Embracing complexity : Co-creation with retired immigrant women

This article discusses a co-creation project carried out by a post-secondary immigrant-serving agency and retired immigrant women. We posit that, by engaging with immigrant women in co-creation, we not only deepen our understanding of the challenges they face in retirement, but also generate valuable insights into the process of participatory design and collaboration; namely, the importance of recognising complexity as a productive, rather than challenging, aspect of knowledge co-creation and collaboration. We show that, by intentionally embedding methods which address issues related to reflexivity, power and difference into the co-creative process, the challenges posed by complexity can be mitigated, leading to positive outcomes for all partners.

As the Canadian population ages, and the baby boomer generation continues to retire in greater numbers, a new group of retirees -senior immigrants -is emerging as a population of interest to researchers and policy-makers.Given the long-term influence of migration on an individual's life, and the over-representation of immigrant women amongst those facing underemployment, discrimination and social isolation, there is a growing consensus that their experiences of retirement may be different from those of Canadian-raised women and therefore merits further exploration (Bauder, 2003, McDonald and Kennedy, 2004, Shin, 2014, Teelucksingh and Galabuzi, 2005).Attempts to fill this knowledge and programming gap tend to treat retirees as an undifferentiated population, use a deficit model of engagement, or fail to go beyond individual level interventions and address systemic challenges to successful ageing and retirement (Shin, 2014, Teelucksingh andGalabuzi, 2005).
This article discusses our efforts to deliberately work differently with retired immigrant women and to engage with them as partners in research and action planning.We posit that co-creation not only deepens our understanding of the challenges they face in retirement, but also generates valuable insights into the process of participatory design and collaboration; namely, the importance of recognising complexity as a productive, rather than challenging, aspect of knowledge co-creation and collaboration.We show that by intentionally embedding methods which address issues related to reflexivity, power and difference into the co-creative process, the challenges posed by complexity can be mitigated, leading to positive outcomes for all partners.
Following a brief discussion of the project background and data collection methods, we explore the idea of co-creation.We then summarise the significant developmental moments experienced during the co-creation process and critically reflect on how our methods to deal with reflexivity, power and difference enabled, and sometimes hampered, the process.While the focus of this article is on the particular experience of engaging retired immigrant women in a co-creative research process, the findings presented here are important for researchers engaging under-represented populations in participatory processes.We identify promising practices for working within a collaborative and participatory framework based on the lessons learned from this project.These insights may be valuable for researchers interested in translating research into more responsive and impactful program and policy designs.

