Abstract

Abstract The article focuses on stories and storytelling practices as explanatory resources in standardization processes. It draws upon an ethnographic study of the development of a technical standard for data sharing in an ecological research community, where participants struggle to articulate the difficulties encountered in implementing the standard. Building from C. Wright Mills' classic distinction between private troubles and public issues, the authors follow the development of a story as it comes to assist in transforming individual troubles in standard implementation into an institutional issue for the 1 ecological scientific community. The authors present the ''hands-on'' social science collaboration in this study as an example of a mechanism for supporting institutionalization of issues. Finally, the authors argue that narratives can serve as effective organizing principles within institutional settings, thereby providing an approach to understand the practical, substantive difficulties that occur in work with data in the sciences. Keywords stories, sensemaking, standards, intervention, trouble, issue Nasreddin sat on a river bank when someone shouted to him from the opposite side: ''Hey! how do I get across?'' ''You are across!'' Nasreddin shouted back. Between 1997 and 2001, a team of information technologists at the National Center for Ecological Analysis and Synthesis (NCEAS) initiated and carried through the first stages of development for the Ecological Metadata Language (EML). The introduction of the standard would serve as a groundbreaking event in ecology, promising to facilitate the interdisciplinary sharing of data sets and new avenues for large-scale collaborations in ecological research. As a ''universal'' language by which standardized descriptions of ecological data could be produced, data would circulate and be shared across disciplinary fields and laboratories. In 2001, the standard was officially adopted by one of the largest research communities in ecology. This adoption marked the high point in a ''success story'' of datastandard development in the sciences. However, individual research sites within the Network had difficulty using the standard when tagging actual ecological data. In particular, information managers, who were tasked with the responsibility of the majority of the work in implementation began to report troubles. It was found that contrary to an idealized image of a ''universal'' language, individual research sites have their own ways of naming, classifying, and organizing their data, making use of specific terminologies and measurement units that were not accommodated by the new standard. Over time, a new story of the standard and the standardization process has begun to emerge within the Network. In this story, the standard is not yet a success, substantial work in implementation remains, and doing this work requires changes to the standard itself, along with renewed access to human resources and time. We ask: What happened in this process of implementation of a standard that led a group of actors to formulate another history of the standard? A story, already recounted and recorded as a success story, was retold as a partial success promising that the greatest gains were to come. We adapt the work of sociologist C. Wright Mills to understand this transition, a shift from private troubles to public issues (Mills 1961), and we draw from the storytelling and sensemaking literatures to trace the storybuilding and storytelling work of participants involved in the implementation process (e.g., Czarniawska 1998; Drawing from ethnography and grounded theory methods, we follow the development of this second story, as an explanatory resource, as participants seek to make sense of their troubles in implementing the standard-highlighting connections between troubles and issues and the shift from individual difficulties or troubles to a story of a collective issue. The authors of this article were observers and participants in this process. We actively intervened in the sensemaking process, helping to shape the success-to-come story. We reflect upon this participatory role and note how recent discussions of ''intervention'' within Science & Technology Studies (STS) do not adequately account for and describe such ''everyday'' and ''on the ground'' forms of interventions. Private Troubles and Public Issues C. Wright Mills first articulated the now classic sociological distinction between private troubles and public issues. Troubles are the experiences of individuals, variously blamed on irresponsible action and poor planning or explained away as unfortunate contingencies: ''they have to do with the self and with those limited areas of social life of which he [sic] is directly and personally aware'' (Mills 1961, 8). In contrast, issues are recognized as collective phenomena: many individuals are swept along in changes that could not be planned for and to whom no responsibility can be laid, ''they have to do with ways in which various milieu overlap and interpenetrate to Millerand et al. 3 at UNIV OF ILLINOIS URBANA on April 4, 2012 sth.sagepub.com Downloaded from form the larger structure' ' (1961, 8). The difference between a trouble and an issue is largely a matter of casting the story in a different light, and assembling information in ways that render individual problems as part of a collective phenomenon. Mills uses examples, such as ''being without a job.'' Within the United States, joblessness is commonly framed as the private personal trouble of an individual and their family-a situation to be resolved by action on the part of that individual, such as finding work. However, in the face of an acknowledged crisis, such as a recession or environmental disaster, a private trouble can be recast as a public issue-''being without a job'' becomes ''unemployment''-a matter of national or international concern, to be resolved by actions of the state, nongovernmental organizations, and/or