A systematic literature review of reciprocity in engineering service-learning/community engagement

3.1 Literature search strategy

To establish our dataset, first we developed search terms based on (i) our research questions within engineering education and (ii) keywords related to SLCE and reciprocity. Our intent in using these as search terms was to capture the breadth of literature connected to reciprocity. We used Boolean operators to incorporate synonyms and link major terms. The complete search string used was as follows: (engineering education) AND (partnership OR collaboration) AND (“service learning” OR “service-learning” OR “community engag*” OR “civic engag*”) AND (university OR college OR “higher education” OR “post-secondary” OR postsecondary).

To identify the most appropriate databases to conduct our search, we consulted a subject librarian in engineering education and selected the following: (i) Educational Resources Information Center (EBSCO), (ii) Academic Search Complete, (iii) Education Full Text (H.W. Wilson), and (iv) Education Research Complete.

After searching each of these databases, an initial review suggested that we were not capturing conference proceedings. Because conference proceedings and other “gray literature” can be a valuable source of information in engineering education systematic reviews (Borrego et al., 2014), we completed an additional ERIC database search using the publisher limiter “ASEE” to yield relevant results from the annual meeting of the American Society for Engineering Education. Because the majority of ASEE proceedings are evidence-based practice papers, by including these papers we enriched and expanded our dataset for analyzing the form and extent to which reciprocity is present in practice.

Our initial database search retrieved 187 relevant records. After duplicates were removed, 142 articles remained.

3.2 Inclusion criteria

Our database search criteria included scholarly, peer-reviewed journal articles and ASEE conference proceedings published between 2000 and 2021. We chose our search period (i.e., 2000–2021) to align with the emergence of engineering education as a field (Froyd & Lohmann, 2014). This window of time also corresponds to a phase of increasing interest in engineering SLCE, and includes publications produced both before and after the Engineering Projects in Community Service (EPICS) program was established (EPICS; Coyle et al., 2005). EPICS has been a highly influential engineering service-learning model throughout the last 15 years, and scholarship on its programming has been disseminated widely to a variety of colleges and universities.

To further tighten the scope of our results for our research questions, we required that all articles explicitly include an engineering context rather than a broad inclusion of the STEM fields. Our inclusion criteria also targeted higher education settings, so we excluded articles that focused exclusively on K–12 programs. Because we are interested in reciprocity as it manifests in SLCE, in our review, selection criteria required that an article have one of the following components: service-learning, community-based learning, or a community-engaged, or university–community partnership. In addition, we chose articles where these components follow one or more of the integration models described by Oakes (2004): integration into existing courses; integration with co-curricular components; new SLCE courses; or SLCE programs. Because prior work (e.g., Salam et al., 2019) shows that SLCE frequently includes cross-national experiences as well as domestic partnerships (and that benefits differ across the contexts), we included both domestic and international SLCE. Table 1 summarizes our inclusion criteria.

3.3 Selection process

To complete the selection process, articles were randomly assigned to teams of reviewers who then screened each by title and abstract by applying the inclusion criteria. Each abstract was evaluated independently by at least two reviewers. If consensus between the initial two reviewers could not be achieved, a third reviewer evaluated the abstract and applied screening criteria to determine inclusion. A total of 65 papers were retained for evaluation for the next stage of the literature review.

After the initial abstract screening, the full text review of each of the 65 remaining papers was then coded independently by two reviewers. From this process, papers were eliminated because criteria for inclusion were not met: SLCE was absent (n = 11) and engineering concepts were absent (n = 2). Our final stage of the selection process left a total of 52 articles for full study inclusion. Details of the search, screening, and selection process are provided in Figure 1.

3.4 Limitations

There are three main limitations to the methods used. First, we acknowledge that our search terms might not have been broad enough to elicit articles describing the full range and variety of ways that institutions and communities may partner and connect. Various orientations to university–community relationships are reflected in the terminology used in publications. We attempted to capture these diverse approaches through our search terms but, in our focus on SLCE in engineering and higher education, it is possible that we have unintentionally excluded some relevant studies.

Additionally, we narrowed our examination of the literature to engineering-specific efforts. Based on our research questions, a broader consideration of STEM-related SLCE efforts was not included in this review. It is possible that a wider examination of the field of STEM would provide a more comprehensive view of the elements of reciprocity that are currently in practice, but because many of these likely include particular disciplines within STEM (e.g., biology or chemistry), we believe that it would skew our ability to make compelling inferences about engineering, specifically.

4.1 Phase 1: Coding

In this section, we present the development and piloting of our codebook, as well as the implementation of deductive codes to analyze the gathered articles.

4.2 Phase 2: Developing emergent analytical framework for reciprocity

This section presents the analysis of the collaborative memos for each framework element and the levels of reciprocity practice that surfaced from the analysis.

4.2.2 Early stages of analytical framework development

Our analysis in Phase 2 focused on examining the collaborative memos from the lens of understanding how the codes informed the form and extent of reciprocity in engineering SLCE (addressing RQ2). This analysis revealed new perspectives on the wide range of reciprocity practices, and we found that leveraging the deductive codes alongside these levels of practice could function as a framework for understanding, evaluating, and informing the manifestation and extent of reciprocity.

For example, we initially derived the element of “context” based on the centrality of time, place, and the people involved in an SLCE partnership: “reciprocity cannot be separated from individuals, families, communities, and generations or the time or place that provide it context and influence it” (Dostilio et al., 2012; p. 24). The salience of context—who, what, when, and where—highlights circumstances influencing how the students, university, and community enter and maintain relationships. However, through further analysis of the memos, we found that context also informs and connects to all other elements of reciprocity because it shapes stakeholder engagement, relationships, and project outcomes.

During the Phase 2 analysis discussions, we identified that the extent to which reciprocity elements were present in an article was indicative of a level of reciprocity. This observation took shape into a gauge for allowing us to interpret whether a partnership was exploitative or an exemplary demonstration of reciprocity. Table 4 presents the levels that were extracted from the discussions on the Phase 2 analysis, along with a ranking scale that would be used to understand the levels of reciprocity present in the articles we reviewed.