Methods
The co-creation approach emerged from a joint research project between a local community college and an immigrant-serving agency.From its inception, this project aimed to do more than simply collect data that would sit in a tidy report on a shelf.Rather, the goal was to move from research findings to action.The aim was for team members from both organisations to work with a group of retired immigrant women to co-analyse data and co-create initiatives.As such, this was imagined as a collaborative process between staff and research participants to inform the creation of meaningful interventions by drawing on the lived experiences of retired immigrant women.The project's methods reflected our desire for a co-creation approach.We intentionally used the data collection phase as a means of building relationships and trust.Semi-structured interviews, for example, invited participants to shape the discussion and their narratives on their own terms.These in-depth conversations built rapport between the researchers and participants.Following the interviews, participants were invited to re-engage with the project through a three-month participant observation phase of data collection.The contribution of participant observation to a deeper understanding of our data is detailed below.However, it also allowed the team to continue building trust and relationships.
Working with our community partner, and using purposeful and snowball sampling techniques, 25 research participants were recruited to participate in the semi-structured interviews.We sought a diverse sample of women from various ethno-cultural backgrounds and with varying educational and professional experience (see Table 1).All participants shared the common characteristic of having a minimum of two years' experience in the Canadian labour market and being retired for at least two years.We were interested in exploring the experiences of retired immigrant women as a subset of senior immigrant women.It is important to note that we do not use 'immigrant' to signify the technical and legal status of not being a citizen.Similarly to Man (2004), we employ the common, everyday usage of the term to refer to people who are seen as immigrants, either by themselves or others, regardless of their legal status.It is clear, however, that the data cannot be considered through the homogenising lens of 'retired immigrant women'.The multiple and overlapping social positions and structural factors that shaped their lives required an intersectional analysis.The adoption of an intersectional lens reminded us to move beyond static, homogenising categories of 'immigrant' and 'women' to reflect on how identities overlap with factors such as ethnocultural background, language skills, education, employment history, year of immigration, and family composition.Further, it encouraged consideration of how these factors interact with existing structures and systems that impact daily lives (Fanjoy, 2017).Our research questions sought to unpack the effects of migration on women's labour market experiences throughout their life course, and thus in later life, their retirement.Questions focused on the experience of retirement and immigration, and how these experiences affected their career and social participation as well as engagement in the community postretirement.Fifteen stakeholders were interviewed, including program managers at immigrantserving non-profit agencies, government policy-makers, and those who work closely with retired populations.The interviews aimed to understand the broader structural context of immigration, ageing and retirement, including policies and services.
To contextualise and triangulate the findings from interviews, the research team carried out participant observation with a sub-sample of 10 women from the larger sample of interview participants.Observation occurred approximately once a week over a period of four months.The research team engaged in periods of observation of the daily lives of the women, allowing participants to guide the process and include the researcher when and where this felt comfortable and appropriate.Observation took place in both public and private spaces and involved a wide variety of contexts, including fitness classes, faith-based activities, volunteer activities, home visits, classes and workshops, and family events.This allowed for an intimate picture of how participants engaged with others, how they built relationships or created ties in the community, what barriers they faced, and their opportunities for social engagement or wellbeing.Researchers recorded the women's interactions, communications, body language and social context.After the observation sessions, researchers conducted social and event mapping exercises to trace meaning and connections.This aspect of the research was key to strenthening our description of our participants' social participation (Geertz, 1973).For example, one participant spoke at length in her initial interview about the importance of her weekly aquacise class and how much she misses it in the winter when the poor weather often prevents her attending.Through the participant observation phase of the data collection, a member of the research team was able to attend several classes with her over a three-month period.The team member mapped the interactions she observed at the classes -extra health and nutritional advice shared by the instructor in post-class chats, tips on grocery deals, clothing sales etc. shared by classmates in the hot tub after class, and the highlight of a weekly lunch with some classmates after class at the local fast food restaurant.Observing and mapping these interactions allowed the team to deepen our understanding of why the participant prioritised her aquacise class and the important interactions that were missed when she could not attend.
Once the interviews and participant observation phases were complete, a group of six women opted to participate in the analysis and action planning phase.We met over a seven-month period to co-create strategies and initiatives aimed at combating isolation and overcoming the challenges associated with ageing.During this time, the co-creation group met bi-weekly to discuss and refine the research findings.The group transitioned from discussion to action by developing and testing two pilot initiatives exploring how to tackle the challenges facing retired immigrant women.This part of the process is described in more detail in the later section on 'Key Developmental Moments in the Co-Creative Process'.
In many ways, the perceptions and experiences of the immigrant women retirees who participated in the project align with existing research on retirees.They were keen to stay active and engaged, and to continue contributing to society outside the realm of paid labour.They believed that social and civic participation was key to their wellbeing as they aged and did not view retirement as a period of decline and withdrawal.However, important points of divergence emerged when we started to explore the intersection of life experiences with systemic contributors and barriers to wellbeing and social participation.Stop and start entries into the labour market, financial precariousness and fuzzier boundaries have always been what retirement looks like for a large number of women and precarious workers (McDonald, 2006).Immigrant women's over-representation in these groups, however, puts them at greater risk of social and economic marginalisation as they age (Man, 2004, Preston, 2014, Premji et al., 2014).For example, participants' experiences of labour market marginalisation through lack of credential recognition intersected with language barriers and lack of familiarity with Canadian pension systems to negatively impact their economic wellbeing later in life.At the same time, the gendered nature of caregiving duties was compounded by the challenges of raising children in an unfamiliar society, as well as cultural norms and values related to spousal caregiving in old age.The accumulated impact of these challenges led to an increased risk of social isolation for many of our project participants.(For details on research findings, see (Fanjoy, 2017)