foreign aid. An issue is often more difficult to articulate than a trouble because its manifestations are not immediately available to everyday experience. A ''high unemployment rate'' is the product of collecting and assembling multiple national statistics from various private and public agencies. New information must be generated and brought to bear on the trouble in order to recast it as an issue. Making an issue out of troubles is also interpretive and argumentative work, difficult for those awash in individual everyday activities. In short, it becomes a case of collective sensemaking. Mills believes that a full formulation of the problem requires understanding ''biography and history,'' ''man and society,'' ''self and world''; or in other words, that a problem is most revealing when troubles and issues are cast simultaneously across multiple frames and/or scales. A national issue of unemployment cannot be addressed without grasping the particular mechanisms and experiences of individual troubles in addition to the broad sweep of history and social change. A sophisticated modeling of a problem draws connective strings between the troubles of individuals and historical and structural transformations. Mills emphasizes the public role of the sociologist in helping to broaden the horizons of ''ordinary men'' who ''do not possess the quality of mind essential to grasp the interplay of man and society, of biography and history, of self and world '' (1961, 4). It is in this respect that we depart from Mills' analysis, greatly tempering the sociological hubris of his arguments. While Mills treats the transition from seeing troubles to understanding an issue as a matter of ''the sociological imagination,'' or bringing to bear a professionally positioned perspective to the question, we treat the transition as a matter of participants' sensemaking, storytelling, and practical work-driven, in our case, primarily by the participants themselves. While Mills posits that ''ordinary men'' ''cannot cope with their personal troubles in such ways as to control the structural transformation that usually lie behind them '' (1961, 4), our field research revealed precisely the opposite. We saw on a daily basis that participants themselves told stories that drew together their individual troubles and began recasting them as collective issues. We take a storybuilding perspective on member's organizational work as the participants seek to make sense of ongoing difficulties in standardization. Stories are deeply implicated in every aspect of organizational life. By turning to storytelling as a sensemaking activity, the ongoing aspect of organizational action can become the object of analysis: ''A focus on stories leads naturally to a concern with themes ranging from fictionality, plurivocity and reflexivity to temporality, intertextuality and voice, all of which are suffused with power'' (Brown, Gabriel, and Gherardi 2009, 324). Stories are ''the basic unit'' of narrative (Fincham 2002, 5) and can be understood as the local activities of sensemaking associated with particular situations, instances, or past events. At its most fundamental level, a story defines a history, a current state of affairs, and then outlines a future direction for the circumstances. More than rhetorical framing devices, Julian Time itself can be framed in the explanation of the object of a story, ''The time of innovations depends on the geometry of the actors, not on the calendar'' (Latour 1996, 88). Timelines, such as the ever-present ''planning and deployment stages'' of technology life cycles are themselves ''changeable'' through narrative formation. The evolution of a project of technological development is understood not according to an inflexible linear time frame (for instance, according to stages of emergence, ripening, decaying, etc., in an evolutionist perspective), but rather, according to the temporalities framed by different actors in the project, which are held and reshaped collectively in stories. For example, the stories of standardization we explore in this article reshape the time frame of when the deployed standard may be considered a success. While initially success is a matter of formal adoption of the standard within the research community studied, we show how storybuilding and storytelling practices come to tie together the work of technological development to its adoption in the definition of success. Millerand et al. 5 at UNIV OF ILLINOIS URBANA on April 4, 2012 sth.sagepub.com Downloaded from Through storytelling, a complete success in standardization becomes part of the future of a data standard, something that can only be claimed after user adoption. The storybuilding perspective can be contrasted with a more ''objectively'' oriented approach that sees only definitive successes or failures. Such rationalist perspectives generate an exclusive focus on outcomes and casts success in stark black and white terms (e.g., in some of the actors' perspectives in the case of the Aramis technology studied by Making an Issue Out of Troubles In fact, as we will see, no one in our study is telling a story of failure, rather, it is a case of retelling a recognized success as a matter of overcoming difficulties and establishing ongoing commitment. As Fincham notes in his analysis of success and failure narratives in technology adoption, ''Rather than being seen as end points (whether a set of causal factors or process), when conceived in narrative terms, success and failure claims form an interactive discourse' ' (2002, 2). Shades of gray emerge, and success becomes a negotiated marker or a future goal. By making an issue out of their troubles, participants gain a new handle on their difficulties. 