The collaborative memos for each code were evaluated using the ranking system in Table 4, independently by two authors. The authors began by analyzing two memos and then met to check for reliability of their approach to assigning levels of reciprocity. Any text that indicated harmful or negative implications, such as downplaying the community outcomes, in the cell was coded as −1. Any text that either was blank (no data) or mentioned that not enough information was included that would indicate reciprocity was coded as 0. Text that suggested reciprocity was present and/or the community was involved, would benefit, or was impacted/changed to any extent at all, even minimal, was coded as 1. Clear language of reciprocity (shared, involved, collaborative, etc.) was coded as 2. Lastly, in cells where there was an exemplary level of reciprocity with respect to the other papers/framework elements, we coded as 3. In cases where the two coders disagreed, they met to discuss each cell of disagreement, and negotiated a consensus when possible. This ultimately led to an interrater reliability score of 99.2%, where agreement was not possible on only 5 of the total 676 cells. The outcome of this analysis is illustrated in Figure 2 and interpreted in Section 6.

5.1 Findings: “Context” code

Through our analysis of the articles, data coded within context was defined by the conditions of the SLCE effort, including the people, places, time, and environment. We further delineated context into three primary subcodes to align with the SLCE stakeholders: student context describes the backgrounds and circumstances of the students; university context describes the characteristics of the university and relevant units within it that concern the SLCE efforts; community context describes the environment in which the SLCE effort is carried out and the circumstances of the associated community partners.

Across the reviewed articles, we found that there were three main characteristics used by authors when articulating contexts: (i) balance of focus across stakeholders; (ii) depth of description; and (iii) thematic emphasis on history, organizational operations, sites, or people.

Within balance of focus across stakeholders, four categories emerged across the articles (Table 5): (i) an institutional context-focused study (prioritized the university, the university students, and their evolution); (ii) student outcomes and classroom-focused; (iii) community solution/product-focused; (iv) long-term context development/integration-focused.

Studies where the balance of focus prioritized the institutional context focused on how the SLCE effort benefited the university, such as the alignment and justification of the course and outcomes with ABET accreditation. Participating universities were often large (and frequently land-grant) public institutions. Descriptions of instructors were often limited to describing the instructors' perspective, their role/lectures, and overall course goals. The curriculum was described alongside a history of SLCE at the university (focusing on justification for the course and procurement of financing). Few studies were part of ongoing courses (those that were, were typically associated with the EPICS program), and these more often had ethics outcomes and the trajectory of the classes described. Again, the course context was typically not a required core course but rather a cornerstone or capstone design course and/or elective. Sustainability of the course was discussed here, as the course may have been designed to meet institutional initiatives (e.g., connecting with an industry/company partner for corporate benefits), the university mission (e.g., Jesuit tradition), and organizational membership. The context referenced the leveraging of institutional commitments, resources, and connections. For long-term SLCE course maintenance, institutional commitments were present (and perhaps required).

Studies with a balance of focus placed on the student context were often nested in the university, where most examples primarily focused on the context of the SLCE course learners. Student leadership was limited (typically not leading the design or decision making of the course). Some were given specific responsibilities (e.g., within a budget), milestones, and requirements for communication with community partners. Assessments for students typically focused on social orientation and satisfaction, but they gave limited descriptions of the students' demographics (e.g., reporting gender, but not race, ethnicity, or ability status). Students often had limited technical experience, and they could take the class to get some practical technical knowledge. Students were often new to service learning, seeing it as an opportunity to volunteer. The student context was isomorphic with the institutional context, in other words, parallel to the aims and behaviors of the institutional context.

The balance of focus placed on a community context was reported least frequently of the four categories that emerged, even at times when university and student context were described extensively. The community context was typically introduced with a superficial description of the people but a detailed geographic description. Often these context descriptions were from a deficit perspective, positioning the community as needs-focused, targeting rural poverty infrastructural issues (e.g., food–water–energy nexus). Some studies distinguished the wider community and local partner non-governmental organizations (NGOs). The NGOs were often already working with the communities, which sometimes separated direct communication between the university/students and the community partners. However, when deeply embedded and committed to the community, the NGOs also facilitated reciprocity among all partners. A few of the studies were situated within a domestic context working with community partners in the United States. We found that, more frequently, the SLCE partnerships were cross-national engagements, which can afford the benefits of building global competence and broadening university students' social, political, and cultural understandings of engineering work (Salam et al., 2019). However, this kind of engagement can also position the community partners as “other,” flattening understandings of complex social and historical factors, such that students can easily identify inequities in other cultures, but not in their own.

The additional two characteristics used to describe the contexts across articles include the depth of description and the thematic emphasis on history, organizational operations, site, or people. Overall, nearly every article had some description of the context of their work, although a handful focused more on describing the nature of the partnership rather than the partners' contexts. The context description in most articles often situated the SLCE in space, geography, time, goals, and values.

The relationship between the various stakeholder contexts—student, university, community—pointed to how a whole study could be characterized: following Dostilio et al., contexts described in the studies as either generative or characterized by a combination of exchange and influence. Examples of coding for context that were supportive of reciprocity include always mentioning the community site and providing background history and efforts to develop sustainability for the SLCE effort and reciprocal relationship. Elements that were not supportive of reciprocity include superficial descriptions of the community, focus on institutional history/mission, and emphasis on university students. Such elements were “necessarily” often a deficit of focus, seeing reciprocity as a clear exchange. Generative reciprocity contexts included descriptions of all relevant resources and motivations across all three stakeholder groups that are very detailed. Participants from the university as well as the community are described in extensive and human detail, explaining their goals and motivations. In contrast, exchange/influence reciprocity relations are often aligned with “service” orientations, and the U.S. institutions (whether the student or university units) are described in operational detail. The community is described superficially and as “in need.”

