ETHICAL BOUNDARY WORK IN CITIZEN SCIENCE:

The concept of boundary work (Gieryn 1983, 1999) has been developed to capture the ways in which scientists collectively defend and demarcate their intellectual territories. This article applies the concept of boundary work to the ethical realm and investigates the ethical boundary work performed by researchers in the field of citizen science (CS) through a literature review and by analysing accounts of ethics presented in CS literature.


Introduction & background
The participation of non-professional contributors in the production of scientific knowledge has been undergoing a process of institutionalisation during the last decade (Macq et al, 2020).
Under the banner of citizen science (CS), this includes recognition and expectations from national and international policy makers; evaluation and standardisation of methods and technology; the launching of a specialised journal; establishment of national and international associations, conferences, courses; and specified calls for CS from national and international funders.
In fact, this development resembles many of the aspects Kuhn (1962) associated with the development of normal science. Individual projects have been showcased as exemplars that successfully mobilise volunteer contributors, make large-scale observations, and explore empirical materials beyond earlier possibilities of science (Follett & Strezov, 2015;Kullenberg & Kasperowski, 2016). CS has also been recognised by national and international institutions such as the United Nations (UN) as important for complementing and improving data for monitoring the sustainable development goals (Fritz et al., 2019;Fraisl et al., 2020).

This process inevitably involves discussions and controversies over
what can be regarded as CS in struggles over credibility, resources, participation, and truth (Riesch, 2010, p. 454;Prainsack, 2014, p. 147). Some argue that CS is "redefining or even disassembling boundaries", but also that "a baseline of mutual understanding of what is meant by the term 'citizen science'" is necessary for the development of the field (Hecker et al., 2019, p. 1). While CS is held to dissolve boundaries between professional scientists and volunteer contributors, the structuring of participation in CS has also been shown to limit inclusion, recreating borders between scientists and volunteer contributors (Hagen, 2020).
The ethical issues of CS are an important part of discussions and controversies in the field of CS. There is little controversy over issues, such as securing anonymity of participants (Woods et al., 2015, p. 25;cf. Cinderby, 2010, p. 249;Blatt, 2015, p. 102). However, there is less consensus on whether existing ethical frameworks are sufficient, as "data collection [left] to citizen scientists involves a risk that ethical requirements may not be taken care of" (Svendsen, 2018, p. 155). Many organisations, platforms, and national portals initiate discussions on ethics, including the journal Citizen Science: Theory & Practice with a special issue in 2019 (Citizen Science: Theory and Practice, 4:1, 2019), and there is a growing exploratory body of academic literature on ethical issues of CS (cf. Resnik & Kennedy, 2010;Riesch & Potter, 2014;Resnik, Elliott & Miller, 2015;Purdam, 2014;Chesser et al., 2019;Tanginiene, 2019;Fiske et al., 2019;Rasmusen, 2019;Vohland et al., 2021). The European Citizen Science Association (ECSA) provides ten guidelines for CS, and when ethics is specifically addressed, "leaders" of projects are pointed out as ethically responsible: 10. The leaders of citizen science projects take into consideration legal and ethical issues surrounding copyright, intellectual property, data sharing, confidentiality, attribution, and the environmental impact of any activities. (https://ecsa.citizen-science. net/wp-content/uploads/2020/02/ecsa_ten_principles_of_ citizen_science.pdf Accessed 20201229).

Purpose and research questions
Against this background, utilising a review of research on CS, the purpose of this paper is to discern how ethical boundary work occurs in CS (Gieryn, 1999;Wainwright et al., 2009Wainwright et al., , 2006. Following Gieryn (1983, 1999, Wainwright et al. (2009Wainwright et al. ( , 2006, Hobson-West (2012), and Frith et al. (2011), we argue that ethical guidelines are socially and culturally enacted. There is no essence to ethical standpoints, only ongoing ethical boundary work, which strives for demarcation lines in CS, where actors and practices are constructed as ethical or unethical. Ethical boundary work creates controversies between scientists. What are the ethical issues in CS, and how are these to be managed? To attain our aim, we will turn to peer-reviewed publications in the field of CS with three questions: • What ethical issues are identified and debated in the scientific literature on CS?
• How are these ethical issues proposed to be managed?
• What ethical boundary work can be identified in the citizen science literature?
The paper starts with an introduction to the concept and the different strategies of ethical boundary work, followed by a brief description of the field of CS, and the specific form of CS that is the focus of the paper. Thereafter, we explain the method used to search for relevant publications for the review and subsequent qualitative content analysis. This is followed by a presentation of our findings in accordance with our research questions, and a concluding discussion of ethical boundary work performed in the CS literature.