1 It is precisely by reframing personal troubles as public issues-as a matter of structural rather than individual concerns-that participants make sense and ''cope'' with their difficulties. The move from private trouble to public issue can be understood through the lens of organizational sensemaking. Viewed as a significant process of organizing, sensemaking unfolds as a sequence in which people concerned with identity in the social context of other actors engage ongoing circumstances from which they extract cues and make plausible sense retrospectively, while enacting more or less order into those ongoing circumstances. (Weick, Sutcliffe, and Obstfeld 2005, 409) Stated in a more concise way, sensemaking involves ''turning circumstances into a situation that is comprehended explicitly in words and that serves as a springboard into action'' (Weick, Sutcliffe, and Obstfeld 2005, 409). Storybuilding is a means by which one makes sense of the world and acts on it. Coherence is a key modifying adjective in our use of story The story must first of all ''work,'' be coherent in any of the many ways stories can be of one piece [ . . . ] The other constraint is that the story must be congruent with the facts [ . . . ] We don't accept stories that are not borne out by the facts we have available. (1998, Roughly speaking, our empirical case begins at the point where there was a single story (success-already), and in this article, we trace the emergence of a new story (success-to-come) through the storybuilding activities of participants. The two stories are related and continue to unfold side by side, one a precursor not only setting the stage, but prompting emergence of the other. This conjoining of stories highlights the sensemaking aspect, in addition to the persuasive element involved in storybuilding A Metadata Standard for the Ecological Sciences We focus on a particular data standard, the EML. 2 In short, the standard provides a shared method for describing data across the ecological sciences in the hope of facilitating data sharing, reuse, and management. The standard was developed at a national ecological center (the NCEAS) and was deployed within an ecological research community (the Long-Term Ecological Research [LTER] Network), a US federation of ecological research sites consisting of more than 2,000 members. We will refer to the NCEAS as the ''national center'' or ''center,'' and the LTER Network as the ''Network'' throughout the article. The social studies of standardization are filled with stories of local resistance and power struggles, of failed plans, and unexpected successes The Network consists of ecological scientists seeking to understand past and present-day ecosystems, as well as anticipating potential futures The standard defines a fixed set of tagged fields that structure the text describing any given ecological project, data set, and/or collection of data sets together with their related references and personnel. Literally ''data about data,'' metadata consists of a set of labels or tags, tag categories, and their relational structure. Tags such as ''title,'' ''location,'' and ''unit'' are used to demark text that provides information about a data set into sections more structured for human understanding, as well as more amenable to at UNIV OF ILLINOIS URBANA on April 4, 2012 sth.sagepub.com Downloaded from automated machine searches. Detailed, standardized metadata can facilitate many tasks, such as searching of relevant data (e.g., requesting all data sets that contain the term biomass in the tagged field title), data availability from multiple field sites (requesting the data location from the tagged field URL), and data integration (requesting data sets with measurements in milligrams per meter cubed in the unit tagged field). Two often ignored aspects of field practices that create difficulties for data sharing are their situatedness and the manner in which data are moved beyond the sites of their production. First, the understanding of scientific field data is closely bound to the local venue or data-collector (e.g., Goodwin 1995). Second, the production of history and context for scientific data is increasingly erased as it moves away from the site of its production, eventually becoming almost invisible as a story completes with frame, interpretation, and limitations upon publication of an article (what Latour and Woolgar 1986 have called the deletion of modalities). The context within which issues of standardization play out has been described from the perspective of the ecological sciences as a growing awareness of the social and technical dynamics associated with synthetic efforts in both basic and applied science While journal publication is part of a well-established scientific process of public community review, publication of data sets and their associated metadata is novel within the ecological sciences, involving new types of work not yet integrated into conventions of existing work and accreditation. It is precisely this gap that the standard seeks to fill, a method of documenting data in ways that capture key features of its collection and methods of production. With rich metadata, data are contextualized in support of both wider reuse and legacy use. That is, the use of data is extended to include others who may be addressing questions beyond the original scientific questions that led to the collection of data (data reuse) and/or recall of the data for use at later times. A Brief History of the Standard The EML was developed by a team of information technologists located at a national center between 1997 and 2001. In 2001, it was adopted as the Millerand et al. 9 at UNIV OF ILLINOIS URBANA on April 4, 2012 sth.sagepub.com Downloaded from official metadata standard of the Network. The problems we investigate in this article focus on the implementation of that data standard within the Network. These problems manifested principally at the divide between those who developed the standard (information technologists at the national center) and those who were tasked to implement the standard by describing existing data sets using the new standard (information managers within the Network). 