5.2 Findings: “Research methods” code

The reviewed articles articulated “research methods” in ways that led us to develop the following definition: the systematic steps documenting the SLCE effort and the ways of monitoring of success/outcomes/impact. This definition agrees with broad understandings of appropriate program evaluation, learning, and reporting (e.g., USAID, 2022). We recognize that methodological choices in manuscripts can vary significantly, and therefore our analysis focused on the ways authors described their methods, without limiting to a particular methodological approach. In our review, three relevant characteristics emerged: (i) the type of method, (ii) focus of the methods, and (iii) the rigor/quality of the application of the methods.

The most prevalent category of methods types was “no description of research methods”, where 12 articles did not explicitly share how they documented, monitored, or made inference about the SLCE effort (see below for further discussion of gaps in rigor). In the next largest category, 10 articles used a multi-/mixed methods approach, utilizing multiple strategies for gathering information. Scholarship of Teaching and Learning (SOTL) papers and surveys were, respectively, each identified in six articles; methods used but not explicit enough to categorize, as well as qualitative or interview methods were used in five studies; and case study methods were used in four. Given the range of methodological choices we observed in SLCE publications, we discuss in our concluding section the implications for strengthening the investigative approaches within the SLCE space.

Within the focus of the methods category, the unit of analysis within the reviewed articles was most often centered on students, followed closely by the community. We see therefore that the subjects of SLCE in engineering research are most often university students. Further, when a community focus was included as a unit of analysis, these articles still often treated the community like a datapoint, rather than as contributors within a partnership. Very few articles focused on the experiences of the instructors. Some papers focused on describing the structures of the actual SLCE partnership itself (Davis et al., 2014; Kant et al., 2014; Thompson & Jesiek, 2014).

Finally, there was also a range of rigor/quality within the reviewed articles. As we were not evaluating the specific methodological choices themselves, we assessed rigor broadly on the common guidelines for educational research (National Science Foundation, 2013), namely, generating evidence about opportunities for learning or knowledge production. For example, some articles were very rigorous and reported exemplary use of educational research methods (Arcipowski et al., 2017; Stein & Schmalzbauer, 2012). On the other hand, a handful of papers demonstrated little to no methodological quality and/or did not address methods in the article. Papers lacking methodological rigor may have lacked alignment between evidence and design choices or drawn conclusions that involved leaps in reasoning or overinterpretations. Research methods were one of the major gaps identified by our review; the type of method was not always explained, nor was the focus or unit of analysis always explicitly articulated.

Incorporating reciprocity as part of methodological rigor is crucial for equitable SLCE; participatory methods align with reciprocity by involving community partners in the research. The descriptions of a community's participation can indicate whether they were viewed as a data point or full partner. However, SOTL papers frequently do not represent reciprocity because they focus mainly on student teaching and learning, which is oriented toward the university. Therefore, partnership should be considered as part of the methods consideration, including in ethics reviews and peer reviews.

5.3 Findings: “Approach” code

The “approach” across the articles can be defined as the steps taken by the institution and/or the community partner to initiate the collaboration, the process followed to build the partnership, and the milestones on the way to achieve the outcomes intended from the institution–community engagement. Among the articles we reviewed, we captured three main aspects of the approach taken by the three different stakeholders (faculty, students, and community partners) to engage in engineering SLCE activities: (i) informal and formal integration of SLCE into the curriculum; (ii) multiple stakeholder involvement in various stages of the SLCE effort; and (iii) efforts taken by higher education institutions to build international SLCE partnerships as part of a major university initiative or mission.

Most of the reviewed papers focused on SLCE efforts formally integrated into the curriculum. However, the approaches taken to integrate SLCE experiences into undergraduate engineering education differed in terms of the target audience and the types of models for integration. Engineering SLCE courses were offered as one-credit courses (often offered to first-year students), project-based learning elective courses, summer study abroad programs, and senior capstone projects (e.g., Abrahamse et al., 2015; El-Gabry, 2018; Onal et al., 2017; Seay et al., 2016). Most of the SLCE experiences were offered as multidisciplinary courses to students from various disciplines in engineering and technology. Some of the articles reported SLCE experiences integrated as informal extra-curricular activities led by non-profit/student organizations such as Engineering for Sustainable Development and Engineers Without Borders (Chisolm et al., 2014; Florman et al., 2009; Stein & Schmalzbauer, 2012).

Another category that emerged via review of the approach to SLCE was the involvement of multiple stakeholders at various stages of the SLCE efforts such as problem identification, design, development, testing of solutions, and final implementation within the community. Most of the partnerships were initiated by the university, while some were initiated by the community partners themselves. There were also a set of articles that were initiated by nonprofit organizations who acted as intermediate facilitators between the university and community. Few articles showcased a comprehensive multi-stakeholder, participatory approach where the faculty, students, and members from the community interacted and jointly collaborated throughout different stages of the SLCE effort (Aslam et al., 2014; Dunkel et al., 2011; Kang & Chang, 2019; Reynolds, 2019; Trott et al., 2020).

The last trend observed was the international nature of many of the SLCE efforts, where many included an international component where the problems to be addressed were identified from different countries, especially from lower income nations (Abrahamse et al., 2015; Budny & Gradoville, 2011; Reynolds, 2019; Seay et al., 2016; Simon et al., 2012). Some articles recognized the importance of building cultural awareness among students before visiting international community partners and organized capacity-building sessions to bridge these gaps. For example, Simon et al. (2012) described a direct correlation and impact of improving cultural awareness and constant communication with international partners to the partners' level of ownership and buy-in to the SLCE efforts. Many articles described the logistics and challenges involved in organizing international SLCE efforts where most of the community interaction occurred as part of summer study abroad or a short immersion experience.