NJSTS vol 9 issue 2 2021
Ethical boundary work in citizen science

15
Theoretical perspective: Boundary work and ethical boundary work Thomas Gieryn (1983, 1999 introduced the concept of boundary work in Science and Technology Studies (STS) to understand the ways in which scientists collectively defend and demarcate their intellectual territories. According to Gieryn (1999), boundary work entails […] the discursive attribution of selected qualities to scientists, scientific methods and scientific claims for the purpose of drawing a rhetorical boundary between science and some less authoritative residual 'non-science' (Gieryn, 1999, pp. 4-5).
The view expressed by Gieryn is that there is no essential definition of "science", only an ongoing debate and rhetorical standpoints among actors to create boundaries that include certain practices or actors, while excluding others. This includes different strategies: expulsion, expansion, and autonomy to create boundaries separating science from non-science to maintain the authority and credibility of scientists (Gieryn, 1999, p. 16-17).
The first of these, expulsion, revolves around demarcations of "real" science from other knowledge-producing activities judged as nonscientific (pseudoscience, amateur science, deviant or fraudulent science, etc.). Here, boundary work becomes a means of social control, denying epistemic authority to those actors who are seen as not belonging to the established cultural boundaries of science (Gieryn, 1999, p. 16). Boundary work also takes place when scientists try to expand their epistemic authority over an ontological domain that is also under influence from rival authorities (for example, lay or local knowledge), who might try to deny science an exclusive right to interpret and to act within this contested domain (Gieryn, 1999, pp. 16-17).
The last of Gieryn's strategies, autonomy, concerns strategies to defend science against efforts from outside actors to exploit the epistemic authority of science in such a way that it compromises the material and symbolic resources of scientists inside the cultural boundaries of science. Here, scientists create boundaries in order to retain sovereignty and autonomy over the selection of scientific problems or the standards used to evaluate research and knowledge-production (Gieryn, 1999, p. 7). To Gieryn (1983), scientists are central to performing and being affected by boundary work. Consequently, as CS is not performed by scientists alone and envisioned to include citizens in several aspects of the research process (Hetland, 2020), this might question the analytical reach work does not necessarily expel non-science (ethics) from science but rather assimilates or even privileges non-science as part of maintaining the image of science (Wainwright et al., 2009, p. 735). is also employed by Frith et al. (2011), who uses the concept to investigate how infertility clinicians construct boundaries between ethical and non-ethical issues in their everyday practice. A similar picture emerges from this study, namely that ethics is an area that is "managed" through the construction of boundaries that differentiate between ethical and non-ethical (Frith et al., 2011).
In the present study, we intend to build upon the implications from these previous studies that have applied the notion of ethical boundary work, specifically how CS and scientists active in this field make distinctions of what counts as ethical issues or not, as well as how these issues are to be resolved (cf. Hobson-West, 2012, p. 661).
The field of CS allows us to extend or modulate the perspectives given by these previous studies further.
First, the field of CS is still in its constitutive phase, especially in relation to the presence of formal ethical frameworks that regulate the field. This situation enables us to apply the notion of ethical boundary work on a context, an emerging scientific practice/field, which differs from the previous studies accounted for above, where debates and controversies around ethics are more established. Second, by its methods and empirical material (see below), the present study differs from the studies mentioned above, which have all utilised interviews in collecting their empirical material. By combining a literature review with qualitative content analysis, this study intends to take a more general view 1 Search string: TITLE-ABS-KEY ( ethic* ) AND ( TITLE-ABS-KEY ( "citizen social science" OR "citizen science" OR "Community-based monitor*" OR "Community-based envir*" OR "civic science" OR "crowd science" OR "civic technoscience" OR "community based auditing" OR "community environmental policing" OR "citizen observatories" OR "participatory science" OR "volunteer monitoring" OR "volunteered geographic information" OR "volun* GIS" OR "neogeography" OR "participatory GIS" OR "street science" OR ( "participatory monitoring" AND "science" ) OR ( "participatory sensing" AND "science" ) OR "public participation in scientific research" OR "locally based monitoring" OR "volunteer based monitoring" ) ) than previous research that have applied ethical boundary work on various contexts. Consequently, by combining the notion of ethical boundary work with our chosen methodology, this study explores the emergence of what is, and what is not, considered ethically legitimate research within a scientific practice that is in a formative phase.

Forms of citizen science
Citizen science (CS), in the form of projects in which scientists recruit members of the public, has recently been creating very visible imaginaries and expectations for an array of such topdown participatory research initiatives (Bonney, 1996;Irwin, 1995;Kasperowski & Kullenberg, 2019;Strasser et al., 2019). However, CS is not one thing, and STS researchers have extensively

Methods and material
This study departs from a search of published, peer-reviewed, papers in the field of CS, motivated by the importance of scientific publications in the formation of a field, and the ability to identify the range of ethical issues that are discussed. It relies on accounts of ethics as they are present in the CS literature. It is therefore largely the realm of professional scientists and usually presupposes top-down initiatives, where citizens are mobilised to perform well-defined tasks in the research process of science. This is a limitation, as discussions and boundary work might also be found elsewhere. The search for literature made use of a search string developed by Kullenberg and Kasperowski (2016). The search term "ethic*" was added to the existing string for the search to capture variations such as ethics or ethical. 1 A search was performed in the Web of Science (WoS) core collection on 17 April 2020, resulting in 232 papers. From an initial reading of abstracts, many of these papers did not address ethical issues but were papers in popular epidemiology, discussing public engagement more generally. A new search was performed removing "popular epidemiology" and "public engagement" from the original search terms but also adding "citizen social science". This resulted in N=84 papers to be checked manually.
The same search string was then used in the Scopus database, yielding 139 results. Sixty-five papers were duplicates, so another N=74 papers were added to the list of papers to be checked manually.
The nature of discussions on ethics in CS is a rather recent phenomenon, displaying a measure of uncertainty among researchers in the field, which makes literature searches a somewhat 'messy' indicator in this case.
In light of this "messiness", we combined our identification of papers with a triangulation approach. Hence, papers were distributed among three readers (authors of this paper) who independently read the abstracts and reference lists of the identified papers and performed a basic concordance search for "ethic", "ethics", and "ethical" in all papers, observing if any publications should be removed from the analysis, and noting ethical boundary work that occurred commonly across several publications for further content analysis.
This resulted in the removal of 18 papers from the WoS dataset of 84 papers on the grounds of being concerned with general research ethics, not discussing specific ethical issues in relation to CS, or only recognising the need for studies of CS ethics. This process resulted in 66 papers from the WoS dataset. For the Scopus dataset, the same procedure resulted in the removal of 10 papers, resulting in 64 papers.
Thereafter, 11 papers were added to this dataset from a recent special issue on ethics in the peer-reviewed journal Citizen Science: Theory and Practice (2019), as the journal is not indexed in the databases of WoS and Scopus. Furthermore, a chapter on ethics in a recently edited publication on the "science of citizen science" was also included (Tauginienė et al., 2021). This resulted in a dataset for a qualitative content analysis of 142 publications. The content analysis followed the triangulation approach and was performed by the three authors, independently analysing each article in full, with the aim of identifying issues that arose across publications. This yielded an identification and preliminary coding of debated ethical issues. These were then compared and grouped together by the three authors, resulting in three themes of ethical boundary work.

Distribution
The main ethical boundary work performed in our empirical material departs from opinions that ethical frameworks in science are insufficient and need to be broadened or renewed for CS (Ehlers et al., 2014, p. 13;Rothstein et al., 2015, p. 902;Winickoff et al., 2016, p. 170). A recurrent narrative is that CS is different from conventional science and cannot be contained in existing institutional frameworks to control and decide on ethical issues (cf. Canosa et al., 2018, p. 400;Tauginiené et al., 2021, p. 408). The main difference is believed to be that of the highly distributed and inclusive character of CS: […] embraces people from a wide variety of backgrounds, with a diversity of values and goals, and uses inexpensive, shared, and/ or open-access technology to enable broader participation. But a lack of gates also might mean a dearth of gatekeeping, the traditional approach to quality assurance. Therefore, as citizen science creates new approaches to scientific discovery, it also must consider new approaches to ensuring research integrity (Rasmussen, 2019, p. 2). To interact with scientists as "equals" in the "research process" (Rothstein, 2015, p.  Citizens cannot be educated, instructed, and tutored extensively before being able to contribute to CS-projects, as it creates unethically excessive demands of time and effort from volunteer contributors. In some exceptional cases, there is concern about the increased requirement of citizens to be more engaged in the process; however, this is seen as an inevitable remission since the ethical issues of CS are complicated (Tauginienė et al., 2021, p. 408).
For some authors, this borders on exploitation, and will eventually be regulated by participants themselves, who will refrain from involvement and thus abandon projects.

