TY - JOUR
AU1 - Pammer-Schindler, Viktoria
AU2 - Prilla, Michael
AB - Abstract A substantial body of human-computer interaction literature investigates tools that are intended to support reflection, e.g. under the header of quantified self or in computer-mediated learning. These works describe the issues that are reflected on by users in terms of examples, such as reflecting on financial expenditures, lifestyle, professional growth, etc. A coherent concept is missing. In this paper, the reflection object is developed based on activity theory, reflection theory and related design-oriented research. The reflection object is both what is reflected on and what is changed through reflection. It constitutes the link between reflection and other activities in which the reflecting person participates. By combining these two aspects—what is reflected on and what is changed—into a coherent conceptual unit, the concept of the reflection object provides a frame to focus on how to support learning, change and transformation, which is a major challenge when designing technologies for reflection. RESEARCH HIGHLIGHTS The reflection object is both what is reflected on and what is changed through reflection. Designing for reflection means to identify what is relevant to users and what can be changed (the reflection object), how to represent this reflection object and which contextual information is necessary to understand the object and its development in time. The object’s representation is a tool for reflection; it triggers reflection and creates communicative context. The concept of reflection object as a conceptual tool creates a focus to think in parallel about that which is reflected on and that which changes; to think in parallel about the sameness of and change in the reflection object; to think in parallel about reflection and the activity for which this reflection is important as the reflection constitutes elements of or a complete linked activity. 1 INTRODUCTION Information systems can be designed with many goals in mind. One such goal is to support work that it is carried out more efficiently and another is simply to entertain. The goal of designing systems that support users in reflection has now come to the attention of research in human-computer interaction (Baumer, 2015; Baumer et al., 2014; Choe et al., 2014; Fleck and Fitzpatrick, 2010, Isaacs et al., 2013; Li et al., 2010; Li et al., 2011; Malacria et al., 2013; Mamykina et al., 2016; Sas and Dix, 2011; Slovák et al., 2017). Ghajargar et al. (2018) have even proposed reflection as a reference design task, i.e. a specific kind of task to which human-computer interaction research should pay attention to because designing for it is substantially different than designing for other types of tasks. A wide range of tools for facilitating reflection has already been designed, evaluated and published, in different domains of usage. These tools are sometimes published under names that do not immediately point to reflection, such as lifelogging (cp. Sas and Dix, 2011), quantified self or personal informatics (cp. Choe et al., 2014; Li et al., 2010; Li et al., 2011), learner modelling (cp. Kay and Kummerfield, 2011) or learning analytics (cp. Siemens, 2013). Across this spectrum, research emphasis varies between technical challenges versus challenges in relationship to usage and effect, and outcome expectation varies in relationship to what is the desired effect, such as changes in knowledge, perception of experience(s) or behaviour. To cite a few exemplary tools that have been developed in such research, Echo (Isaacs et al., 2013) lets users describe significant situations and events in their lives textually and with multimedia data and prompts users later to remember and reflect on them. KnowSelf (Pammer and Bratic, 2013) captures and visualizes application and document usage on computers and thereby creates a data basis for reflecting on time management. The Mobile Diabetes Detective (Mamykina et al., 2016) enables users to document their blood sugar level and provides data analytics functionalities in parallel to reflection scaffolds and prompts. TalkReflect (Prilla, 2015) enables users to describe significant situations textually, to share descriptions with others and to collaboratively and asynchronously reflect on such situations. Narcissus (Kay and Kummerfield, 2011) combines log data from a software code repository, an issue tracking system and a project management wiki to provide an overview and data basis for synchronous and collaborative reflection on progress and collaboration patterns. All exemplary tools listed above are useful to reflect about something concrete, such as personal memories, time management, lifestyle as relevant for diabetes, difficult conversations or collaboration in software engineering. All of the tools represent these reflection objects in some manner. It is clear from the descriptions that the designers of these applications have devoted careful thought on the issue of what is actually reflected and furthermore how to represent this in their tools. The same is true for most reasonably well-designed and well-published tools that aim to support reflection: They are intended or proven to be useful to reflect about something concrete; they represent that which is reflected on in some manner; and of course, substantial effort has gone throughout research and design into identifying something that is useful to reflect on, and how to represent this. However, neither foundational theoretical works on reflection and reflective learning such as Boud et al. (1985) or Schön (1983) nor design-oriented