Co-Creation
Co-creation research, which engages participants in meaningful ways throughout the process, is increasingly seen as having value, especially when that research aims to address social inequities (Mulroy, 2004).Collaborations between educational institutions, community organisations and people with lived experience of the research subject are understood to be 'vital sources for teaching, research and practice' (Strier, 2011).Scholarship on collaborative research projects suggests that inclusion of community residents and participants with lived experience relevant to the investigation may generate different kinds of knowledge and information from that of traditional research praxis (Silka, 1999).In sum, 'by including community residents in research and planning, researchers can create programs that have immediate relevance and policy implications' (Farquhar and Dobson, 2005).The participation we seek in co-creation is more involved, however, than in most participatory projects.If participation is viewed broadly to consist of the spectrum from passive through to active engagement, co-creation is firmly planted on the active end (Voorberg et al., 2014).It also involves a shift from dialogical to transformative relationships.Rather than simply seeking input from participants, the researcher seeks to co-create a transformation, or shift, in social conditions through the process (Beebeejaun et al., 2013).
While collaborative research agendas and knowledge co-creation are laudable goals, they also raise methodological and ethical challenges.Much of the research in this area points to the role that power relations play in such processes.While the 'co' in co-creation suggests an equal sharing of responsibility, such processes -like all social relations -are shaped by the way stakeholders are positioned in relation to one another.One of the challenges with knowledge co-creation lies in the way knowledge and power are conceived, and these conceptions can limit or impede the outcomes of research (Olesen and Nordentoft, 2013).For example, often co-creation efforts are shaped by ideas of 'democratic collaboration and knowledge production' but this collaboration is based on fixed positions and power relations (for example, research participant and paid staff person).There are also epistemological challenges with co-creation: to what extent are we truly able to understand another's world view across profound difference in age, gender, class, or immigration experience?A critique of collaborative knowledge creation processes offers that they are challenged by the reality that 'we want to let go of control; at the same time, we want to stay in control' (Olesen and Nordentoft, 2013).Despite the ideals of democratic participation and shared knowledge production, research projects take place within a real-world context of limited resources, short timelines and unequal power relations, all of which come to play a role in how knowledge is produced, shared and experienced.
There are also questions about the efficacy of co-creation methods in achieving systemic or structural change.A review of co-creation projects in the public sector found that, while most projects report on processes and factors that influence the success or failure of the work, they rarely touch on outcomes (Voorberg et al., 2014).Strier (2011) also raises questions as to whether partnerships and collaborative research are a means to solve complex social challenges, or if the main result is merely an endless cycle of relationship building.
We acknowledge that collaborative research and knowledge co-creation is complex and imperfect; however, at its core it is an emancipatory process that seeks to include the voices and experiences of research participants in the problem identification and resolution finding process.Therefore, while this work is complex, it should not be abandoned in favour of less participatory -and more traditional -research methods.We argue that, by acknowledging and addressing rather than glossing over the complexities inherent in the co-creation of knowledge through collaborative research, these challenges can be mitigated.In the following section, we describe the key developmental moments that emerged from our co-creation process.These moments represent the main challenges to, as well as opportunities for, knowledge co-creation that we encountered throughout our process, and how the research team responded to them.

Key Developmental Moments in the Co-Creative Process
In the section that follows, we identify five 'key developmental moments' in the life of this project.We use these moments as a way of articulating the productive challenges and key learning experiences that emerged through the process.Dozois, Langlois and Blanchet-Cohen (2010, p. 41) define developmental moments as 'instances when the initiative shifts or moves forward in some significant way; moments of clarity, strategic insight, serendipity, connections, and/or movement'.It is important to emphasise that these five moments can only be identified in hindsight -this was not a timeline developed in advance of the group but rather emerged in collaboration with the co-creation participants, the facilitators from both the immigrantserving agency and the college, and in relation to external events.

BUILDING THE GROUP
A sample of women from the 10 research participants that were involved in the focused ethnography formed the co-creation group.In hindsight, the research team reflected that it was unlikely that research participants would have had the same level of interest and buyin had they only been engaged in the interviews.Of the 25 women who participated in interviews, 10 participated in the participant observation, and of those 10, five became active members of the co-creation group.One action team member participated in the interviews, but not the participant observation, making a total of six participants in the analysis and action planning phase.When the co-creation began, the research team hoped to engage 10 women as members of the group.It became evident, however, that despite expressing interest in being involved, these participants faced personal barriers that prevented them from being engaged.Reasons for not participating included personal health issues and declining health of a spouse and the need to be available for care.There were also unstated or unclear reasons for not participating.One participant expressed concerns around social anxiety and geographic isolation during the interview, so this may have been a factor.
In the end, the six participants who attended the first and following co-creation meetings were eager to be involved and engaged.There are several key factors to note about the cocreation group that emerged early on and were ongoing throughout the process.First, the presence of pre-existing peer groups was a factor that likely facilitated engagement in the co-creation process (people are more likely to attend an event if they can go with friends).This was especially the case for three women who were more recent immigrants and attended all their social activities and community activities as a group.These peer groups were positive in that they encouraged participation of individuals who may not have been involved if not for their peers.However, these peer groups also challenged the building of a cohesive shared group identity (discussed further below) and occasionally created challenges for facilitation.Another positive/challenge was the strong bond that had developed between the research assistants (RAs) and the co-creation participants during the participant observation phase of the research.This factor also likely contributed to the women's engagement and participation, yet created an expectation that the RAs would remain involved throughout the co-creation period (which was not the project team's role and, as expected, diminished during this stage of the research).It proved a delicate operation for the RAs to carefully and thoughtfully extract themselves from co-creation once the group had met a few times and there was greater comfort participating without their presence.Both these factors -the presence of pre-existing peer groups and the trust between the RAs and the participants -point to the importance of trust-based relationships for securing participation in a new initiative.They also point to the challenge of balancing operational realities (the end of RA contracts, for example) with the need to sustain relationships built over time and through research.