Our analysis observed that the extent of reciprocity can be examined via the approach and the extent of stakeholder involvement at various stages of the SLCE effort. Approaches were considered least reciprocal when the students had no interaction with the community partners and worked to solve pre-identified problems provided to them by the faculty. Here, the SLCE efforts were completed within the boundaries of the university and therefore had limited elements of reciprocity within the entire experience. The extent of reciprocity was observed to be more involved in articles where the problems being addressed were identified and provided to the university by the community partner. In this type of approach, the community partner may have had greater, but still limited, involvement in the entire process. Further reciprocity was showcased in SLCE efforts where students actively engaged with community members, contributing to the development of meaningful relationships between the stakeholders. The most thorough form of reciprocity was highlighted in articles that reported following a systematic and/or participatory approach during all the various stages of the SLCE effort. Within these approaches, there appeared to be attempts to build a long-term partnership with the community through frequent face-to-face interactions that can allow the development of trust, respect, and meaningful relationships between the stakeholders.

5.4 Findings: “Impacts” code

Impacts, defined as the outcomes achieved for each of the stakeholders involved in the SLCE partnership, were observed to cluster across three sets of characteristics, similar to the three sets for context. As with the context code, these three areas of impacts—student, university, and community—were further delineated to align with the three stakeholders: student impacts describe the outcomes and benefits directed at students; university impacts describe the impacts associated with the university and relevant units within it that concern the SLCE; community impacts describe the benefits and outcomes directed at community partners. The analysis of the impacts explored the implicit and explicit outcomes of the SLCE efforts described by the authors and resulted in three trends across the reviewed papers: (i) impacts most frequently centered students; (ii) university and student impacts were most often the driving factors behind developing SLCE partnerships and experiences; (iii) community impacts were often implicit in the reviewed papers and were generated from the perspective of the authors.

The first trend we observed was that, in describing outcomes of the SLCE effort, impacts most frequently centered the outcomes achieved by the students. The primary areas of impact on students were the technical and professional skills that were developed. In the reviewed papers, impact was evaluated using techniques such as pre and post surveys (e.g., Budny & Gradoville, 2011) and guided reflections (e.g., Florman et al., 2009), which were most often used for the purpose of assessing student learning outcomes. Moreover, the evaluation techniques predominately focused on learning outcomes associated with technical skills (i.e., problem-solving and engineering design), while professional skills (i.e., cultural competency, leadership, communication) were often not explicitly evaluated, but an assumed outcome of the immersive nature (i.e., cultural, geographic) of the SLCE experience.

The second trend that emerged from analyzing the impacts of the reviewed papers suggested that SLCE efforts were developed to fulfill the need of achieving specific outcomes for universities and students. Across the reviewed articles, the primary need identified was to provide experiential learning opportunities for students. Some papers alluded to unassessed long-term university and student impacts, such as graduating students who were giving back to society more broadly. The impacts on the universities varied, but largely focused on increasing access to funding opportunities, positive media and public exposure through curricular advancements, an enhancement to recruitment and retention efforts, and vague or watered down allusions to broader impact that align with societal benefits.

The third trend revealed that there were few instances where community impacts were described, and even fewer where these impacts were centered. Further, when impacts on the community were described, they were written from the authors' perspectives rather than collected from the community perspectives themselves. The most common community outcomes that authors did include tended to make assumptions of positive outcomes for the community by nature of the service students provided (e.g., community will benefit from having a product installed for them), the product of the engineering design project, community being educated on engineering topics broadly or on the specific product or service being provided, or community being empowered to use the installed product. Lastly, impacts on communities were not prioritized by authors when providing a description of the overall SLCE efforts. This outcome could be a result of the value systems at institutes of higher education, as mentioned earlier in the paper. However, this could also be a result of not centering community perspectives when approaching the design of a SLCE effort, particularly the evaluation aspect of the effort. Only two of the articles (Davis et al., 2014; Stein & Schmalzbauer, 2012) intentionally collected feedback from the community partners as evidence for the community impacts that were described in the paper.

When considering the findings of the impacts code through the lens of reciprocity, characteristics of mutuality in the partnerships could be inferred, but not confirmed, in some cases by observing how and in what ways impacts were framed and prioritized in SLCE efforts. For example, when the impacts only focused on student and/or university outcomes, it is possible that community impacts were not considered, or only seen as a byproduct at the conclusion of the student service experience. This inference can indicate an exchange orientation of reciprocity, where student learning is one of the impacts of the service and the community also experiences an impact from the service, resulting in a direct exchange of benefits during the experience. However, when community perspectives are excluded from the evaluation of the impacts of a SLCE effort, there is a risk of actions from students and universities that could produce negative outcomes on the community. These outcomes could undermine reciprocity, resulting in an imbalance in the exchange of resources, inauthentic relationships, and a partnership lacking transformative outcomes for the partner community.

5.5 Findings: “Relationships” code

In our review, coded data highlighting relationships addressed how the authors (from university or community) describe the roles of stakeholders (university, students, faculty, programs, community) and their interactions with each other. For example, we observed the form and frequency of communication between student design teams and users/clients in the partner community site. Three trends emerged in the literature when describing the relationships within SLCE: (i) relationships were reported in a student-centric and task-oriented way; (ii) relationships among all stakeholders were established through organized interactions; and (iii) relationships were mutually beneficial, collaborative, and multilateral between multiple stakeholders in SLCE efforts.

The first category includes articles where the authors reported the relationship in a student-centric and task-oriented way. Here, the content in the articles suggests these relationships were formed to facilitate student learning outcomes. For example, in one case, students were engaged in relationships with beneficiaries who served a supporting role in the student-centric goals of learning a technique, process, or approach. In other studies relationships were often established to exchange the benefits of learning outcomes and to finish projects. Furthermore, the community is often described as a “beneficiary,” indicating a distance and distinction in the relationship between the students and the community as provider and receiver.

