Service providers, government agencies and other entities gather data on immigration and settlement for myriad reasons. In Canada, newcomers to the country are required to provide personal information to access essential services from community-based organizations and government agencies. Individuals who handle immigration data hold valuable yet under-examined perspectives on these data collection and sharing activities. This study therefore seeks to answer the overarching question: What information practices are prominent in the work of different groups who collect, analyze and steward newcomers' data?
Our interview-based study reports on the practices of individuals supporting immigration and settlement (i.e. settlement service providers, migrant justice activists, immigration researchers, government staff and designers of digital systems and services oriented toward newcomers) through their use of newcomers' data.
A dual narrative and thematic analysis interprets participants' reflections on their information practices and responsibilities, showcasing variation despite their interdependence and shared priorities for newcomers' well-being. We propose the concept of “data care” to draw attention to experiences and tensions inherent in stewarding newcomers' data. This inquiry reveals conflicts over responsibilities, differences in ethical reasoning and the need for multi-stakeholder negotiation.
Findings bring greater clarity to the intricacies of respecting migrants and their privacy. The study contributes to a theoretical lens on information practices in care work by drawing from feminist care ethics and sociotechnical scholarship.
Introduction
In 2025, there are over 281 million international migrants globally, as crises of forced displacement deepen due to climate change, armed conflict and economic collapse (McAuliffe and Oucho, 2024). Responding to shifts in immigration trends, governments and corporations have introduced an array of novel information technologies including “AI lie detectors” at border checkpoints, automated visa processing and algorithmic risk calculation to sort people on the move into categories, and large language model-generated personae of refugee applicants to “train” government officers. These are but a few of the immigration information technologies recently tested or implemented in Canada or the United States (Dastin, 2024; Molnar, 2024; Nalbandian, 2022). These types of technological experiments have led to life-altering mistakes, misrepresentations and misunderstandings in migrants' lives (e.g. Aizeki et al., 2023; Amoore, 2006; Benjamin, 2016; Latonero et al., 2019; Metcalfe and Dencik, 2019; Molnar, 2023; Walia, 2021). How, then, may such harms be prevented? Who is involved in day-to-day handling of newcomers' data? While information researchers have sustained a strong focus on newcomers' information interactions, there has been less attention to the activities of other groups who collect, analyze and steward immigration data.
We seek to analyze how individuals work with sensitive immigration data and make sense of their ethical responsibilities, contributing to a theoretical framework on information practices in care labor. In doing so, we examine the following:
What information practices are prominent in the work of different groups who collect, analyze and steward newcomers' data?
What would they like to learn from newcomers' data?
What are their concerns for how newcomers' data is used?
How do they suggest that newcomers' data should be cared for ethically?
What is their sense of whether their practices are aligned with other groups who work with newcomers' data?
To address the research questions, we sought to learn from individuals who work with immigration data daily through a series of semi-structured interviews, as part of a broader project (Shankar, 2023). Findings indicate opportunities for theoretical development through recognizing conflicts over responsibilities, differences in ethical reasoning and the need for multi-stakeholder negotiation in information practices.
Context
In Canada, the federal government creates contracts with, and funds, non-profit community-based organizations to provide programs for newcomers. Individuals are typically eligible for these programs during the first five years of their arrival in the country. These programs are settlement services, offering newcomers targeted opportunities for civic engagement, resources for job seeking and education, and language training (Caidi et al., 2010; WelcomeBC, 2025). In this context, “newcomers” is a commonly used term which can inclusively refer to migrants of varied statuses, including asylum seekers, undocumented individuals, immigrants and refugees. Newcomers may be asked to provide details such as immigration status, identification documents, reasons for leaving their country of origin and gender identity when seeking services. Staff in community-based organizations or in the federal government agency Immigration, Refugees and Citizenship Canada (IRCC) frequently collect newcomers' data through intake or registration forms for many programs and services. For the purposes of this study, the phrase “newcomers' data” included qualitative and quantitative evidence, such as auditory, material, numerical, textual and visual records of immigration and newcomers' lives.
Literature review
A robust strand of scholarship highlights migrants' interactions with information and service providers (e.g. Allard, 2015; Caidi et al., 2010; Caidi and MacDonald, 2008; Ekmekcioglu, 2024; Lloyd, 2020; Mabi et al., 2023; Ndumu, 2020; Shuva, 2023a; Srinivasan and Pyati, 2007). Researchers have studied libraries' roles in supporting newcomers' settlement (e.g. Dali, 2022; Shuva, 2023b). Some have also investigated political and cultural facets of information practices that influence migrants' lives, examining difficulties associated with missing, incorrect or withheld records such as identification documents (e.g. Sheikh et al., 2023). Scholarship on migrants' information behaviors, interactions and practices is well established, an area which we gesture toward yet do not synthesize due to space constraints. We notice that fewer information researchers have explored the inverse: that is, the network of groups that directly or indirectly influence newcomers' experiences.
There are signs that such a focus is warranted. Vannini et al. (2019a, b) found that humanitarian aid workers in the United States had scarce privacy and cybersecurity guidelines and resources for supporting migrants with precarious legal status. Anwar et al. (2024) emphasized the need for differentiated privacy and cybersecurity instruction for migrants with varied prior knowledge and lived experiences. Guberek et al. (2018) demonstrated that undocumented migrants to the United States experienced uncertainty about the risks to their digital privacy and the consequences of government surveillance, yet often expressed high levels of trust in social media platforms. Needle and Fleischmann (2024) have followed immigration attorneys' difficulties navigating surveillance. Hultin and Introna (2018) observed officers' and asylum seekers' information interactions within government-run reception centers in Sweden and found that discursive and material elements of these interactions cast individuals into subject positions with relative agency. Lamdan (2019) revealed migrants' privacy and security is jeopardized by corporate data brokers and certain institutions, including public libraries, that hold contracts with such brokers. From existing evidence, public and private sector actors' information practices with immigration data appear consequential, although their experiences and understandings of these practices are unclear.
Theoretical framework: information practice and care ethics
Our study is grounded in information practice, particularly the tradition with ties to practice theory (Cavanagh, 2013; Greyson, 2018; Hui et al., 2017; Huizing and Cavanagh, 2011; Ivanov, 2017; Lloyd, 2010; Pilerot et al., 2017). Information practices are actors' routine uses of information and technologies enacted within situated cultures and environments (e.g. Anderson, 2007; Cavanagh, 2013; Díaz Andrade and Doolin, 2018; Dourish and Anderson, 2006; McKenzie, 2003; Nathan, 2009; Olsson and Lloyd, 2017; Rivera and Cox, 2014; Talja et al., 2005). Information practice counters individualistic conceptions of information and technology use. There are, however, still limited theoretical resources for studying information practices shared across social groups. Polkinghorne (2023) has argued that interconnected information practices need to be examined in more depth through vocabulary afforded by practice theory. Many researchers have investigated information practices within bounded community, demographic or professional groups. Cumulatively, this has led to a focus on single or co-located populations with relative consistency (e.g. Gomez, 2016; Lea French and Williamson, 2016; Lloyd, 2007; Olsson, 2016). This leaves open questions about how information practices are negotiated across groups and potential effects of conflict and difference, power relations and labor distributions.
We suggest that care ethics offers one avenue to strengthen some of these undertheorized aspects of information practice. Developed in political science and feminist scholarship, ethics of care is a normative ethical theory (Tronto, 1998, 2013). Care ethics investigates moral reasoning in context, attuning to situations between groups with specific responsibilities and roles. A care lens is particularly useful for exploring actors' reliance on each other to meet their needs. Qualities that constitute care include attentiveness (caring about), which relates to the needs and vulnerabilities of the receivers of care; responsibility (caring for) perceived by carers; competence (care giving) or practical capacity to care; and responsiveness (care receiving) on the part of individuals (Tronto, 1998). More recently, Tronto (2013) added a fifth quality called plurality (caring with others jointly), upheld by communication and efforts for building solidarity amongst groups. The conduct of care is not neutral nor ideal, it can encompass conflicts, discomfort and political or economic motives (Murphy, 2015; Puig de la Bellacasa, 2017). Researchers have employed care ethics in many contexts, studying sociotechnical systems in care labor over the last 2 decades (e.g. Gray and Witt, 2021; Mol et al., 2010).
Care work can be performed for entities such as ecosystems, knowledge, people and social fabrics. Scholars who have examined care work at the core of activities with information and technologies include Anderson and Fourie (2015), Caswell and Cifor (2016), Cushing (2023), Dalmer (2018), Dalmer and Huvila (2020), Gibson et al. (2021), Gibson and Martin (2019), Greyson (2017), Lawson (2004), Liew and Lipscombe (2024) and McKenzie (2003). In these studies, caregivers' and receivers' place, positionality and power play into whose needs are recognized and met, and other needs which may be exacerbated, misunderstood or neglected.
Methodology
Our methodology elicited participants' stories of working with newcomers' data. The first author's work experience prior to this study indicated that actors such as academics and designers of systems played an important role in shaping priorities for data collection by other groups like settlement service providers, while different actors, such as migrant justice groups, for example, were collecting, analyzing and sharing data sets of their own to lobby for change in Canadian immigration policies. The activities of groups beyond service providers appeared important to include. Our sampling strategy sought to expand the focus on a wider range of stakeholders engaging with immigration data and their interconnections, whether individuals worked with newcomers directly or not. We therefore employed purposive sampling to recruit settlement service providers, migrant justice activists, academics working with newcomers' data, government staff and designers who build information systems that are used by newcomers.
Participants were recruited through their public affiliation with an interdisciplinary migration studies network or related organization. The first author held virtual semi-structured interviews with participants from July 2020 to May 2021, each of which were approximately an hour. Interview questions inquired into participants' roles and information practices, objectives with newcomers' data, concerns about data use and perspectives on caring for newcomers' data. During transcription, all personally identifying information was redacted.