conceptual works such as Baumer (2014,2015), Fleck and Fitzpatrick (2010), Krogstie et al. (2013), Pammer et al. (2017) or Slovák et al. (2017) have a name beyond the concrete (e.g. personal memories, time management, lifestyle as relevant for diabetes, difficult conversations, collaboration in software engineering), nor an explicit concept, for this. In the present paper, we develop the concept of the reflection object systematically using activity theory: when we understand reflection as an activity, then the reflection object is the object of this activity. This means that the reflection object is that which is reflected on, and that which is changed through reflection. The reflection object is therefore different from a particular past experience, or a set of data on which people reflect. Further, the reflection object connects reflection to other activities in the life of those who reflect. When designing technology that should facilitate reflection, the reflection object serves as a conceptual tool that supports the analytical (cp. Löwgren and Stolterman, 2007) and dialogical (Fallman, 2003) aspects of design. The paper is structured as follows: Gap identification: We first describe existing theoretical works on reflection with an emphasis on design-oriented theoretical works from research on computer-mediated reflection and identify the existing gap. Core concept: We develop the concept of the reflection object using activity theory as a conceptual tool and framework. Concept development along four themes: Subsequently, we elaborate the reflection object by transferring characteristics of the object in activities according to activity theory to the reflection object. We have structured this concept development along four themes, namely (i) what is the reflection object, (ii) what is relevant context, (iii) the development of the object in time and (iv) the role of its representation. In this elaboration, example cases of designing for reflection are given to explain the concept and the four themes. Discussion and conclusion: We conclude the paper by synthesizing the contribution and discussing it in relationship to existing design-theoretical works. Finally, we discuss the methodology used. 2 BACKGROUND AND RELATED WORK This section reviews literature that structures how to think about reflection; first in general, and then in design-oriented contexts. Subsequently, the existing literature is synthesized and the gap of a missing concept for that which is reflected on and changed through reflection is pointed out and developed. The body of literature discussed in this section is the body of literature that our work aims to complement and contribute to. In the context of this paper’s line of argumentation, the role of this section is also to show the gap in literature that we identified, and which we argue to be able to close. 2.1 Reflection Reflection means re-thinking existing knowledge, values, behaviour and practice with the goal of deriving insights on and potentially changing aspects of knowledge, values, behaviour and practice (cp. especially Boud et al., 1985; Schön, 1983). Boud et al. (1985) conceptualize reflection as returning to past activities, ideas and feelings, re-evaluating these and drawing new insights on them, including new perspectives and changes for future behaviour. This shows important characteristics of reflection: firstly, reflection is based on reconstructing or recalling past activities as well as on representing ideas and opinions. Second, in contrast to activities such as rumination and venting, or concepts like sensemaking (e.g. Crowston and Kammerer, 1998; Weick et al., 2005), reflection has a clear focus on outcomes, that is, learning and change. In most of the literature, reflection is understood as an individual activity and essentially a cognitive process. This view is dominant in the context of reflection as a means of learning, e.g. in Boud et al., 1985, and Schön, 1983, and is also taken in Daudelin, 1996, in the context of work practice even while acknowledging the social context in workplaces. HCI literature also frequently takes this view, for instance in research on quantified self (e.g. Choe et al., 2014; Li et al., 2011), adaptive help (Malacria et al., 2013) or when supporting remembrance (e.g. Isaacs et al., 2013). In collaborative and organizational settings, reflection can also be conceptualized as a collaborative activity and social process (Prilla and Pammer, 2012; Prilla, 2015). Krogstie et al. (2013) and Pammer et al. (2017) finally bring together both perspectives, that of reflection as an individual and as a collaborative activity in the context of workplaces. 2.2 Computer-mediated reflection This section discusses how others have conceptualized computer-mediated reflection. The below works do not focus on a single piece of technology, but rather put forward structured ways of thinking about computer-mediated reflection. We first discuss works that do so by separating reflection into different phases, then a work that describes different levels of reflection in terms of reflection quality and then a work that describes what are good conditions for reflection. Finally, we discuss works that have contributed in terms of how to think of the role of computational tools in reflection. Several works theorize about designing for reflection from the perspective of separating reflection into different phases. Based on background theory, Baumer (2015) has proposed three principal dimensions along which reflection should be understood by designers, namely breakdown, inquiry and transformation, based on a literature