REFLECTING ON RESEARCH FINDINGS
The co-creation group's first activity was to review the key findings from the interviews and observations, and to provide comments and feedback.The goal of this activity was twofold.First, to hear from retired immigrant women themselves about what they thought of the key research findings -were they applicable to their life experience?Did these findings make sense?Were they accurate?Did they reflect their life experience?Was there anything we missed?This process would add depth and strengthen the initial findings.The second goal was to encourage co-creation participants to take ownership of the research data and to feel like they were active contributors to the research process.The reflection meeting was structured to encourage ownership.The lead researcher shared the key findings on PowerPoint.The group was then split in two to facilitate small group discussion.Each group was given a copy of the presentation, which they went over slide by slide, discussing the key points and considering their relevance/applicability.This strategy was effective in allowing the participants to feel their perspectives and opinions mattered.It also allowed the research team to go deeper into the research findings, adding context and specificity to the initial findings.Reflecting back on the process, several participants commented that they appreciated having the opportunity to be involved in the generation of research knowledge and making systems and policy recommendations.One participant commented: 'I had the opportunity to give suggestions, resolutions.I'm glad I went through something like this because I can see the overall [picture].'

IDENTIFYING PRIORITY AREAS
The initial co-creation meetings focused on building group cohesion and refining the research findings.Because the goal of the co-creation process was to create two activities or pilot initiatives related to supporting retired immigrant women, it was important to move from this initial group-building process to a process that would lead to the development of activities.The process adopted identified several priority areas related to the civic engagement and social participation of immigrant women.These were brainstormed by the group and included a variety of challenges and barriers facing retired immigrant women.Following this process, the research team and co-creation participants voted on the areas they felt were the most important to address.They settled on 'loneliness and isolation' and 'caregiving challenges'.Over the course of several meetings, the co-creation participants worked with the research team to dive deeper into the causes, solutions and actions related to each of these issues.The goal in this process was to work as a group to explore and identify possible opportunities for action and program development.