Relevance
This theme revolves around what kind of relevance, scientific or social, should guide and drive CS. Here, issues of empowerment, research integrity, marginalised groups, funding, and translating scientific results into community relevant knowledge, and promises of change, come to the surface (cf. Chesser et al., 2019, p. 2). These kinds of issues are driven to the foreground, particularly in relation to CS-projects that are directed towards and involve local communities. The main ethical issue for authors concerned with top-down CS extending into local communities is that such projects usually have as its main goals to contribute to science, i.e. to achieve a scientific rather than a social relevance for the local communities. For that to be realised (producing scientific publications), such initiatives must rely on institutionalised ethical frameworks of science (cf. Vayena & Tsaioulas, 2015). Some authors, displaying a strong science centred position, argue for an expansion of research ethics to also include CS initiated outside scientific institutions. In their view, citizen scientists operating outside of standard institutional contexts [need] mechanisms of oversight to ensure compliance with relevant scientific and ethical standards. Only in this way can CS responsibly achieve the goal of making a socially recognized contribution to scientific knowledge (Vayena & Tsaioulas, 2015, pp. 482-83; see also Oberle et al., 2019, p. 6).
Thus, through arguments which claim that ethical review boards must guide and protect participants in CS and that most scientific journals require ethics approval (Resnik, 2019 (a), p. 22), this reasoning presupposes that community-driven projects should emulate institutional CS.
From an ethical point of view, this is seen as highly problematic by other authors as it undermines and affects the knowledge gained and its relevance for the local community (Wiggins & Wilbanks, 2019, p. 10;Morello-Frosch et al., 2009, p. 1). This is of particular concern for CS that mobilises members from marginalised groups, with respect to cultural diversity, traditions, and beliefs. Some  (Glover et al., 2008, p. 397). If members of a local community are engaged, local benefits must be provided. "Expertdriven research within resource-poor communities that does not provide some local benefits is unethical to conduct" (Quigley, 2016, pp. 731-732;cf. Chesser et al., 2019), and will eventually be regulated by participants themselves, who will refrain from being involved or refuse to be mobilised. For CS, this means that in order for various local communities to be engaged ethically, a finer granulated and context sensitive ethics than what conventional research ethics can offer is needed (cf. McGowan, 2017, p. 513; see also Tauginienė et al., 2021). A general framework for all forms of CS, and for all tasks undertaken by volunteer contributors in projects, might therefore be difficult to develop.
Coalitions of convenience may emerge out of expediency and affinity for specific issues, but shared ethics may not be central; coalitions of conscience are more likely to emerge out of necessity, commitment, and personal or collective identity, and ethics is central to the formation and critical praxis of a participatory coalition (Aungst et al., 2017, p. 362).
"Micro-ethics" is referred to in this context, illustrating the ethical decision-making required in a CS project negotiating the "fluidity" of research (Rashid, 2015, p. 525). This creates demands for a "dynamic" ethic (Tauginiené et al., 2021, p. 409), which is pitted against a research centred "macro-ethics" (Simialka & Samways, 2018, p. 637;Verplancke et al., 2016, p. 310). Multiple modes of ethical engagement and strategies that are tailored to a given situation are called for. Given the growing diversity of CS approaches changing the research landscape (Wiggins & Wilbanks, 2019, pp. 11-12), each project should address the concerns most relevant to volunteer collaborators (Fiske et al., 2019, p. 621).

Members of local communities have important local knowledge
about the ethical issues associated with particular practices, places, or communities. Thus, they may be able to collect data or conduct experiments more ethically than professional scientists (Duxbury, 2018, p. 436;Elliott & Rosenberg, 2019, pp. 50-51;cf. McConchie, 2015;cf. Guerrini et al., 2018, p. 136), as local people are often the best placed to take action on local issues, they can complement, extend, refine, monitor, or initiate conventional science and do so with an ethic of care (Carr, 2004, p. 841).
Similarly, frequent use of terms like "participant-driven", "democratisation", or even the term "citizen science" itself, is problematised in discussions of the relationship between citizen science research and local communities as they are […] increasingly being used to encourage the public to become involved in research ventures as both subjects and scientists.
Originally, these labels were invoked by volunteer research efforts propelled by amateurs outside of traditional research institutions and aimed at appealing to those looking for more "democratic," "patient-centric," or "lay" alternatives to the professional science establishment (Woolley et al., 2016, p.  One example is capitalising on CS ideologies, drawing on an idea of a social contract and greater societal good, but in fact exploiting volunteer contributors' efforts to collect and classify data, which is then stored without reach of citizens (Riesch & Potter, 2014, p. 117). Thus, neither individual autonomy nor greater societal good are automatically realised by current ethical frameworks of science applied to CS (Woolley et al., 2016, p. 14), and ethical issues are  (Woolley et al., 2016, pp. 13-14).