3 A first version of the standard saw the light in 1997 at the national center. It was the product of a small team of information technologists trained in computer science and ecological research. The standard fit within the core mission of the center, which is the support of cross-disciplinary research that uses existing data to address major scientific challenges in ecology. 4 The information technologists working at the center were engaged in various technological projects, developing tools and techniques for the environmental science community. We will refer to them as the ''developers'' of the standard throughout the article. Information management is a formal body within the Network. Each of the twenty-six research sites has an information manager, tasked with carrying out data and information management. Notably, at the Network level (of all twenty-six sites), there is an Information Management Committee with one member from each site. Thus, the information managers are responsible for managing data and a data repository at the site level and also for collectively planning data curation and integration at the Network level The recasting of the standard's implementation as a matter of success-tocome was largely reasoned and articulated by participants of the Network itself, primarily by information managers. It was their hands-on experience in attempting to implement the standard, and the continuing interaction among themselves as an organized subunit of the Network that provided the raw materials for reinterpreting their troubles as issues. field is ongoing-still continuing more than six years after its inception. However, within this article, we focus on the period where the second narrative (success-to-come) emerged and took form, essentially between 2004 and 2006. We participated in more than 200 events relating to work with the Network over the period of the study. Specifically relating to the standard topic, we conducted ten interviews, participated in nine conference call discussions and six working meetings, and attended several design sessions. Research Design and Methods 5 Interviews were with representatives from the main groups of actors involved (i.e., information managers, developers, and scientists), some of them we interviewed repeatedly. All the interviews and selected sections of conference calls, working meetings, and design sessions were transcribed and coded with a qualitative data analysis software (NVivo). Data analysis followed grounded theory methodology, from coding to categorizing to theorizing, developing from memo writing informed by participant observations notes. The quotes presented in this article are marked as information manager (IM), developer (D), and scientist (S). Document analysis was carried out longitudinally, and included standard documentation, e-mails, information managers' publications, and Network reports and publications. Being physically present at the research site, in this case at two sites (Palmer Station and California Current Ecosystem) located at Scripps Institution of Oceanography in San Diego, CA, allowed for two authors of the article (Millerand and Baker) to engage in participant observation almost on a daily basis. One of the authors of this article (Baker) is the information manager for the two sites; also trained in STS, she brought to bear a ''sociological imagination'' to the troubles in implementation. The authors of this article were observers and participants in the process of storybuilding and storytelling around the enactment of the standard. We actively contributed in the sensemaking process, in particular helping to shape the success-to-come narrative. But we, as social scientists, by no means credit ourselves with the bulk of building the emergent interpretive narrative. We were not privileged actors ''unveiling'' the truth of an issue to those mired in a situated view of their troubles; rather, we were one kind of participant in a highly diverse, largely expert mix of participants seeking to enact a standard. We were a sounding board, providing context and language, prompting dialogue, and participating in joint reflection. Within STS, such forms of participation by the researcher in the shaping of the object of study are known as ''interventions'' (i.e., the social researcher partaking in the unfolding of the research object). Of late, much has been written on the topic of STS scholars intervening within policy or legal spheres Telling Stories: Making an Issue Out of Standardization Our empirical study begins at the point where a new story (success-tocome) was developed, thus challenging the dominance of the previous one (success-already) in its attempt to account for the standardization process. The two stories share a factual understanding of the point of adoption of the standard and of the importance of this moment, but the interpretation of the significance has come to differ substantially. Is adoption the end point in the story of standardization, now considered a success, or is adoption an important milestone toward a final goal of data practices standardization in the Network? The success-to-come story differs in three distinct ways from successalready. First, the success-to-come story emphasizes the difficulties that arise in implementing the standard-requiring additional resources and expertise. In contrast, the success-already story demarcates a transition point in which the standard has been successfully developed and which ''merely'' leaves the task of implementation ahead: the greatest investment of resources, expertise, and time had come and gone. Second, it is expected within the success-to-come story that some difficulties in implementation are insurmountable without changing the standard itself, thus, calling for some redesign of the standard or of its use by the information managers. Third, following the previous two points, the success-to-come story emphasizes that the process of the standard's enactment and the solutions to related problems requiring significant work and innovation are