survey. By breakdown, Baumer understands situations or moments that constitute the starting points for reflection. Breakdowns include elements of surprise, puzzlement, conflicts or ‘explicit consideration of that which was previously unconscious and implicit’ (ibid, p. 6). Inquiry means the process of ‘generating, testing, revising, and further testing hypotheses’ (ibid, p. 6), a ‘re-examination of things—concepts, ideas, theories’ (ibid, p. 6). Transformation is a change in understanding, conceptual schemata, values or decisions in the sense that ‘reflective decision making (...) enables decision making that is not subject to (...) biases’ (ibid, p. 7). Li et al. (2010) investigate reflection in quantified self, where people self-track relevant aspects of their lives with the goal of changing and improving own behaviour. Based on background theory and their own empirical work, the authors have identified five stages in this process. First, in a preparation stage, people decide to start self-tracking and prepare their own method (tools, procedure) to do so. Second, in a collection stage, people are engaged in collecting data, either manually via calendar entries for instance or automatically such as when analysing phone usage. Third, in an integration stage, people ‘prepare, combine and transform (data)’ (ibid, p. 5). Fourth, in the reflection stage, people reflect on data, and fifth, in the action stage, users ‘choose what they are going to do with their newfound understanding of themselves’ (ibid, p. 6). Krogstie et al. (2013) and Pammer et al. (2017) investigate reflection in the context of workplace learning. Based on background theory and their own empirical and design work, the authors have identified four phases in reflection, such that in a first phase, people ‘plan and do work’. In a second phase, reflection is set-up. This includes goal setting, involving relevant people and planning the actual reflection. In a third phase, people reflect. In the fourth phase, plans for the future are made (‘apply outcome’). This includes planning how to apply insights in future work practice, as well as planning follow-up reflection sessions. One unique notion put forward by the authors is that of triggers that initiate reflection. Triggers can be internal or external, the first are perceived discrepancies, and provide the motivation and individual or collective rationale to reflect. The second are external impulses for starting to reflect, which essentially constitute impulses to take time and space to reflect, albeit only to reflect briefly on whether there is something worth reflecting on. Fleck and Fitzpatrick (2010) structure their discussion around levels of reflection. Levels of reflection can be understood as the extent to which an activity (or an outcome of an activity) is reflective. The authors’ model is based on background literature and their own design work. Level zero, revisiting, consists of reporting the past, without adding significant ‘reflective thought’. This level is not understood as reflective in the understanding of the authors. Level one, reflective description, includes ‘justification or reasons for action or interpretation’ (ibid, p. 3). In level two, dialogic reflection, ‘relationships between pieces of experience or knowledge’ (ibid, p. 3) and generalizable insights are sought. In level three, transformative reflection, ‘fundamental questions’ are asked, and change of future practice is explicitly intended. In level four, critical reflection, ‘social and ethical issues are taken into consideration’ (ibid, p. 3). Slovák et al. (2017) discuss conditions that help learners to reflect. Based on background theory and their own empirical and design work, the authors identify as one condition for reflection learning that learners are enabled to construct knowledge by themselves in a setting that provides them with the experiences, which are real enough as well as safe enough to learn from. At the same time, the reflection process should be guided. In contrast to the reflection approaches described above, the authors discuss designing not only technology for reflection but also the overarching reflection activity. In parallel, researchers both in HCI and in educational technology have thought about the role of computational tools for reflection (Ghajargar and Wiberg, 2018; Ghajargar et al., 2018; Krogstie et al., 2012). In this, Krogstie et al. (2012) have focused on functions that computational tools can deliver for reflection, such as collecting or visualizing data, and supporting documentation. Both Ghajargar and Wiberg (2018) and Ghajargar et al. (2018) discuss the role of computational tools for reflection in the context of smart physical objects. Firstly, the authors develop a conceptual understanding of the role of tools as enablers for reflection that has the goal to lead to behaviour change (Ghajargar and Wiberg, 2018). Secondly, the authors argue that reflection should be considered as a reference task in HCI, designing for which is distinctly different from designing for other (kinds of) tasks (Ghajargar et al., 2018). In this work, the authors focus on the relationship between the users (=reflective practitioners) and the artefact (=computational tool intended to support reflection). 