ACTION PLANNING
Based on the brainstorming during the previous sessions, two possible activities for the group to work towards were identified.Rather than asking the co-creation participants to come up with an idea for an activity on their own, the facilitators were to present an idea and have the participants provide feedback.This decision was made based on the facilitators' sense of the group's readiness and capacity to develop an idea from scratch.In the past, it had proven effective to make a series of suggestions and have the group critique them and provide insight and expertise.This process helped to identify possible activities for the co-creation group to pilot.The following activities were presented to the group: Issue: Loneliness and isolation > Activity: Peer support group Issue: Caregiving challenges > Activity: Toolkit for service providers to identify signs of caregiving challenges When presented to the group, nearly half of the participants were enthusiastic and eager to move forward.The others seemed confused and unsure about both the process and the activities.This division in the group happened to fall along peer lines.Participants who had been in Canada longer, and had stronger English language abilities and a higher level of comfort with this kind of action planning, jumped in and started making suggestions.The participants who had been in Canada for a shorter period of time, and had weaker English language abilities and a lack of familiarity with this kind of action planning process, seemed confused by what was being suggested and disengaged.The facilitator opted to divide the group along these lines, with the participants who were 'ready to go' in one group, and the other participants in another group.In the past, mixing the groups had been successful in helping bridge the two group dynamics.In this case, however, the facilitator felt it was important for the former group to move forward with their momentum, and for the other group to have the opportunity to ask questions and ultimately move towards understanding the process in which they were engaged.It was also the case that the group that was more eager to move ahead included participants with the capacity to offer peer support, and had already demonstrated these skills in relation to the other participants in the group, whereas the other group had more experience with caregiving issues and could perhaps (the facilitator hoped) focus on those issues.This moment revealed several challenges within the co-creation process.First, until this point, the group had remained at the discussion and brainstorming stage of the process and this was the first opportunity to move from ideas to an action-oriented activity (i.e.help plan and execute the two pilot activities).While half the group had the capacity to jump in and take this step (and expressed an eagerness to do so), the other half of the group expressed uncertainty and a lack of clarity around the proposed activities.Second, this tension revealed the challenge presented by the diversity within the group.In previous meetings the differences between the participants (in terms of English language capacity, experience working in groups and leadership skills) had been evident but productive.Those participants who had been in Canada longer offered support, guidance and recommendations to those who had been in Canada for a shorter period of time.As the group moved into action planning, these differences became more salient as the different abilities within the group came to the fore.The facilitators had to make decisions about how to support the very different needs and abilities of the participants within the group while also keeping everyone engaged and interested in participating.
It was clear that the idea for a toolkit for service providers to address the needs of caregivers did not have the same level of interest or buy-in as the peer support model.Therefore, this idea was dropped and the group focused on peer support training.The peer support training and pilot allowed for all participants to be engaged but at different levels.The facilitators also focused on a second pilot initiative that would be more facilitator directed but still provide opportunities for the participants to engage and offer feedback.Ultimately, this became a knowledge exchange event between the City of Calgary's Age Friendly team and the cocreation participants (discussed further below).
Two key learnings stem from this developmental moment.First, it is critical to acknowledge -rather than ignore -the differences that exist within a group.This may be especially pressing when the group is vulnerable or faces barriers to engagement (i.e.retired immigrant women).While the goal of this project was to develop and test two actions to address the barriers to social and civic participation, it was also critical that participants not be left behind in this process.There were moments when it would have been tempting to simply push on and inform participants of what was going to take place, but the research team felt that would risk alienating the more vulnerable participants.This points to the second key learning, which is that it is important to be flexible and adaptive in these emerging processes to the needs and desires of the group, as well as to the group's ability to take action collectively.In being flexible and allowing plans to shift, we were better able to keep participants engaged while also meeting the desired project outcomes.

EXECUTING ACTIONS
'In the midst of complexity, taking action can be daunting because there is always more to do [which] might increase your chances of success.However, you won't learn much if you never actually try anything.Developmental evaluators play an important role in helping groups understand the importance of quick iterations; of learning by doing' (Dozois, Langlois & Blanchet-Cohen 2010, p. 45).

Peer model
The peer support program was formalised as one of the pilot initiatives for this project.The group was enthusiastic about a peer support program and identified training they felt important to include: where to make referrals, boundaries and expectations, cross-cultural communication, confidentiality and privacy, and effective listening.Working in partnership with an immigrant-serving agency was advantageous as they had a program for retired and/ or elderly immigrant women.This program, My Community, My Home, offered a good space Staff also attended the program to observe interaction between the two groups.It was evident that the co-creation participants were able to offer support and assistance to the other women.This program draws a more vulnerable group of senior and retired immigrant women -many of whom struggle with English, are not familiar with public transport and/or don't drive (so depend on others to leave the house), and have less experience attending programs.The co-creation participants were able to offer a gentle form of mentorship by speaking about their own experiences.An unintended but valuable consequence of building the peer support program within an existing program was that participants from the co-creation group had a program to continue going to after the co-creation process ended, which provided important continuity for those seeking social connection and continued interaction with one another.