The second category lends insight to articles where the authors reported the relationship by mentioning formal ways of facilitating interactions among stakeholders (primarily students and community partners). In these interactions, for example, formality was introduced by generating a list of the stakeholders, roles, or responsibilities involved in the SLCE efforts (e.g., Chisolm et al., 2014). Additionally, a structured approach for fostering relationship-building was implemented through organized meetings, rather than having students engaging with the community in spontaneous ways. For example, Onal et al. (2017) describe formal meetings initiated when the instructor visited the local community organizations where student teams could observe the challenges associated with the projects firsthand and discuss possible solutions with local community leaders. The project teams then continued meeting with the community organizations to share their suggested solutions and implement the agreed-upon revisions. As a result, the project teams and the community developed relational connections, facilitating conversations and promoting deeper relationships through regular communication, and using meetings as formal spaces for interacting.

The third category of articles includes content where the authors report relationships to be mutually beneficial, collaborative, and multilateral between SLCE stakeholders. In this category, the key purpose of such a relationship is to support both university and community goals. These efforts recognize that SLCE cannot exist without community buy-in and collective engagement. For example, Bratton (2014) acknowledges that their global program is only made possible through relationships in which “students learn to see ‘clients’ as ‘partners’ without whose cooperation, the solutions created, no matter how innovative or affordable, would be less likely to succeed in the long-term. The needs of partner organizations must be satisfactorily addressed for the relationship to flourish and the project to advance” (p. 207). The value of understanding such purpose is to facilitate the students' appreciation of knowledge from the community and enable all stakeholders to feel a part of and contribute to the project.

To summarize, the ways relationships are approached and described may reflect the authors' intentions to support reciprocity. If the relationship is positioned as student-centric and task-oriented, it can be challenging to determine how reciprocity is present via this framework element. Our analysis suggests that, in addition to investing time and approach, partners may authentically and systematically structure relationships through formal interactions and also invite mutual engagement and increased levels of reciprocity. Similarly, mutually beneficial, collaborative, and multilateral interactions among multiple stakeholders in SLCE efforts embody highly reciprocal relationships (DeBoer et al., 2022; Delaine et al., 2023).

5.6 Findings: “Problem identification” code

Data coded under problem identification led us to define this code as how stakeholders determined the issue, project, or course goals that were addressed/pursued by the SLCE effort. Problem identification existed along a continuum that involved varying degrees of collaboration among stakeholders. Some papers described how projects were selected, yet omitted who was involved in the selection process. These articles provided little insight into how the project's focus was determined and which stakeholders were involved. In other cases, when problem identification was described, it could be differentiated into four categories.

Through our data analysis, four categories of problem identification emerged: (i) one stakeholder is the primary driver of problem identification; (ii) collaboration among stakeholders to identify problems is evident, but limited; (iii) external stakeholders, such as intermediary organizations, drive problem identification; and (iv) problem identification is derived through a systematic, multi-stakeholder approach.

In the first category, one stakeholder was the primary influencer when determining what problem to address, even if opinions of other stakeholders were solicited. For example, Onal et al. (2017) described a partnership in which either students or the community independently submitted industrial engineering projects ideas to faculty members, who ultimately decided which projects to focus on in the course.

A second category emerged when there was some input from multiple stakeholders in identifying the problem. In these cases, there was community involvement or co-development in the generation and selection of projects. For example, university faculty and engineering students collaborated with middle school principals and students to propose new projects (Anderson, 2005). In a similar effort involving multiple stakeholders, Dewoolkar et al. (2009) involved the community and university faculty in the scoping stages of the project, in which problems were identified to also align with the aims of the course and requirements of the university.

External or additional input, such as the involvement of intermediary organizations or community-based research, represented the third category of problem identification. NGOs connected communities with university partners or provided insight into what local needs the SLCE effort might address. In their work with an Egyptian squatter village, an NGO provided information about which projects might benefit the children in the community (El-Gabry, 2018). Market research and needs analyses also contribute to identifying significant problems for the community. Arcipowski et al. (2017) described the identification of the problem of access to clean water through input from the community and research, which indicated that this issue substantially impacted the community.

The fourth category of problem identification occurred when multiple stakeholders were systematically involved in defining problems and selecting projects for SLCE efforts. Previous partnerships represented one potential way stakeholders worked together to determine which problems were most appropriate for SLCE. For example, in the development of the Bowman Creek Academy, a multi-year partnership was used to identify local environmental concerns (Boukdad et al., 2018). Intentional collaborative listening sessions can also support these methods of problem identification. For example, Dale et al. (2014) prioritized focus group sessions involving the university team, the village council, partners from a local NGO, and other community stakeholders. Additionally, transnational partnerships such as that between EWB-MSU and the Khwisero community created structures to support collaboration among stakeholders, enabling the project team to better serve the community (Stein & Schmalzbauer, 2012). An advisory board was established with community members to work with students on surveys and send them to local residents. Results from the surveys were then used to determine which schools would benefit from the SLCE efforts. These examples reflect attempts at partnerships that involve the continued and systematic co-generation of projects.

How problems were identified is shown to influence the degree of reciprocity embodied in a partnership. Of the articles reviewed, 13 did not explicitly address who identified the problem, 10 represented one stakeholder as the primary influence in problem identification, and 9 showed systematic, multi-stakeholder involvement. The range of stakeholder involvement reflects how projects are determined, how stakeholders can be involved in the selected projects, and who benefits from the outcomes and relationship. Involvement from the community that is superficial or marginal can result in projects that are not aligned to community needs, that are unworkable or prioritize student coursework over community issues. These patterns pointed to minimally reciprocal relationships. Conversely, the perspectives of multiple stakeholders and their active involvement in identifying problems can increase the relevance of projects to the community and incorporate their core values. Additionally, sustained partnerships may support the development of realistic timelines and achievable goals. These examples illustrate the opportunity for collaborative problem identification to increase reciprocity. Given these trends, we argue that addressing how problems in SLCE are identified provides transparency and supports efforts to create reciprocal partnerships.

Evaluating the approach to problem identification within an SLCE effort provides insight into how reciprocity is supported. Reciprocity may be reflected in how problems and needs are identified and by whom. Given these trends, addressing how problems in SLCE are identified provides transparency and supports efforts to create reciprocal partnerships.