2.3 Synthesis: conceptualizations of the reflection object in related work This section synthesizes the above described conceptual contributions to computer-mediated reflection. It shows their commonalities and rooting in existing theories for reflection and points out the gap that there is no generalized concept for that which is reflected on and changed through reflection. The foundational works of Boud et al. (1985) and Schön (1983) talk about experience(s) and practice as that which is reflected on; knowledge, behaviour or individual perceptions as that which is changed; and future experience(s) and practice as that which is impacted through reflection. Note that as ‘experiences’ occur only once—either in the past or in the future—it cannot easily be said that future experiences are changed through reflection, but they are impacted in the sense that it makes a difference for the future course of things whether reflection has occurred or not. Subsequently, in this view, immediate change and hence effect through reflection is at the level of the reflective practitioner who learns. In the works discussed above, the authors write about experience, events, actions, knowledge, interpretations, ideas, values, thoughts as being reflected on as well as about concrete examples such as financial expenditures (Li et al., 2010) or nursing practice (Krogstie et al., 2013). A coherent concept is missing, however. Subsequently, a systematic relationship of these exemplary reflection objects to the reflection activity is missing. In the above discussed literature, there is an agreement that a key element in reflection is the derivation of learning outcomes. The respective phase is termed transformation by Baumer (2015), transformative reflection level by Fleck and Fitzpatrick (2010), reflection stage by Li et al. (2010, 2011) and applying outcomes by Pammer et al. (2017). The implicit understanding that underlies the above cited works follows Boud et al. (1985) and Schön (1983), namely that work practice, experience or more generally speaking behaviour is impacted through reflection and will differ in relevant characteristics from experience, work practice or behaviour before reflection. In addition, the knowledge, perception, attitudes and values of those who reflect are also understood to change. However, none of the above works clearly binds together that which is reflected on with that which is changed. To give a few examples from aforementioned related work, Fleck and Fitzpatrick (2010) mention the use of pictures taken throughout the day but leave reflections and changes to the communication of people; in parallel, reflection is on pictures, but change is intended at the level of future (general) life. Li et al. (2010, 2011) describe people as reflecting on data with the goal to change behaviour; again, that which is reflected on is conceptualized separately from that which is changed even while the authors of course make concrete arguments in single cases of how data are connected to that which should be changed. Ghajargar et al. (2018) discuss how smart, physical objects can support reflection in their role as a mediating artefact but do not connect that which is reflected on (not conceptualized explicitly in this work) with that which is changed (behaviour in this work). This conceptual gap makes it difficult to connect design decisions for triggering reflection and representing that which is reflected on with design decisions about how to support transformation. We propose the following: to use the concept of reflection object to talk and think at the same time about that which is reflected on, and that which is changed through reflection. We see this as one step towards being better able to support transformation, which is the most challenging phase and level to design for (cp. Baumer, 2015). 2.4 Using activity theory as a tool in order to conceptualize the reflection object In this section we introduce the concept of reflection object. We have used activity theory as a conceptual tool for thinking about reflection, and in particular the reflection object. We have chosen activity theory because it is a mature framework that is accepted in HCI (see Kaptelinin and Nardi, 2009) and information systems literature (see Karanasios, 2018) as a tool and framework to think at the same time about human agency and about the (computer-based) tools that are designed with the intention to support such agency. Central to its usefulness in HCI is the core proposition that (computer) tools substantially shape how human activities are performed: every tool enables some ways of performing an activity and makes some other ways of performing it more tedious or impossible. Naturally, the goal of design is to enable good ways of performing an activity, and making bad ways of performing an activity hard or impossible. This seems to us essential also when designing for reflection. Moreover, activity theory understands all elements of an activity to develop over time and has originally been developed as a framework to understand (human) learning. This resonates with reflection, in which learning and behaviour change are desired outcomes. At the same time, activity theory adds perspectives to reflection as we argue below. Finally, activity theory has also influenced Ghajargar et al. (2018)’s thinking about how (smart) things can support reflection. Below we first review activity theory very briefly and then frame reflection in terms of activity theory. In this context, the role of this section is to introduce the concept of the reflection object. In order to be able to do so, activity theory needs first to be introduced sufficiently such that the core concept of the reflection object can be derived from it, which is done via a simple but of course fundamental step of framing reflection as activity. 