Knowledge exchange
In addition to the development of the peer support program for retired immigrant women, the co-creation group settled on planning and organising a knowledge exchange event between policy-makers and research participants.This idea emerged out of the group's desire to gain a better understanding of how policy decisions are made and how they might engage with decision-makers.The facilitators felt that a knowledge exchange event would be a better fit since the original idea (a toolkit for support workers) had not gained traction within the group.
The co-creation group identified Calgary's Age Friendly Strategy as a key policy lever for meeting the needs of retirees from both immigrant and non-immigrant backgrounds.In November 2017, the team set up a meeting with staff from the City of Calgary (City) to learn more about the Age Friendly Strategy and civic engagement more broadly.The facilitator met with the City of Calgary staff in advance of the knowledge exchange event in order to identify areas of shared concern.The City expressed an interest in learning from the research participants about the facilitators and barriers to attending available programs in the community.They also suggested doing a presentation on civic engagement and the Age Friendly Strategy.To prepare for the meeting, the City gave the research team sample discussion questions regarding social participation.The co-creation group worked together to answer these questions and prepare for the discussion with the City staff.On 22 November 2017, staff from the City of Calgary, the co-creation research participants, and members of the research team from the college and immigrant-serving agency came together for a productive three-hour knowledge exchange meeting.The event itself was an opportunity to see how differently positioned stakeholders can come together to share information and learn from one another.The knowledge exchange format provided space for the participants of the research group to share ideas and ask questions directly of City staff (and vice versa).As such, it was a positive two-way learning opportunity.The event also sat well with the overall goals of the research project of influencing systems and policy by connecting those with lived experience to those who help shape the systems and structures that impact older adults from immigrant backgrounds.

Discussion
The five developmental moments identified above highlight some of the challenges encountered in the process of moving from data collection through to action implementation.Upon reflecting on the five moments, and how challenges emerged and were dealt with, three important themes became apparent: reflexivity, power and difference.In the remainder of the article we discuss how, in hindsight, the methods we chose to embed in our co-creative process addressed all three.

REFLEXIVITY
'Reflexivity enables us to highlight the political dynamics of our endeavours, attention to which would otherwise tend to be absent from the representation of our project' (Orr and Bennett, 2009).
Reflexivity addresses the need to be attentive to the process of knowledge creation and the various positions collaborators inhabit within the process.Building on Strier's (2011, p. 95) contention that successful partnerships depend on 'the capacity of the leaders to provide a learning and reflexive organizational culture', we strove to ensure reflexivity was built into the process, from data collection through to project implementation.Developmental evaluation was chosen as the means through which intentional reflexive practices would be built into the research and project design.
Evaluation can take many forms.Summative evaluation looks back on the arc of a project and assesses project impact through outputs and outcomes.Formative evaluation takes place during the project, but is also oriented towards understanding and evaluating impact.In contrast to these approaches, developmental evaluation is embedded alongside the project as it progresses.It is best suited to projects that 'work in uncertain territory, developing and testing their strategies as they proceed' (Dozois et al., 2010) and is oriented around supporting learning and growth as projects emerge and their purpose is clarified.
While the team had a general idea of the goals it hoped to achieve, the exact path forward was not predetermined.As a participatory project that sought to include the perspectives and lived experiences of participants, the research team wanted to leave the 'how' of the project open.In order to do so, it was critical that researchers be flexible and responsive to the ideas of the research participants.This required establishing a balance between the structure of the project and the power differential and competing interests embedded within it, and the desire to be participatory and open to new ideas and suggestions.To maintain this balancing act, the research team drew on the principles of developmental evaluation to support the learning of the project, to offer course correction when needed and to document the process of this emergent design for future initiatives.Methodologically, this included journaling by team members, reflective interviews with all participants at various stages throughout the process, and ongoing reflection on process through the review of meeting notes.As detailed in the developmental moments captured above, this reflexive process contributed to the regrounding of the action plans in participant priorities and the move away from a service provider toolkit to the development of a peer support program.

POWER
'We want to let go of control; at the same time, we want to stay in control' (Olesen & Nordentoft 2013).
Power dynamics are an inevitable component of collaborative work.The research team found it more productive to identify and acknowledge the dynamics of power within the group rather than ignore them.From the outset, the group with the most power in the meetings was the research team, who were leading the meetings, planning the discussion topics and managing the overall execution of the co-creation process.The other actors with considerable power were staff at the partnering immigrant-serving agency who held power as content experts as well as having roles as co-facilitators.The privilege accorded to the facilitators was likely reinforced by the participants' previous experiences at immigrant-serving agencies, either as clients or volunteers.Participants had varying degrees of experience in similar groups, but many were used to being in service-provider/client relationships and this relationship was reproduced through these meetings.This is not to say that the relationship between participants and facilitators was totally predetermined by this dynamic, but just to emphasise that the structure of this relationship did influence the interactions between facilitators and participants.
As discussed above, research participants were invited to engage as 'co-analysts' at our bi-weekly meetings.Their engagement in reviewing data, commenting on themes emerging from the literature review and sharing ideas and feedback allowed for the demystification of the research and analysis process.The goal was to open the role of knowledge creator to all participants and remove some of the privilege from the research team.Our work, however, was also shaped by the institutional parameters that surrounded the research.For example, the meeting location was predetermined, while the timing was somewhat flexible.As described, with the change from toolkit development to a sole focus on the peer support model and meeting with municipal staff, the participants were encouraged to take ownership of the work, but external timelines shaped the pace and nature of that ownership.Therefore, the project was largely successful because of the ability to be flexible and adaptable within the context of these institutional parameters.This meant a degree of flexibility on the part of all participants, including organisational partners, facilitators, researchers and research participants, and recognition through reflexive practice of the limits of our attempts to mitigate the power imbalances we had set out to erase.
There were also power dynamics between participants.Power comes from a variety of locations and subjectivities, and shifts depending on context and relations.At the beginning of the process, it is fair to say that the three participants with higher levels of English language capacity tended to have more power in the group discussions.They occasionally positioned themselves -or were positioned by other participants -as experts on particular issues (for example, other programs available to seniors and retirees in the city).Their superior knowledge of English, as well as having been in Canada longer, provided them with a level of comfort in participating openly in the group.At first, the other three participants were somewhat quieter and more withholding.Over time, however, this dynamic shifted.The more reticent participants became more confident within the group and at times came to dominate the conversation, taking this opportunity to ask numerous questions, occasionally unrelated to the topic at hand.As this group opened up, the women who had previously held dominance were required to take a more active listening role and to be patient as the conversation occasionally veered off course.This shifted the power dynamics somewhat and at some points led to tension or frustration.Such tension was mitigated to a certain degree through the approach taken to acknowledging and working with difference described in the next section.

DIFFERENCE
'The strength of these partnerships depends on the capacity of the leaders to provide a learning and reflexive organizational culture and a participative organizational structure capable of making room for the supplementing, competing or even conflicting agendas embodied in these partnerships' (Strier 2011).
As a collaborative initiative, the project's success was contingent upon the active participation of both institutional partners.As an educational institution and an immigrantserving agency, these partners were differently positioned, with different expertise, organisational structures, and shared -but distinct -project goals.A key learning drawn from the research was the importance of valuing the different contributions that each partner brought to the work.Strier (2011) suggests conceiving the differing perspectives within a group as frames.These frames offer a way of thinking about the different perspectives and experiences that stakeholders and partners bring to the group (Strier, 2011, Gray, 2004).Identifying, acknowledging and, where possible, addressing the differences in frames across a research collaborative allows for more generative and realistic outcomes.From the inception of the project we mapped out the perspectives of each of the three groups (researchers, service providers, community members) and continued to check in with the varying perspectives, attempting to view project priorities and milestones through multiple frames.For example, while the researchers from the college were interested in supporting the research and knowledge generated through this process, the immigrant-serving agency was interested in learning about how to leverage the research and co-creation process into opportunities to improve programming and service delivery to the population they serve (immigrant women).These goals were not mutually exclusive and overlapped in generative ways, allowing both groups to meet their goals.Regardless, it was important to acknowledge these differences and work towards achieving both visions simultaneously.
A learning from the broader findings that related to retired immigrant women's wellbeing was the need to acknowledge the complex intersections that shape the retirement experiences of immigrant women.This meant being attentive to the way class, race, immigration trajectory, work experience, physical ability and English language proficiency (among many other variables) impacted their access to services and supports.Recognising these differences was as significant as staying true to our co-creation approach throughout the project.As the process brought together a diverse group of retired immigrant women, with the aim of having them work together towards shared outcomes, it was not possible to gloss over the differences in life experiences between them.One of the key differences that played out within the group was the different immigration trajectories of the participants.Some of the participants had been in Canada for close to 50 years, others had arrived more recently, within the last decade.Those who had arrived more recently had been sponsored by adult children, while those who arrived decades ago had come as economic immigrants or with their spouses and had raised children in Canada.Those who had come earlier had had a considerably different 'settlement' experience from those who had arrived more recently.This was visible within the group through things like English language proficiency (verbal and written), comfort and experience working in a group setting in English, and knowledge of services and resources available in the community.While all the participants had worked in Canada, the type of work they did -as well as the length of time they had been employed in Canada -varied widely from 40 years down to two years.Some had had professional careers in the oil and gas industry, others had worked in day care.This wide range of personal, immigration and professional experiences -as well as distinct cultural and linguistic backgrounds -meant the group was rich in diversity and experience.It also meant that facilitation, group activities and action planning had to take into account the varying levels of comprehension, experience and comfort within the group.
Participant feedback identified an overall theme: a sense of accomplishment in having seen the project through to completion.One woman commented: 'I am the type of person that I want to see through a project from beginning to end.And I feel a sense of accomplishment if I do that, so with this project I have this sense of accomplishment.That, yes, I was able to contribute.So that's a sense of accomplishment'.Others reflected on their contributions and ideas being valued: 'They made us feel as though we're contributing to something, to some sort of research project which is really nice.They even said that we could look at it after it was finished which was really nice…I liked breaking out into groups and forming an action plan -all sorts of ideas coming from everybody coming together.'For other participants, their enjoyment was shaped simply by the opportunity to 'meet other nice ladies and share stories'.Thus, the motivation for attending and the reasons each continued to participate were shaped by different factors.For some, this was an important opportunity to contribute ideas and possibly inform service delivery and policy; for others it was a valuable social occasion to interact with other women who shared similar experiences.From this feedback, it was important to acknowledge both the diversity in motivation and the need to create opportunities for participants to take ownership of the process -whatever that looks like for them.The feedback from participants in our knowledge exchange event with the City of Calgary was overwhelmingly positive.Participants shared that they appreciated having the opportunity to have their ideas heard and to learn more about the policy-making process.Many of the participants also shared that up until that point their experiences had been as recipients of services and support and rarely had they had the opportunity to work at 'a higher level' (as one participant described it), that is, at the level of crafting programs and policy.
Overall, the project team was successful in adjusting the in-group activities to the different abilities and interests within the group.It was important to manage expectations of the level of ownership that different participants might take within the group.While some participants were more eager to take an active leadership role, others were more interested in the social interaction and camaraderie the group provided.Rather than trying to have everyone at the same level, it was beneficial to create space within the group for different interests and abilities to take shape.Also important was taking time to work together as a group to lay parameters around how discussion could and should work (through the establishment of ground rules), as well as ongoing reflexive conversations to re-ground all team members in the overarching goals of the project -immigrant retirees' wellbeing -despite the differing frames brought to the table.

Conclusion
While 'emergent design', 'participatory research' and 'knowledge co-production' are all distinct concepts, they draw on similar approaches to addressing complex social problems by seeking collaboration between differently positioned stakeholders.In order to be successful, then, collaborative and co-creative work 'requires a constant, on-going investment in increasing trust and face to face knowledge between partners' (Strier 2011).As detailed in the above account of our five developmental moments, challenges that could have derailed the co-creative process arose frequently.However, by intentionally building methods aimed at encouraging reflexivity, recognising and mitigating power relations, and embracing difference into our praxis, we were able to build, and maintain, a level of trust between partners that kept us grounded in our co-creation principles from project inception through to action implementation and evaluation.Therefore, while collaborative work often focuses on finding commonalities between stakeholders, our work with retired immigrant women demonstrates the productive space that can be opened up when the multiple frames and forms of power held by differently positioned stakeholders are acknowledged.Rather than glossing over the complexity created by power and difference, reflexively recognising differing priorities and positions and then working together to find parallel or intersecting interests contributes to a deeper understanding of the challenge at hand and, subsequently, to the development of programming and policy recommendations that are adaptable and responsive to a more diverse set of needs.
Journal of Community Research and Engagement, Vol. 12, No. 1, May 2019 Journal of Community Research and Engagement, Vol. 12, No. 1, May 2019 Journal of Community Research and Engagement, Vol. 12, No. 1, May 2019 Embracing complexity: Co-creation with retired immigrant women Gateways: International Journal of Community Research and Engagement, Vol. 12, No. 1 Journal of Community Research and Engagement, Vol. 12, No. 1, May 2019 Embracing complexity: Co-creation with retired immigrant women Gateways: International Journal of Community Research and Engagement, Vol. 12, No. 1peer support process.Participants from the co-creation group, once trained, would attend My Community, My Home and act as peer mentors to the program's participants.The co-creation group underwent two formal trainings to prepare them for attending My Community, My Home.This included cross-cultural communication training and training on active listening.Following the training, participants from the co-creation group attended My Community, My Home.
Journal of Community Research and Engagement, Vol. 12, No. 1, May 2019 Journal of Community Research and Engagement, Vol. 12, No. 1, May 2019