beyond the reach of a single individual site but are within the scope of the Network or even the domain of ecology. The two narratives frame differently the particulars of the problem 6 of standardization, and thus, suggest different kinds of individual and collective action. We begin with the framing of the problem associated with the success-already story, which casts the story in terms of private troubles. Having Troubles in Implementing the Standard As soon as implementation started at the research sites within the Network, problems emerged. When difficulties in implementing a standard are cast as troubles, they are perceived as unique and exceptional circumstances (i.e., a problem for an individual information manager at a site to be addressed and solved locally). Immediately following the adoption of the standard, both developers and information managers accepted this formulation of the situation. Below, we outline in detail an instance of an implementation problem, cast as a trouble. Because such troubles were not completely unexpected by the developers, in the next section, we illustrate the individually targeted solutions that they planned, and in this case, a set of workshops for the information managers. Being in Trouble In order to characterize the nature of a trouble, we outline a single instance of a problem, as articulated by an information manager describing biochemical data using the standard. This event, and the interview selections that are based on it, occurred shortly after the official adoption of the standard as the implementation phase began. Jane is an information manager working at one of the twenty-six LTER sites in charge of managing the data collected by scientists at her site. The site is a biome with research focusing on the impact of human development on the quality and quantity of water. Jane is attempting to describe an existing measurement within a nutrient data set using the newly adopted standard. For her site's ecosystem, nutrients are any of the organic and inorganic substances that serve as nourishment for plants; these are commonly composed of, for example, phosphates, silicates, nitrites, or nitrates. They are a crucial component of any ecosystem and can be a limiting factor for a biological system. A common unit for the measurement of phosphates is microMoles; a unit used in chemistry for the amount of a substance. In applying the standard to her data, Jane finds that microMoles are not included as a metric in the standard. Instead, Jane uses a naming convention that provides a guide for capitalization and ordering of the parts of the name at hand (capital M on moles): I was getting nutrient data, and my units came in as micromoles with the micron symbol and capital M, microMoles. When I started having to go into Jane runs into two troubles: the standard does not provide guidance on biochemical units and, when she does research the unit name on her own, she finds that the measure used at the site is a ''custom'' unit that the site's scientists use as a shorthand convention. Here, ''custom'' means that it is a locally used unit, rather than one common to the Network or to ecology more broadly: And in digging deeper and going to our lab that processed these data, I found out it's not microMoles, it's microMoles/liter. And I am not a chemist so it just didn't mean anything to me. You know, I am just organizing and posting this type of data, and so it really opened my eyes that I have a bigger issue here than I thought, you know, because here we've got people reporting things as microMoles, which is not proper. But that is just the way the work is done, and shared, and no one ever questioned it. (IM) Jane realizes that the shorthand convention used at her site lacks the completeness required to be understood by those outside the site, a key goal for metadata. The naming convention sufficed for use at her site, but the full formal name including the ''per liter''' (that makes explicit that this is a measure of density, an amount per volume, and not simply an amount) was missing. While the shorthand is not ''wrong'' per se, that is, it is sufficient for the needs of scientists at the local site, for the purposes of the metadata standard, this specification is inaccurate or as Jane says: ''not proper. '' 7 In other words, the unit is not proper for communicating the data to the broader Network. Jane began to compile a list of the units used at her site that could be reviewed by scientists and made available to site researchers. Like Jane, most of the information managers perceived their difficulties with the standard implementation as unique and exceptional occurrences, or rather, as individual troubles they needed to address and solve on their own. As we show in the following section, the developers of the standard also perceived emerging difficulties as private troubles experienced individually at each site. Targeting Solutions Individually That there would be troubles in implementing the standard was not in itself a surprise for the developers. They were familiar with the heterogeneity of the sites and the data in the Network as well as the differences in the backgrounds of individual information managers. However, they perceived these problems as troubles, that is, as difficulties to be addressed through individually targeted actions at each site. Below we describe their solution: tutorials and training sessions for information managers. This solution, targeting information managers' deficiencies, would train individuals at sites in how to implement the standard. Developers related difficulties and lag in the implementation of the standard across the Network directly to the variation in the sites' information systems. Only a few of the ''ideal'' sites were able to implement the standard quickly because their data were stored in highly structured databases. Other sites used ''semistructured'' files and a lot of the sites had files with very little structure. As a developer describes: