. Related Work Match, the AARP Volunteering platform, or volunteer recommendations on local websites.
In analyzing our interviews, we found that retirees typically do not use online tools to find volunteer op
our research on retiree volunteerism and the
development of our resulting interaction mechanism. We used
Charmaz’s textbook on grounded theory to guide our
interview study and data analysis [
In probing the user experience in seeking and react- were signed. We conducted individual interviews with six
ing to recommendations, we followed the guidance of subjects, with one or two researchers present. The
interKnijnenburg et al. [
Finally, while we found prior research on the phe- cepts, which informed the development of theories about nomenon of word-of-mouth communication in many how retired individuals learn about and pursue voluncontexts, we found none specific to retiree volunteers. teer opportunities, personalize their roles, and encourage The current paper makes important contributions on this others to volunteer. Our full data repository is available front. upon request.
We employed a lightweight version of Charmaz’s guide- Our initial coding eforts resulted in 287 instances of 124
lines for grounded theory research [
nities, such as the medium used to share information to older adults.
Using contrastive coding practices, we then developed several axial codes based on these groups and categories. The major takeaways from each axial code are presented in the following subsections.
Participating in any activity (even those not related to volunteering) can organically cultivate an interest in volunteering and help retirees discover new volunteer opportunities. For instance, U01 and U02 discovered volunteering opportunities by participating in lifelong learning classes, and were later approached about those opportunities, which they accepted. These individuals now regularly learn about new volunteer opportunities through word-of-mouth because of their involvement in the organization and organized events.
Some organizations take advantage of this organic recruitment efect by actively recruiting new volunteers at events—participants noted that the Lifelong Learning Institute had made eforts to get people to congregate, such as socials and book exchanges. Finding volunteer opportunities is often not a case of retirees outside an organization looking in, but rather current volunteers and staf inside an organization looking outward to find new volunteers. U01, for example, noted that she regularly "makes a plea" for new volunteers at such events, and several interviewees remembered being approached at events about opportunities within the organization. These findings highlight the importance of existing organizational structures in recruiting retiree volunteers.
personal motivations, they are more likely to volunteer.
For example, participant U03, who used to play polo, now volunteers to teach children how to ride horses and play polo. They also volunteer at grade schools to provide art presentations and encourage others to pursue the arts, saying "I guess I’ve used volunteering as more of an escape" and "to give back." U02, a former English teacher, volunteers to teach reading and English, stating "Well, I’ve been a teacher for 42 years, and they were ofering me the opportunity to continue doing what I always loved" and "It’s been very important. It’s changed my life. It’s kept my brain alive." U05 believes that good volunteers "select volunteer opportunities in which that volunteer has a deep interest," with volunteering being where interests meet needs. The degree to which these volunteering interests and motivations are related to the users’ past experiences varies, however. For instance, in the case of U01 and U06, these users chose volunteering 4.3. Organizational Culture roles that were not so closely related to their prior careers. U01 moved from a career in education to a role in Organizational culture significantly influenced retirees’ administration, and U03 moved from being an assisted decision to participate in volunteer opportunities. For inliving administrator to being a leader for the local church stance, U01 stated that when she brings friends to see the and community. In the case of U03, in addition to vol- lifelong learning institute and they meet the organizers unteering for opportunities related to her interests in and volunteers, and experience the general culture, "they horseriding and polo, she is primarily involved with art are sold!" Conversely, some volunteers make substantial events and wishes to get involved with the National Parks contributions to the organizational culture, attracting Service. While interests and motivations may vary from many members and potential recruits. For example, U01 person to person, these findings highlight the importance mentioned volunteers with notable careers, such as a of personalizing volunteer experiences to the individuals’ doctor who fought HIV and distinguished academics, interests and motivations. who deliver interesting lectures and draw in potential volunteers. This means that recruitment eforts not only increase volunteer numbers but also enhance the organizational culture. For example, U01 mentioned that the doctor who fought HIV decided to start an annual re- misinformation as a reason for avoiding social media, cruitment event for the organization, thereby making and stated that she is trying to avoid “getting hooked” a positive impact. Again, these findings emphasize the to email. Most participants expressed a level of trust in importance of existing organizational structures. emails from organizations and individuals they know, but signaled a lack of trust in other resources. 