5.7 Findings: “Knowledge” code

The coding of the articles for knowledge resulted in our defining this as information identified, exchanged, and generated as a result of stakeholder engagement in SLCE efforts. We utilized the term “information” broadly, including data gathered to understand the community and the context, the project-specific information exchanged between the participating stakeholders, and finally, new knowledge generated, or existing knowledge validated through research. We evaluated the articles for what knowledge is gathered/exchanged/generated and who participates in knowledge-related activities. Four forms of knowledge were represented across the articles reviewed. These include (i) process and product knowledge, (ii) knowledge of community, (iii) SLCE programmatic and theoretical knowledge, and (iv) SLCE course/program content and design knowledge. Additionally, embedded within each of these categories is an additional characteristic that promotes understanding of knowledge: loci of control, ownership, and/or creation of knowledge. Two more categories support this understanding: (i) knowledge generated and controlled by faculty and/or students, and (ii) community partners as providers of feedback within knowledge generation.

Process and product knowledge highlighted the engineering design process, problem-solving-related information, and information of products (mostly technical) that were designed. We observed two variations in process and product knowledge: one was observed in the articles that provide dedicated and detailed descriptions of the engineering design process applied in the program. These articles highlight the various ways the steps in the design process were completed by the students participating in the program. In this type of representation, students and faculty are described as predominantly responsible for this knowledge with no representation of the community. The other type of process and product knowledge was observed in descriptions of the community as an end user in the design process. The articles under this type highlight the critical role played by the community during the needs assessment, problem scoping, data gathering, validation of solution and results, and in the other design stages to provide feedback (Bratton, 2014; Brockman et al., 2017). These articles also commonly present the type of their program or process as user-centered or human-centered, or humanitarian engineering (Arcipowski et al., 2017; Florman et al., 2009; Stein & Schmalzbauer, 2012).

The second category of knowledge we observed—knowledge of community—relates to descriptions of community history, demographics, and other contextual information. Similar to process and product knowledge, knowledge of community is also commonly represented in two different ways: In the first type, articles addressed the community's geographic and demographic information and highlighted the significance of this knowledge in creating relevant designs and meeting user needs. However, these first type of articles omit reference to how the community knowledge was obtained and integrated into the design of the SLCE effort or project process. In contrast, in the second type, knowledge of community is both acknowledged and applied in the overall collaboration and implementation of the SLCE effort and project process. Here, the processes are commonly highlighted by describing it as “collaborative design,” and “sustainable approach for partnership” (Arcipowski et al., 2017; Florman et al., 2009; Rodriguez et al., 2014; Thelenwood et al., 2020) and the community members as “local experts” (Arcipowski et al., 2017; Stein & Schmalzbauer, 2012). Based on these descriptions, this category might best be described as “knowledge with community.”

The third knowledge category—SLCE program and theoretical knowledge—connected information from the SLCE literature and theories to the conceptualization and approach of the effort. The program and theoretical knowledge are shown as being gathered and communicated by the university members, that is, students and/or faculty participating (Cannon et al., 2016; Duff et al., 2014). Belonging directly to the academic community, faculty and students take on this role to perform the literature reviews relevant to SLCE and the specific nature of their own efforts and draw theoretical connections to support their program and research study. It is difficult to conclude reciprocity in this form of knowledge, as the process of obtaining and adapting this information is not discussed or represented in the articles. Community participation is only highlighted after the faculty and/or students have generated the program and theoretical knowledge.

Finally, the fourth category of knowledge—course/program content and design knowledge—highlighted information on the SLCE effort's structure, format, and design process. Knowledge specific to the program/courses is represented based on the course structure and learning goals: that is, first-year design (Dale et al., 2014), capstone (Cannon et al., 2016), projects as part of a course, or others. This is again presented in two ways. The majority of articles discuss and highlight the role of faculty as a significant contributor to the design of curriculum and course content. Although typically faculty were the key stakeholder in providing this kind of knowledge, NGOs and students were knowledge makers in some of the studies. For example, NGO partners provided feedback on the overall structure and content within which they hold expertise (El-Gabry, 2018).

Two critical themes emerged from our analysis of knowledge and its relationship to reciprocity. First, reciprocity is dependent on the form of knowledge highlighted in the articles—SLCE efforts where knowledge across these four categories involve the community to a greater extent tend to be more reciprocal. Second, the locus of control/responsibility for knowledge among the stakeholders varies based on the form of knowledge, which further demonstrates the nature of the effort's reciprocity. Except for programmatic and theoretical forms of knowledge, all other forms present some community role in the knowledge and, therefore, present elements of reciprocity. It is important to note that articles in which community partners were creators/owners of knowledge were not observed. Foregrounding community knowledge demonstrates increased efforts toward reciprocity.

Furthermore, collaborative and sustainable development may result from positioning community members “as experts in their own history, culture, community challenges, and strengths” (Arcipowski et al., 2017, p. 5). To summarize, the different forms of knowledge presented in the articles demonstrate reciprocity at varying levels based on the participation of the stakeholders in the process. While certain forms are traditionally limited to the involvement of university stakeholders and are more commonly represented, the few articles that did show examples of community inclusion through equitable participation demonstrate reciprocal relationships, which can result in sustainable SLCE efforts.

5.8 Findings: “Power, privilege, and oppression” code

The power, privilege, and oppression code emerged from the data to be defined as the ways in which the language used to describe perceptions of stakeholders and their actions indicate the privileging or asserting of authority of one stakeholder over another. The main characteristics shown to impact this element include (i) who is in control of the SLCE decision making and implementation, (ii) how power is conveyed through language, and (iii) control of and access to resources.