Expulsion
In accounts of the historical development of the field of CS, ethical issues of invisibility and displacement of outsiders are a longlasting legacy, usually resulting from the preoccupation with the dangers of bad data, low quality research, and bias (Haklay, 2013;Cooper et al., 2014).
While outsiders are not barred from taking part, or being defined as rivals, but in fact recruited and mobilised, they are expelled and made invisible after their contributions to science, taking unethical advantage of public engagement (Tauginiene, 2019, pp. 122-123). This is in accordance with historical accounts which demonstrate that the visibility of actors in scientific work is not a given, but subject to cultural values. Shapin's (1989) notion of "invisible technicians" denotes the vital scientific worker as being part of the invisible infrastructure of science. The reliance on outsiders has been found to decrease during the professionalisation of the sciences during the 20th century (Star & Griesemer, 1989;Goodchild, 2007;Miller-Rushing, 2012).
However, this account is questioned as the history of CS is now beginning to be written. Did the practises of relying on outsiders actually decrease? Or were outsiders, volunteering for service in the sciences, made invisible and not acknowledged for their contributions, and if so, why? Cooper et al. (2014) show that observations and data on avian migratory patterns and climate change still make contributing citizens invisible, despite advances in securing data quality.
The use of citizen science data in an active field of ecological research, such as migration phenology, is strong evidence that any stigma associated with the use of data collected by volunteers is unwarranted. Yet, the contributions of citizen science were not readily detectable in most cases. Thus, the stigma may persist unless researchers begin to draw attention to the citizen-science elements in their research papers (Cooper et al., 2014, p. 3).
It is difficult to answer the extent to which the practice of invisibility has ensued (Cooper et al., 2014). Some authors argue that "hundreds of scientific papers" in bird migration studies would not have been possible without volunteer contributions, yet without acknowledging these efforts (Haklay, 2013, p. 113).
Several explanations have been offered for making volunteers invisible in CS, as well as why this has not previously been considered being an ethical problem. General ethical frameworks for science, such as informed consent, exploitation, and the benefit of projects for participants, did not formally exist as guiding principles, even in medicine, until the mid-20th century (Reiheld & Gay, 2019, p. 2). Furthermore, discussions on ethical frameworks that apply specifically to CS, in comparison with the long history of recruiting volunteer contributors in science, are few and mainly occur from 2015 onwards. The main argument, however, is that "outside" contributors have been managed both as an asset and as a problem for science.
Whether outsiders could violate scientific standards has been, and still is, a continuous issue in participatory practices of science. The danger of bad data, affecting the possibilities of CS contributing to science, is foremost associated with the "citizens", as they are more (than professional scientists) prone to becoming activists or falling into advocacy as they are included in scientific work. However, according to some authors, CS cannot be regarded as being of a lesser quality than any other form of science. It has to be based on "the details of [the] specific research context", as with any evaluation of scientific work. To waste scarce resources on research of low standards is as problematic as it is to dismiss research of high value, regardless of who performs it (Elliott & Rosenberg, 2019, p. 9).

Concluding remarks: Paradoxes of ethical boundary work
The purpose of this study has been to investigate the ethical Finally, what ethical boundary work can be defined in the CS literature? Our empirical material cannot, in its majority, be subsumed under the concepts of boundary work suggested by Gieryn (expulsion, expansion, and autonomy). This is a result that is largely consistent with earlier studies of ethical boundary work in bio-medical research (see Wainwright et al., 2006;Hobson-West, 2012) and medical practices (Frith et al., 2011). Our view, therefore, is that the conceptual resources for studying ethical boundary work would likely be best regarded as a heuristic device, under which more fine granulated analytical work is still required. In particular, this concerns how actors, and not only scientists, deal with current ethical dilemmas and boundary work in their day-to-day practice of the distributed work that is taking place in CS.
On the basis of our limited data, we suggest that ethical boundary work in CS is, to a large extent, a matter of managing ambiguities without any clear boundaries drawn between the unethical and ethical, ending up in paradoxes that seem very difficult to resolve. It is in these judgments and managing of paradoxes that ethics "happen" in the field.
Ethical boundary work gives an insight into the complex role of ethics in the field of CS, which is important to bring forward and further explore, as CS is currently undergoing a process of institutionalisation, including also evoking positive visions and high expectations of democratisation, trust, and scientific literacy raised as societal benefits of CS. As Gieryn notes, "the boundaries of science are ambiguous, flexible, historically changing, contextually variable, internally inconsistent and sometimes disputed" (1983, p. 792). This also seems to hold for ethics in CS.

Funding:
This research was funded by Sweden's innovation agency, Vinnova, DNR 2017-03527 for the project Arenas for building relations for co-operation through citizen science (ARCS).