4.4. Required Commitment Overall, interviewees emphasized that word-of-mouth is the key to finding and recruiting volunteers. As U01 How much time, energy, and overall commitment is re- stated, "I have no idea of anybody who finds volunteer quired for an individual to volunteer great impact on jobs via computers. [...] The way you find volunteers is their ability to participate. For example, as U01, U02, through other people, word-of-mouth." and U03 noted, retirees lead very busy lives—many visit family members that live far away, are involved with 4.6. Volunteer Habits Change Over Time organizations around them, or are doing other activities with their retirement time—making regular time com- Lastly, abilities of retirees change as they advance in age, mitments can be dificult for some. Furthermore, some which may afect their desire and ability to participate in volunteers may experience age-related issues that reduce volunteer activities. For example, U06 participated in a the feasibility of participating, with U02 noting the re- horse welfare camp but she felt she was getting too old strictions of arthritic volunteers. Another restriction is for this high-energy commitment, so she terminated her that some retirees eventually return to work in some participation. Additionally, personal motivations may capacity—U04 and U05 both expressed that they would change, and volunteers may decide they want to try somedo more volunteering if their job did not take up so much thing diferent. For example, U03 stated that she would of their time. like to explore new volunteering opportunities and that
Some people have personal motivations that may af- she has always wished to work for the National Parks fect their desire to commit themselves to volunteering Service. Similarly, U06 said she would like to volunteer opportunities. For instance, U01 started a center for help- for the Dollywood Foundation and also would like to ing abused children. This created a high-stress, high- take care of sea turtles. commitment role for herself, but she remained motivated It is in organizations’ best interest to accommodate by her strong commitment to causes that help children. changes in volunteers’ motivations and abilities, as this On the other hand, if the commitment required by an may lead to them leaving a greater impact on an orgaopportunity exceeds one’s threshold, they may move on nization than might have otherwise been possible. For to other volunteering opportunities that better fit their example, U01 mentioned that the doctor who fought abilities. For example, U03 mentioned that their volun- HIV had originally made plans to give a lecture about teer work sometimes felt more like a job than something COVID-19, but eventually decided to be more interested she was passionate about, and she often left events feel- in organizing a variety of fundraising eforts that ended ing dissatisfied. Likewise, U01 and U02 noted that people up drawing large crowds. sometimes “get up to here!” with responsibilities and just say “I’ve done enough,” ultimately quitting their volun- 4.7. A Theory of Retiree Volunteering teering role. These findings emphasize the importance of matching the required commitment with the abilities of the retiree volunteer.
Figure 1 summarizes our findings in a Theory of Retiree Volunteering. Upon retiring, people often turn to volunteer activities as a means of adapting to new life 4.5. Word-Of-Mouth Recruitment circumstances. The process of finding and committing to a volunteer opportunity is a deeply personal process Many of our participants reported rather limited tech- driven by retirees’ activity of seeking relevance. This nology usage, relying on email and word-of-mouth for process crucially depends on retirees connecting with communication more than social media or other web- available volunteer organizations—something that prisites. This makes recruiting volunteers through word-of- marily happens through interpersonal relationships and mouth more efective than other recruitment methods. communications. Additionally, volunteering opportuniFor instance, U03 and her friends learned about volunteer ties must be a good fit for the volunteer (i.e., be related possibilities mostly by word of mouth, and she suggested to their interests, fulfill their goals, and be accessible to that many in her age group do not use online resources them). In many cases, this means personalizing the when looking for opportunities. experience—something that also ascertains that the role
Another key aspect of the retirees’ technology choice remains a good fit for the volunteer, even as their moinvolves the desire to protect their privacy and to avoid tivations, interests, and abilities change. The ultimate scams. For instance, U01 cited specific concerns about goal of this process is contributing to the volunteer We designed an interaction mechanism informed by the ifndings of our interview study (see Figure 2). Our Personal Community Volunteer System (PCVS) builds on existing volunteer opportunity email newsletters with the addition of a personalized social recommender-based organization, not only through the volunteer activities engine. This design reflects the interviewees’ feedback themselves, but also by contributing to the organizational that volunteers find opportunities organically, and that culture and by recruiting others to join the organization. the preferred tool for communicating within their social