The first characteristic, namely who is in control of the SLCE decision making and implementation, emerged within four categories across the articles. These categories reflect who is empowered and/or to what extent (and how) power is shared as described within the SLCE efforts. The first category reflects articles in which students and/or faculty held more power. In this category, power dynamics may be problematic, as there is potential for exploiting community partners. Additionally, this category includes cases where power imbalances were discussed as being addressed through the “long game”: Here, authors justified the privileging of students, assuming trickle down to communities over time as these students developed into professionals embracing social responsibility. The second category reflects articles where effort is taken to empower the community. In this category, power dynamics was acknowledged, and attempts made to level power imbalances/efforts taken to empower community partners were presented. (Arcipowski et al., 2017; Aslam et al., 2014; Boukdad et al., 2018; Nieusma & Riley, 2010; Stein & Schmalzbauer, 2012). The leveling of power dynamics can be a complex challenge and, in some cases, power imbalances persist in favor of students/faculty despite efforts to level them. The third category reflects SLCE efforts where community members were involved as full partners in the SLCE efforts (Florman et al., 2009; Kang & Chang, 2019). Lastly, the fourth category includes efforts where power is retained by the community (Ing et al., 2012). In this scenario, the community partners held authority over all decision-making processes. Additionally, articles were reviewed in which power dynamics were not addressed or could not be determined.

The second characteristic centers the ways in which language reflects the power dynamics within the SLCE literature, highlighted by three categories. The first category involves instances when subtle language cues disempowered the community partners. Here, language such as “developing world,” “beneficiaries,” and describing the SLCE efforts as “fun” and “a good time” for students reflect ways in which university stakeholders hold more power than community partners. Additionally, deficit language was often used to describe the community context as well as community skills/abilities. In the second category, language cues that indicate an equitable consideration of power were used within the articles. Terminology such as “peers” and the use of descriptors that highlighted assets rather than deficits were shown to center the strength of partners and opportunity within the community. Additionally, the privilege of the students, such as the use of “socioeconomic advantage,” was sometimes declared in meaningful ways that acknowledged the power imbalances and sought to minimize the impact. Lastly within this category, language was used that reflected priorities within the SLCE context. Here, “soft skills” can highlight how professional skills may be considered secondary to technical skills in an engineering context. Additionally, some articles showed a desire to move away from the “service-learning” language, as “service” may connote a more powerful stakeholder serving the other.

The last characteristic to emerge reflected control of and access to resources. This characteristic highlights how resources reflect power and control within SLCE. Here, the ability to obtain, access, and utilize funding was the primary way this characteristic was manifested.

To summarize, funding, decision making, and outcomes were areas that indicated privileging of one stakeholder over another. Overall, the language use provided a direct way to determine how power was distributed or withheld from stakeholders. Ultimately, this element mirrors the essence of reciprocity, as in certain ways reciprocity can be defined as the sharing and/or equitable distribution of power within and across SLCE efforts. With equitable power, a partnership can strongly reflect reciprocity.

Through the coding process and the results of our data analysis, the trends we see suggest an emergent analytical framework. In the next section, we situate the overall trends, explain the analytical framework, and recommend its usage for others.

6.1 A breadth of engagement with reciprocity across the framework elements

Engagement with reciprocity is shown spanning from exploitative to exemplary. It also provided a variety of entry points for understanding how reciprocity is or is not being actualized. To one extent, there is evidence of singularly beneficial SLCE efforts (actions and benefits of engagement initiated and received by only one group of stakeholders), which could benefit from more reciprocity. Conversely, SLCE efforts that demonstrated more reciprocal elements were observed to not be singularly beneficial to one stakeholder or performance of a specific service. Rather, they were efforts that were co-generated across stakeholders, collaborative in all phases of development, execution, and sustainable over the longer term.

We show in Table 6 that nearly every cell (framework element by level of reciprocity) has an example excerpt from our literature review that exemplifies the breadth of engagement with reciprocity. This breadth also represents a lack of consistency in forms and levels of reciprocity, leaving individuals and institutions to define for themselves what is acceptable, often with few benchmarks. We see some general strength areas (approach, with high levels of reciprocity described in the overall program plan, and student impacts, in that students take away skills and knowledge about cross-institutional exchanges) as well as some generally exploitative areas (power relations implicitly or explicitly described, treatment of knowledge, setup of relationship, and the elements related to community). We see some areas of overwhelming lack of information about reciprocity (methods, student context, university impacts).

To further articulate the breadth of engagement with reciprocity, we created a heat map across a matrix of the 52 reviewed papers and the framework elements, presented in Figure 2. This figure illustrates the extent to which reciprocity is present within the reviewed engineering SLCE literature, using horizontal and vertical dimensions to intersect findings. The horizontal axis represents the framework elements (listed in Table 3), while the vertical axis lists the 52 reviewed papers in a color-coded fashion indicating reciprocity levels. Papers are sorted from least to most reciprocal, with paper 13 being the most reciprocal and paper 116 being the least. This heat map provides a snapshot highlighting the range of highly reciprocal to exploitative practices across studies, the extent to which few studies embodied the same level of reciprocity across elements, and the most consistent trend being a lack of information on specific elements that would clarify the level of reciprocity.

As a synthesis of Figure 2's heat map, we provide an alternative model (Figure 3) to illustrate the extent to which each framework element is present across the reviewed papers. This model shows the attention given to each framework element individually, summing the number of papers that address it at each indicator level. Three aspects are emphasized: vertical axis arranges levels of practice (−1 to 3), horizontal axis lists framework elements, and three shapes represent paper counts for each element's reciprocity indicator. For example, in the first column, out of 52 papers, 35 lacked sufficient student context information, 12 papers showed some reciprocity, while only 5 exhibited clear or exemplary reciprocity levels.

Several discussion points emerge, such as 14 papers demonstrating evidence of harmful community relationships when analyzed with respect to power, privilege, and oppression. Further, many papers do not provide sufficient detail within many of the elements, indicating a need for greater clarity around the extent of engagement with reciprocity. Trends emerge from these figures, illustrating the ways in which SLCE is positively engaging with, struggling to engage with, and neglecting to articulate reciprocity (discussed in more detail below).