Participants were limited to those living within Canada, primarily the province of British Columbia. To maintain participant privacy, we agreed not to report on their organization, location, demographics or tenure in their role, as this could render individuals identifiable within a small-scale region-specific study. Career stages in Table 1 reflect participants' references to their 0–5 years of relevant experience (Early), 6–10 years (Mid) and 11+ years (Established). We acknowledge the limitation that some participants were the only individuals we could reach who held a very specific role. The study was approved by an institutional review board, and all interviewees gave prior informed consent and were offered a $30 honorarium. We provided participants the opportunity to review transcripts and make amendments for accuracy and privacy.
Participants by pseudonym, role and career stage
| Participant | Role | Career stage |
|---|---|---|
| Finn | Settlement service provider | Mid |
| James | Settlement service provider | Mid |
| Kim | Settlement service provider | Established |
| Noor | Settlement service provider | Early |
| Zora | Settlement service provider | Mid |
| Lea | Migrant justice activist | Early |
| Mark | Migrant justice activist | Early |
| Ash | Academic studying migration | Early |
| Cora | Academic studying migration | Established |
| Hamid | Academic studying migration | Established |
| Ira | Designer of information systems that are used by newcomers | Mid |
| Tanis | Designer of information systems that are used by newcomers | Early |
| Valerie | Consultant on the integration of digital systems and technologies in settlement service design and delivery | Established |
| Wendy | Federal government staff member in a leading role | Established |
We sought richness of insights for theoretical sufficiency and conceptual depth, and collected data until we reached this point rather than aiming for representativeness or exhaustiveness (O'Reilly and Parker, 2013; Staller, 2021; Tight, 2023). In idiographic research, participants' distinguishable voices strengthen the credibility and dependability of the findings to bolster trustworthiness (Robinson, 2014; Schwandt et al., 2007).
To honor distinct participant voices, we paid attention to the stories they shared. Stories shape moral reasoning, standpoints and sensemaking about information practices (Frank, 2010; Kaczmarek, 2023). Adapting Frank's dialogical narrative analysis, we explored stories participants told during interviews in response to any of the interview questions. As a first step, narrative analysis allowed us to zoom in on individuals' accounts of their practices. As a second step, we employed reflexive thematic analysis to zoom out on their practices (Braun and Clarke, 2021a, b; Nicolini, 2009). We conducted three inductive, iterative rounds of coding and theme development, through which we carried out immersive line-level analyses of overt descriptions and implicit ideas in interview transcripts in constant comparison across transcripts and code memos. This process allowed us to identify gaps, overlaps and alignment with evidence as per Braun and Clarke. We held regular debriefings, cross-checked findings during and after this process with alternative interpretations, and sought feedback on interpretations from scholars at a distance from the project, guided by qualitative standards for trustworthiness (Schwandt et al., 2007; Shenton, 2004).
Findings and discussion
Each participant reported working with data collected from, or about, newcomers to Canada in the course of their daily work. Their uses of newcomer data were specific to routine tasks and respective roles. We offer details for clarity below, yet we do not claim generalizability. Service providers and designers of systems collected, analyzed, managed and shared data on newcomers accessing services (e.g. collecting personally identifiable information on forms for intake and eligibility screening; documenting interactions, assessments and referrals; sharing data with funders; conducting internal evaluations of services delivered to newcomer clients; telemetry of newcomers' interactions with apps and websites to gauge user experience). Academics worked with much of the same data when they collaborated with service providers on immigration research, and used a broad array of other sources (e.g. fieldnotes, observations, interviews, social media data sets, surveys, statistics provided by government agencies, linked data on health, tax returns and service usage). Activists' interactions with data depended on their campaign and audiences (e.g. collecting survey data on newcomers' experiences of discrimination; analyzing institutional data sources such as policing agencies' records; sharing and visualizing data on discrimination to inform journalists, community members and government agencies). The government staff member analyzed, managed and shared data collected from service providers, designers and academics (e.g. settlement service delivery and evaluation data; researchers' data on newcomers' economic, health and other outcomes). While specific data sources and tasks were unsurprising, participants commonly discussed the prominence of ethically challenging labor, underrecognized in the literature to date. In this paper, we therefore turn our attention to what we propose constitutes “data care.”
Our findings are organized by two themes and associated sub-themes signposted in italics. We interweave stories shared by participants in extended quotes throughout the findings. The first theme is data care, as it is shaped by the second theme, labor conditions. Figure 1 features themes and sub-themes.
The concept map shows two text boxes arranged horizontally at the center. The text boxes are labeled as follows: “Data care” on the left and “Labor conditions” on the right. From “Data care”, four text boxes branch out on the left, each representing a sub-theme. The four text boxes are arranged vertically and are labeled, from top to bottom, as follows: “Demands and expectations”, “Questions of credibility and data quality”, “Inequitable dynamics”, “Valuing labor with data”. From “Labor conditions”, four text boxes branch out on the right, each representing a sub-theme. The four text boxes are arranged vertically and are labeled, from top to bottom, as follows: “Distinctions of data care labor”, “Boundaries and connections”, “Differential access to data, infrastructure, and resources”, “Challenges and conflicts”.The figure above is a map of the two themes and supporting sub-themes identified through reflexive thematic analysis (Braun and Clarke, 2021b). Figure by authors
The concept map shows two text boxes arranged horizontally at the center. The text boxes are labeled as follows: “Data care” on the left and “Labor conditions” on the right. From “Data care”, four text boxes branch out on the left, each representing a sub-theme. The four text boxes are arranged vertically and are labeled, from top to bottom, as follows: “Demands and expectations”, “Questions of credibility and data quality”, “Inequitable dynamics”, “Valuing labor with data”. From “Labor conditions”, four text boxes branch out on the right, each representing a sub-theme. The four text boxes are arranged vertically and are labeled, from top to bottom, as follows: “Distinctions of data care labor”, “Boundaries and connections”, “Differential access to data, infrastructure, and resources”, “Challenges and conflicts”.The figure above is a map of the two themes and supporting sub-themes identified through reflexive thematic analysis (Braun and Clarke, 2021b). Figure by authors
Data care as information practice
Participants' activities could enact care for data on behalf of newcomers. We conceptualize participants' information practices with immigration data as data care because they align well with elements of attentiveness, responsibility, competence, responsiveness and plurality as theorized in ethics of care scholarship (Puig de la Bellacasa, 2017; Tronto, 2013). Data care included activities at each phase along the data life cycle, encompassing collection, creation, access, ownership, stewardship, management, analysis, reuse, sharing, retention and destruction of data. Some engaged in most or all stages of a typical data life cycle from collection through destruction or re-use, while other participants said they relied on others for activities they did not perform (e.g. James, Kim, Lea, Mark).
Although each participant said that attention to ethical handling of immigration data is essential, they expressed ambivalence about demands and expectations of performing data care (Cora, Hamid, Noor, Tanis, Valerie, Wendy, Zora). Data care could be a “burden” (Lea and Mark), “duty” (Hamid), or long-term responsibility (Ash, Noor, Tanis). Participants identified rising demands for data collected from newcomers with mixed feelings. Hamid reflected:
Ten years ago, there weren't as many people working with forced migrants […] There is a lot of interest in it. It is very politicized, and because of this I think that the situation has changed for the people with whom we're working, who also get asked by many other people – other researchers, journalists, government workers, the whole range of groups […]. There is also a general information fatigue. People are less inclined to want to speak with you because they have already spoken to other[s].
Given higher volumes of data collection, several participants emphasized mounting obligations to imagine implications of data use or misuse and unintended consequences (Hamid, Lea, Mark, Zora). Some expressed concern that they were not trained for the most difficult matters of data care. Cora said:
I’ve never really been able to sort this [ethical matter of immigration data] through because of the intricacy of the issues involved. For example, what if group x has a higher rate of spousal violence and group x is a highly marginalized group? […] Certain kinds of people will suffer if that information is made available, and other kinds of people will suffer if that information isn’t made available. There’s an ethical trade-off if ever there was one. I am not equipped personally to sift through those kinds of things, because I know those problems exist, but I’ve never studied [them], I’ve never gained expertise on how to deal with that.
Cora reported that the work of ethical decision making with immigration data was not often explicitly trained nor evaluated. Participants said expectations of their own and others' involvement in ethical decision making could be perplexing, especially when individuals share some of the same data sets and information systems yet exercise different ethical perspectives.
Others noted that newcomers' data should be stewarded with a responsiveness to plurality in cultural norms (Ash, Ira, Noor, Tanis, Zora). Noor reflected on collecting data from newcomers while conducting service design and evaluation research. She was attentive to newcomers' agency in the informed consent process:
We had a form that people read in their first language to understand what we were doing there. And they signed those forms and consented to it. […] I think there are so many complex things happening with power dynamics, and cultural norms, that maybe from our cultural lens we would read, “Well that’s agency, they wanted to tell us stuff” but from their cultural lens I really reflect back on interviews we did over two years ago with Syrians, and see “Oh, of course they would never say no to sharing that information with us, because that is so rude in their culture” to say no and set boundaries and ask people to leave after we’ve been in their house for five hours, it won’t happen.
Noor's story reveals her process of learning about differences in cultural norms that inflect information practices, such as hospitality, politeness and privacy. Here, caretaking is culturally specific. One can exercise care for an individual's data by appreciating the context for an interpersonal interaction in which it is shared. Noor's analysis of constraints on newcomers' agency parallels that in the literature, put forward by scholars such as Díaz Andrade and Doolin (2016, 2018), Gilliland (2017), Hultin and Introna (2018) and Kaurin (2019).
Noor and other participants were attentive to the fact that their demands and expectations for newcomer data could indirectly affect newcomers. For example, an academic named Ash said that through a small gesture such as liking or commenting on a migrant informant's social media post, they realized they could inadvertently make them vulnerable to state surveillance in their countries of origin. Our analysis offers evidence that groups influence each other to varying degrees through data care. We take an interest in this interdependence, seeking to move away from individualistic logics of responsibility for uses of data and technologies. Interdependence is a state of interconnection within networks of responsibility (Puig de la Bellacasa, 2017; Tronto, 2013). Though groups involved in or affected by data care may be interdependent, there is a need to recognize power relations between them. In addition to the interdependence of researchers and participants, links were evident between groups who performed interdependent parts of the life cycle of immigration data.
As referenced earlier, participants from different groups articulated questions of credibility and data quality when they judged others' work with immigration data (Cora, Finn, Hamid, Ira, James, Lea, Mark, Noor, Tanis, Valerie). Lea and Mark were both community-based volunteer activists working on anti-discrimination initiatives with Asian newcomers. They said that other groups working with data on migration and discrimination, particularly journalists and law enforcement agencies, claimed epistemic authority over activists. Lea put it this way:
We have definitely had issues with journalists who have said, “Well we can’t fact check your information, therefore it is going to be a problem for my editor,” or something to that effect. Or, what else have we heard – “This information needs to be made publicly available, if it’s to be considered credible.” […] It is very interesting when [the police department] says “x number of anti-Asian incidents happen in [our city],” then the media say “Oh, great, [our city’s] police say this many incidents increased by this much percent.” But when a community group that also collected incidents in a way that is more accessible to community then says, well we collected double the number of incidents, that information is not seen as valid. […] Or people have asked, “Please send us the data so we can see if there are other trends than the ones you've identified.” Why? Why do you get to decide that? These are very odd, and sometimes people don't reply at all, they just get kind of upset.
By exercising limits on data sharing, Lea and Mark resisted journalists' and other groups' doubts about their epistemic authority as analysts and stewards of their community's data. In this way, co-participants in data care exercised epistemic authority by judging each other's practices, justifying grounds for their own expertise, and defining or deconstructing potential media and policy narratives of immigration. This issue parallels what Crooks and Currie (2021) have called “agonistic data practices” whereby community groups intentionally contest the data and interpretative frames of institutional actors such as law enforcement or government agencies, a strategy of resistance also found by Muñiz (2022).
Other participants similarly identified inequitable dynamics between themselves and other groups handling immigration data (Ash, Kim, Mark, Zora). As an academic, Ash said they find settlement service providers are burdened with the labor of caring for data:
Most of the burden is on the settlement organizations, not on the government. Many settlement organizations in BC and in other provinces, they have all these problems that they don't know how to tackle. […] So, the main problem is that the government is not responsible for it, but the settlement organizations are responsible for it.
Demands and expectations for the collection, analysis and sharing of immigration data were perceived as bringing benefits or burdens to multiple actors involved, distributed differentially between them. Multiple participants pointed toward a bespoke Canadian information system, called the Immigration Contribution Agreement Reporting Environment (iCare), which they saw as problematic. iCare is a data entry system used to record interactions between newcomers and service providers within settlement service organizations. The same system is also employed by the federal government to attempt to track outcomes of settlement services, although alternative and future uses of the data remain unclear from publicly accessible government statements about the system (e.g. CIC Research and Evaluation, 2015). James, a service provider, was concerned with the lack of control he has over the means of producing immigration data through the iCare system:
We turn our information into iCare, but we are not able to generate any reports. It is all managed by IRCC, so it is not useful for us. […] I am not sure how useful it is for the people who use it, who manage, and I know that it continues to grow. Previously we used to enter our narrative reports and our annual reports on it, but recently it has been changed because they heard from the sector how onerous it was to have so many reports that needed to be entered. Now they have sent us this year a PDF document that we need to fill out instead. It is not useful for us, because we can't generate reports from it.
Features of the information technologies individuals use appear to condition their data care labor. Participants discussed working with immigration data through data entry systems such as iCare, relational database management systems, enterprise software such as the Microsoft suite or proprietary third party tools (e.g. Facebook, Google Forms, WhatsApp). Few participants reported working directly with automated or algorithmic technologies. Yet, two migration researchers described their colleagues' design of such “AI” tools. Hamid, for example, told a story about problems he saw with a predictive system another scholar had designed. Hamid said:
They were trying to predict the next areas of mass migration. The danger in that approach is that governments could actually use it to then try to prevent those mass migrations from happening, but not in a way that is trying to resolve the underlaying issues of human rights or economic inequalities, but more in a kind of fortress Canada or fortress Europe perspective in the sense that if we know the next wave is going to come from this country, let's already start to put in place extra-territorial and externalization mechanisms so that individuals cannot get out of that country in order to claim asylum in our country.
Hamid flagged several risks he was particularly worried about: the forced pushback of people on the move to zones in which they face threats to their freedom or their life – prohibited under the 1951 United Nations Refugee Convention – in addition to discrimination, and distorted data. These problems go beyond privacy, echoing the concerns of other experts (e.g. Achiume, 2020; Maitland, 2018; Metcalfe and Dencik, 2019; Molnar, 2024). The wave of attention to artificial intelligence has grown in the period since we interviewed participants. Government agencies' and technology corporations' information practices with automated and algorithmic technologies will need to be more closely monitored going forward, particularly in the realms of immigration research and service provision.
Participants suggested that the availability of funding, staffing and training for work with data indicates how this labor is valued within their local organizational contexts. Participants expressed views of abundance and scarcity for funding, staffing and training that affected valuing labor with data (Ash, Finn, Hamid, Kim, Lea, Mark, Wendy, Zora). Whereas abundance could support valuing of this labor, relative scarcity could devalue and detract from this labor. Academics typically expressed their experiences of abundance, as well as their capabilities and uses of data given their ability to apply for grants and determine the direction of their research on immigration and settlement (Ash, Cora, Hamid). Others, most notably service providers, were quick to contextualize their information practices within the strictures of funders, IRCC policy and limitations on resources (Finn, James, Kim, Noor, Zora). Service providers also discussed experiencing alienation from interpersonal aspects of welcoming and interacting with migrants as individuals, not just data subjects with whom they had to fill in forms. Some talked about the challenge of having to neglect friendly conversation because of demands for collecting data. The same individuals discussed frustration over the limited control they had in the means of producing immigration data (e.g. Noor, Zora).
Kim pointed out that service providers might seek to measure different indicators than funders. Prioritizing evidence of migrants' well-being, while also needing to fulfill demands for evidence of return on “investment” valued by Wendy at the federal government level, added to the complexity of this multi-group information practice:
We collect our own data for the different indicators that we see as valued, that we see as success, so we are trying to collect data to measure self-efficacy and hope, and things that are important that funders do not think are important. How does that interplay with all the other indicators that we’re collecting information for?
Service providers and migrant justice activists consistently expressed frustration that their information practices were beholden to the interests of better funded or resourced entities and institutions such as academics and government agencies.
Labor conditions of data care
Participants reported very different contextual factors that we see as labor conditions shaping data care. We employ the term labor to illuminate data care as a critically important information practice that requires skillful action and specific supports, informed by prior work on human labor with information systems and technologies (e.g. Ekbia and Nardi, 2017; Warner, 2005). When immigration data care is interpreted as labor, variations in its conditions become more noticeable. Labor conditions shaping data care practices included differential access to data, infrastructures, and resources, boundaries and connections between actors, and conflicts and challenges which complicate cross-group practices. Together, these circumstances influenced how capably actors said they could perform their information practices.
Individuals across roles made demands and had expectations of one another, creating or reinforcing divisions of data care labor (Ash, Cora, Finn, Hamid, Ira, James, Kim, Noor, Tanis, Valerie, Wendy, Zora) and boundaries and connections between them (Finn, Hamid, Ira, James, Noor, Tanis, Zora). These two sub-themes are interrelated because divisions of labor could exert a boundary or a connection between groups. Academics and government staff, for instance, typically relied on settlement service providers to collect data from newcomers in the course of everyday service provision (e.g. Finn, Wendy, Zora). Settlement service providers indicated that the iCare system operates on a one-way model of data entry, so that data are entered into the system which is managed by the federal government. Thus, although they collected immigration data early on in the data life cycle and these activities were essential for other groups' work with it, service providers consistently articulated challenges with accessing and using it after collection. Wendy, who is a Canadian federal government staff member in a leading role on immigration and settlement, explained how this system was intended to operate:
There are over 600 of them [service provider organizations across the country] who enter data into the settlement database. Settlement data is so important because that is the only data we have to make sense of whether the settlement investment – about $1 billion a year – that the department makes: Is it really helping, is it creating results? To answer that question more systematically or scientifically, not anecdotally, one needs solid data. The iCare data is the core of it […] But the problem is that the service provider organizations themselves do not have access to the data that they enter. The only thing that they have access to is a monthly static PDF report that we send back to them, which is useless. It has been like this for many years, and we have not really been able to overcome that big barrier yet. We are working on a modernization project right now to aim for that access, which will definitely increase the capacity of the organizations who enter the data, we owe it to them.
This story of the iCare system illustrates differential access to data, infrastructures and resources (Ash, Cora, Finn, Hamid, James, Kim, Noor, Valerie, Wendy, Zora). Wendy noted that government agencies and funders tend to have the most power to organize information practices with immigration data. Another participant, Valerie, who advises organizations in the settlement sector on their use of information systems and technology, pointed out the mismatch between her expectations and actual experiences of using iCare:
The promise of iCare was that [it would be] two-way – give us the data and we'll take care of analyzing it and sending it back out to you. People have complained for many years that the two-way has never happened. They submit data, IRCC does run reports, and does some research on it that they put out. But the data does not go back …. That promise has never really happened. The onus is still on organizations to try to figure out how to mine that data. Because of that initial promise, I hear that people didn't feel the need to invest nor were they given the resources to invest in data analytics, because it was the promise of iCare that was going to do that.
Many demands on service providers were enforced through the iCare system, which Valerie said was designed to collect certain streams of immigration data from service providers. A service provider, Kim, said that data work was a burden for their organization:
It is administratively burdensome, especially for organizations that do not have CRMs [customer or client relationship management systems] as well. We spend a lot of time having to aggregate data manually and to then feed it back. What makes it really frustrating and demotivating is that the data we produce then doesn't tell the full story of what we are doing and what community members are going through, and speak to the systemic issues or assets.
While required, the outcomes of the labor were not seen to be very useful to Kim because the data that they wanted to analyze was not being collected and analyzed in the first place. In this way, participants working in varied roles suggested that the iCare system is exemplary of information asymmetries in the benefits and burdens of labor with immigration data, contributing to conflicts and challenges (Cora, Hamid, Ira, Kim, Lea, Mark, Noor, Tanis, Valerie and Wendy).
Dominant groups, like funders, government agencies, police forces, larger and better resourced service organizations, appeared to impose demands on other groups for newcomers' data to be delivered or evaluated on their terms (Hamid, Ira, Lea, Mark, Noor, Tanis). In their role as a service designer who works with settlement organizations, Tanis told a story of one such incident:
We wanted to know what culture [newcomers] were from. And what language they speak. […] But other than that, from a UX [user experience] design perspective, there has been a lot of problematic data collection. But I think the best practices now are don't take data you are not going to use. The less data you have, the more secure it is – because there is less chance for misuse. So, we were taking a very data-light perspective to data collection, and we started partnering with the [management and staff] at [a settlement service organization …] So, registration became a big sore spot, or a pain point between us and the team. And they wanted to collect PR (permanent resident identification) numbers to get them registered, for IRCC to count the numbers. […] They know the data that they usually collect and didn't have any reflection on what data it is that they really needed. […] We didn't see that data as necessary to the mission of what we were asking, but that is not really how the other team saw it and it was not the hill we were going to die on. So, we did let them take it.
Tanis expressed a concern for over-collection of data, conflicts in professional norms and tension between groups' information practices. Differences in ideas of data care appear irreconcilable in this account.
Many conflicts centered around the unevenly distributed benefits and burdens among groups. Some participants said that the collection of immigration data was valuable in and of itself, regardless of how it was used. The motivation of data collection for the sake of immigration data becoming a “corporate asset” (Wendy) would benefit few groups, primarily IRCC and academics. Powerful, well-resourced stakeholders may shape practices with immigration data, advancing their own interests and narratives. Ekmekcioglu's (2024) findings on accountability mechanisms in the Canadian settlement sector are relevant to our study in this regard, especially insights on hierarchical pressures at times exerted by funders on service providers to collect data for reporting purposes, which could detract from social interactions between frontline workers and migrant clients.
On-the-ground matters of data stewardship appear complicated and subject to disciplinary and role differences (Huvila and Sinnamon, 2024; Oliver et al., 2023; Pasquetto et al., 2024). Our study adds evidence that experiential elements of performing data stewardship are well worth investigating further. Aiming to steer away from under-specified or universalistic claims of data ethics, we found that an intersectional feminist ethics of care can support reflexive yet pluralistic inquiries into information practices, such as data stewardship, which influence community well-being (e.g. Garcia et al., 2022; Mohanty, 2006; Mol et al., 2010; Puig de la Bellacasa, 2017; Tronto, 2013). Similarly, we look to work on the CARE Principles for Indigenous data governance (e.g. Carroll et al., 2020; Taitingfong et al., 2024) and scholarship on knowledge management in the Māori context that employs constructs of relationality, responsibility and respect from care ethics to widen norms of stewardship (Liew and Lipscombe, 2024).
Our proposed concept of data care emphasizes peoples' experiences of stewarding data, recognizing that the existence of data stewardship policies or principles is not necessarily sufficient to bring about their enactment. Data care acknowledges factors of agency and power, material affordances and constraints, tensions between multiple stakeholders' roles and responsibilities, as well as differences in moral reasoning, each of which we found in the performance of information practices with immigration data.
Below, we articulate contributions of this study to theoretical refinement of information practice.
Information practice theory contributions
This study inherits a great deal from information practice research. In line with this tradition, it identifies an information practice within wider complexes or layers of social practices (e.g. Cox, 2013; Polkinghorne, 2023). Findings, however, complicate some methodological approaches and notions of practice as a unit of analysis. Researchers of information practice have tended to sample from participants who are co-located, and share community or group norms, professions or cultures of activity (e.g. Lloyd, 2007; Nathan, 2012; Olsson, 2016; Savolainen, 2008). No single role or agency is responsible for immigration data in Canada. Rather, a web of actors engages in data care. Managers depend on frontline staff to provide up-to-date data on clients served by the organization. Federal government staff and the immigration agency (IRCC) need service providers to enter data into their databases. Academics depend on migrants or service providers to be able to collect, create and steward data, while crafting research ethics protocols to guide these relationships (e.g. Clark-Kazak, 2017). To continue to exist, service providers rely on funders for material resourcing and are required to provide data to justify their spending. While information practice scholars have made leaps in nurturing a sociocultural perspective to move away from the dominance of cognitivism and individualism, there are further steps we can take. Attending to the influences of multi-group participation is a move in this direction.
Varied ethical and political views, professional backgrounds and cultural norms can co-exist in an information practice, even if not comfortably. Immigration data is used in settlement service provision, migrant justice activism, academic research in migration studies, design of digital technologies for newcomers, as well as immigration policymaking and analysis. In part because of this, there is a plurality in what groups seek to do with immigration data. Groups share some of the same means (i.e. data sets, information systems) although they pursue dissimilar ends. Some service providers did not regard the federal government's uses of data within interpretive frames of return on “investment” (Wendy) as ecologically valid or epistemically just. Information practices manifest differences in what groups care about, as well as how they seek to perform care for communities and their data.
This inquiry also makes contributions to information practice theory by drawing on the constructs of feminist care ethics. We found a care lens especially useful for exploring interdependence where actors rely on each other's labor to meet their needs to use immigration data for evidence, to seek funding and to safeguard privacy. A care lens sensitized us to varied needs, power dynamics in who can have their needs met, and conflicts over how care should be performed in the interests of newcomers' well-being. There is strong potential in continuing to better understand data or information work as care work by leveraging existing vocabularies from critical theory in concert with information practice theory.
Limitations
Conducting this study during the early lockdown stages of the COVID-19 pandemic influenced its research design. Other methodological factors also informed findings. Participants' role and tenure in their position varied. We note that this contributed to understandings of data care which differed. Collecting further information on participants' backgrounds and organizational contexts could have been useful for understanding underlaying reasons for these differences, yet was not feasible given priorities for preserving participant confidentiality. In addition, observational data would have allowed analyses of everyday activities embedded within task performance, rather than relying on recollection. Furthermore, geographic concentration of participants in Western Canada influenced findings because provinces have different immigration and settlement service agencies and models, and other contextual factors such as privacy legislation vary.
Future work
Practices with precarious communities' data is instrumental in their ability to survive and thrive (Benjamin, 2016; Costanza-Chock, 2018; D'Ignazio and Klein, 2020; Eubanks, 2018); however, there have been few studies of practices with newcomer community data to date. We see the need for inquiries into data care with other stakeholder groups and in different geopolitical settings beyond the Canadian nation state. Further case studies would be theoretically useful for developing and extending the concept of data care, whether in immigration or other domains in which multiple groups negotiate ethical responsibilities while collecting, analyzing and using data of groups in vulnerable conditions.
Practical implications of this study point to opportunities for a wide variety of stakeholders – from government staff to researchers, service designers and service providers – to engage in stronger coordination of shared data care labor through dialogues and joint decision making. Findings from this inquiry also seed questions for information practice and migration studies: What are newcomers' expectations for other stakeholders' handling of their data? How might newcomers' visions of care inform the activities of other stakeholders handling their data, as well as information policies governing immigration data and technologies? We see opportunities for participatory research with newcomer communities to examine their perspectives on information practices with immigration data, as well as the effects of these practices in newcomers' everyday lives.
Conclusion
Within information practices of collecting, analyzing and stewarding newcomer communities' data, we identified frictions between participants' ideals and experiences. In this article, we propose the concept of data care to describe the ethically intensive labor we found embedded in handling newcomers' data (RQ1). This concept aims to bring ideas from information practice and care ethics into conversation for future theoretical development. Our study highlights fraught expectations about what can be gleaned from immigration data sets and their value for decision making, which at times surfaced tensions between groups (RQ2). Individuals acting as data stewards were concerned about deterring newcomers from accessing services by asking for too much data, adding to or hiding vulnerable conditions, data misuse and a lack of awareness of newcomers' distinct needs and preferences (RQ3). Participants reported that care can involve limiting the amount of data collected by default, communicating about cross-group duties and responsibility sharing (or gaps), and better understanding intended outcomes and unintended consequences of shared information practices (RQ4). Data stewards struggled with confusion, competition and compromise when their conceptions of data care conflicted with those of other groups (RQ5). Accounts of cross-stakeholder coordination were uncommon, suggesting a need to encourage dialogues among groups involved in or affected by data care.
A burgeoning array of stakeholders and sociotechnical systems inform decisions about immigration in Canada and other parts of the world. These information practices require continued attention. For newcomers, there are life-wide ramifications of others' information practices with their data. Living in a time of compounding displacement, we find that ways of knowing, learning and caring about people on the move are evermore crucial.
I would like to thank my co-author and advisor, Lisa Nathan, for her guidance and support throughout the dissertation research on which this paper is based. I am also grateful to Sean Lauer, Eric Meyers and Heather O'Brien for their contributions as committee members. We appreciate the participants who shared their insights on care. We also thank the anonymous reviewers for their time and for offering feedback on this work.