From finding opportunities, to participating, to recruit- circle is email. ing others into an organization; every step in the volun- The PCVS shows opportunities sourced from organizateering process is deeply personal for retired volunteers. tions the user has signed up for. In its full implementation, A decision support mechanism must therefore support organizations would register with the service, and users this highly personal process, rather than try to replace it would sign up for organizations in order to receive inforwith a new process that will likely be abandoned. Indeed, mation about volunteering opportunities. Additionally, our results suggest that existing attempts to replace the to allow users to organically learn about new opportuniprocess have failed, as seen by the low prevalence of ties through volunteer social networks, we built in a peer website-based volunteer recruitment systems being used group system (see Figure 3). That is, the second grouping by retirees. of volunteer opportunities is based on preference-based communities: Aided by a social recommender engine that matches users to groups of volunteers with similar 5. Interaction Mechanism Design interests, users can create, join, and add events to these peer groups. The events that have been added to a peer group are included in the newsletter of each person in the peer group. This peer group system enables users to create small networks of volunteers that share opportunities with each other, just like how our interviewees told us they find opportunities. an interaction mechanism to support retirees in these
To help users understand how well an opportunity tfis, activities. The findings of our semi-structured interview we display a description of the opportunity, the location, study show that the process of finding and learning about the number of people signed up, the date/time/frequency, new volunteering opportunities is a deeply personal and and an image of the type of volunteer work for each organic process for retirees: Information about volunopportunity. Lastly, users can join an opportunity by teering opportunities is spread through word-of-mouth, pressing the “sign up” button, which takes them to a and retirees must make sure that the opportunities are— page where they can complete any formal registration and remain—a good fit for their interests, motivations, required for the event. and abilities. Furthermore, we found that older adults do not use the most modern communication methods, preferring email and face-to-face communication over 6. Discussion, Limitations, and social media or websites.
Future work Based on these findings we designed the PCVS: an email-based newsletter system that displays lists of volunIn this paper, we present our user studies to better under- teer opportunities based on existing organization memstand how retirees find, join, and disseminate volunteer berships and user-defined preference-based communiopportunities. Based on our findings, we then developed ties (i.e., peer groups). For the latter, opportunities are displayed in groupings labeled by the preference-based experience. This research was supported in part by the community they are associated with, and a social recom- NSF award IIS 2045153. mender engine helps users find or create new communities, view users within the communities, and recommend events to them. The focus on smaller communities sup- References ports word-of-mouth recruitment while respecting their privacy.
Our research was constrained by a small sample size
and a short time frame of three months to design the
study and complete data collection. Additionally, all of
our interviewees were college-educated with previous
professional careers, most with connections to a local
university. Future work should expand this sample to be
more diverse, particularly in education level, past
employment roles, and geographic location. Such work can also
study the volunteer decision processes of people that are
not retired, to better understand the diferences between
those populations, and how our interaction mechanism
could be generalized to a broader audience. Future
research may explore how system preferences change with
a new generation of retirees. With the rise of ubiquitous
computing, newer retirees with more technological
experience may prefer outlets such as social media platforms
and instant messaging services for finding and sharing
information about volunteering. In the future, we aim
to implement the presented system and conduct a user
experiment [
In this paper we studied what compels retirees to volunteer, how they find and select volunteer opportunities, and how they share opportunities with their personal social network. We used our findings to design an adaptive decision support mechanism to help retirees find volunteer activities that integrates social recommender technologies into existing email newsletters.
Our results are not only useful in helping older adults look for meaningful activities and social groups but may also have broader implications for social decision support systems in general. Importantly, we conjecture that allowing users to craft their own preference-based communities may be a broadly-applicable approach to supporting decision contexts that are heavily guided by close, interpersonal relationships. Recommender technologies can further support the creation of such communities and catalyze the word-of-mouth recommendation process.