Such breadth of engagement with reciprocity complements previous work in this area that demonstrates how relationships in community engagement can operate differently with varied approaches to reciprocity, including conceptualizations of thin and thick reciprocity (Jameson et al., 2010), transactional and transformational partnerships (Enos & Morton, 2003), and transactional, cooperative, and communal models of partnerships (Thompson & Jesiek, 2017). These models acknowledge that inequity and reciprocity cannot coexist in the long term, which aligns with assumptions proposed by the Dostilio et al.'s (2012) framework. Our work complements and expands on existing work by demonstrating the ways reciprocity can exist across multiple levels and forms of engagement at the same time.

7.1 Intentional pursuit of reciprocity

In the literature we reviewed, conceptual understanding and explicit acknowledgement of reciprocity as a construct and a goal for engineering SLCE was largely assumed, implicit, or missing. Scholars and practitioners must be aware of the various forms of reciprocity to meaningfully pursue reciprocal relationships. The wide breadth of engagement with reciprocity presented in this work enables reciprocity to be envisioned and reflected upon to advance in SLCE efforts. For example, narrow conceptualizations and enactments of reciprocity can indicate an imbalance of power and pose the need for a critical reevaluation of the goals and structure of an SLCE effort. Our framework can serve as a checklist along a multidimensional spectrum of reciprocity, for thoroughly evaluating SLCE efforts. Further, this research provides indicators that stakeholders can use to gauge their progression toward more reciprocal partnerships. Integrating and assessing reciprocity across the elements can enhance the design, implementation, and evaluation of SLCE efforts.

Reciprocity must be explicitly acknowledged and pursued in SLCE efforts. In the SLCE conceptualization phase, centering contexts, knowledge structures, as well as issues of power, privilege, and oppression are critical in aspiring toward advanced forms of reciprocity (Claussen et al., 2019; Delaine et al., 2023; Leydens & Lucena, 2017). For example, partnerships that value community knowledge pursue high-level geographic and demographic information about the community and also deliberately seek to understand the context from the community stakeholder perspective and culture. Additionally, when power dynamics are acknowledged, partnerships can be pursued as collaborative, power-sharing relationships in which reciprocity is a cornerstone (Dostilio et al., 2012). When SLCE is enacted in such a relational approach, it enables moving beyond the exchange of benefits and supports stakeholder transformation.

7.2 Language of reciprocity

Our analysis revealed an imbalance in articulation of reciprocity that is skewed toward a university- and student-centric perspective, positioning the university as the central stakeholder. In our review, language connected all elements of the analytical framework. Language choices reveal implicit beliefs about power, privilege, and oppression. Reynolds (2014) notes that deficit language can reveal whose knowledge and contributions to a partnership are most valued. Certain words can marginalize the community or its members, while indicating how members from the university view themselves from a position of authority. For example, referring to a community as “needy,” “disadvantaged,” or “underprivileged” reflects an “implicit disacknowledgement of the knowledge in the community” (Reynolds, 2014, p. 81). As such, language that “others” the community partner might indicate a view of reciprocity that is transactional or minimally reciprocal. It is important, therefore, for authors to consider the language used in dissemination as well as in practice, as it could highlight the extent and orientation of reciprocity achieved.

7.3 Broadening the scope of engineering SLCE research

Our findings indicate that most articles report SLCE efforts from the university perspective. Emphasis on framework elements such as student and university context, impact on students' learning, university and faculty approaches to designing and facilitating SLCE courses, and challenges encountered by faculty to initiate SLCE courses highlight this focus. This approach is perhaps promoted by the scope of engineering education journals being maintained or perceived to be linked to the professionalization of engineers. Therefore, orientation toward university stakeholders is expected. Yet, as SLCE is explicitly oriented toward community partnership, the scope of research conducted by engineering educators should be expanded to include outcomes, impacts on, and benefits for external stakeholders (in this case community partners). To support higher levels of reciprocity, we recommend that journal editors request, and researchers provide, more detail about the community: their involvement, their contributions during the various stages of the partnership, the impact of SLCE efforts on the community, and any change or transformation of the stakeholders. We also encourage journal editors to enhance the scope and mission of their journals, incorporating requirements for acknowledging stakeholders in partnerships. Lastly, reciprocity can be further supported by integrating theoretical perspectives in SLCE research and practice. Leveraging existing theoretical and conceptual frameworks can provide evidence-based insights into the various stages and forms of engagement in SLCE efforts (Clayton et al., 2010; Delaine, 2021; Delaine et al., 2019). Prior literature reviews in similar domains have also concluded that using theory in SLCE settings can support the efficacy of intended partnership outcomes (Holsapple, 2012) and suggest that drawing on other fields can inspire SLCE research and practice directions specific to engineering (Shermadou & Delaine, 2022). By making a conscious effort to center and report on community perspectives and experiences, as well as intentionally integrating theoretical perspectives in SLCE, highly reciprocal partnerships can be more readily pursued.

7.4 Reciprocity statements

A limitation emerging from our review is the often vague or unclear articulation of reciprocity, despite its centrality to SLCE. SLCE efforts could benefit from authors explicitly articulating the presence and form of reciprocity, fostering a broader and deeper understanding of its achievement in engineering SLCE. Therefore, the final implication of this study is the recommendation for authors to include statements of reciprocity in their articles, similar to the positionality statements included in qualitative research (Martin et al., 2022). The elements in our analytic framework can provide a meaningful starting point for such statements. The reciprocity section could elaborate on the approach taken by the university to initiate, build, and sustain reciprocal partnerships with the community. Authors should detail their intended and actual impact, and provide direct evidence of efforts toward reciprocal SLCE. Leverage existing frameworks and orientations from reciprocity scholarship can support articulating its elements in engineering SLCE efforts (e.g., Dostilio et al., 2012; Thompson & Jesiek, 2017). Existing scholarship can be particularly useful for practitioners and researchers seeking a nuanced understanding of the dynamics of meaningful versus limited relationships among stakeholders. An explicit reciprocity statement can support clear intentions to engage in SLCE in more reciprocal ways.