3 THE BRIEFEST OF INTRODUCTIONS TO ACTIVITY THEORY Activity theory is a philosophical framework that postulates ‘activities’ as relevant units of analysis. Activities are overarching endeavours; they are long-term processes that ‘consist of actions or chains of actions’ (Kuutti, 1996, p. 30). An activity consists of a human actor (=the subject) who manipulates an object and is motivated by that object. Activities are mediated (shaped, influenced) by tools, so that depending on which tools are used, different ways of acting are possible or impossible (ibid, p. 27). Tools can be both physical and conceptual, such that a useful formula, concept or method is equally a tool as an item of software or a hammer. An object can be ‘an object of eating, an object of labor, an object of contemplation, etc.’ (Leontiev, 1981, as translated and cited in Kaptelinin and Nardi, 2009, p. 140). By object one understands both the conceptual target of an activity and the material object that is manipulated by the activity (Kaptelinin and Nardi, 2009, p. 138ff—Objekt versus Predmet). When constructing a statue for instance, the material statue is the physical object of the activity. This is external to the subject. The concept of the statue is the internal object of the activity. Both the externally existing statue and the internally existing representation of the statue are aspects of the activity’s object. This object directs the overall activity, in that it motives subordinate actions, such as buying material, hiring workers, etc. We understand the object therefore as combining physical, tangible and external as well as conceptual, intangible and internal aspects. An object of an activity is different from basic human needs and also different from motives that drive human actors, such as the need to earn a living, or the motive of being well respected or making a career. However, objects satisfy or address actors’ needs and motives, in that they motivate and give meaning to activities (Kaptelinin and Nardi, 2009, p. 143ff—Objects versus motives). Activities consist of subordinate units, actions, which again can be decomposed into operations. However, the meaning of and motivation for actions cannot be understood when looked at outside the context of the wider activity. In this sense, an activity is the smallest possible unit of analysis based on which human agency can be understood—so an activity can be decomposed into smaller units (actions, operations; subject, object, tools), but these cannot be understood on their own and instead need the unit of ‘activity’. On the other hand, a single actor is typically involved in many activities, and activities themselves are interlinked and influence each other (Kuutti, 1996, p. 30). Subsequently, while a single activity is postulated as being the smallest unit of meaningful analysis, analysis may also need to consider other activities in which a subject is involved. In the present work, we refer to activities that are linked to the reflection activity via the same subject and an additional connection via the reflection object as ‘linked activities’ (see next subsection below). Further, activity theory understands individuals as acting within a social context. Firstly, tools capture cultural and historical context in that tools are artefacts in the most direct sense—they are man-made and are tools not by nature but because they are used as such in the context of activities (Kaptelinin and Nardi, 2009, p. 71). Secondly, humans participate in collaborative activities in which multiple actors work on and towards a shared object. The relationship between the subject and the community is mediated by rules, and the relationship between the community and the (shared) object is mediated by and organized using a division of labour with distinct roles (Kuutti, 1996, p. 27ff). The relationship between individual actors and a community of actors is in some works also framed as a ‘collective subject’ (Kaptelinin and Nardi, 2009, p. 187). We do not use this conceptualization in the present work, however, so when we talk about the subject of an activity below, it is always an individual; however, embedded in a community. Last, but really not the least, we highlight that in all these discussions the development of subjects, objects and tools over time is a key and re-appearing concept in activity theory. Through and by means of an activity, all entities of the activity change over time. Subjects develop their knowledge and skill through activities, tools and how they are used are created or further developed through activities, while the object of the activity is manipulated and transformed via the activity. In the present work we emphasize the development of the object. We have visualized the core concepts in Fig. 1 (top figure) using the by now iconic pyramid picture of activity theory as in Kuutti (1996). FIGURE 1 Open in new tabDownload slide Framing reflection as activity—using the by now iconic pyramid picture as in Kuutti (1996) on the top as underlying representation of an activity. Linked activities are those activities from which the reflection object is picked. (a) The object of an activity can also be the object of reflection. (b) Tools or concepts that mediate the linked activity can be the object of reflection. (c) The community of an activity can be the object of reflection. Any other constituent (subject, community rules, division of labour) or combination thereof could also be the object of reflection. (d) In particular, the reflection object may also fully encompass another activity. FIGURE 1 Open in new tabDownload slide Framing reflection as activity—using the by now iconic pyramid picture as in Kuutti (1996) on the top as underlying representation of an activity. Linked activities are those activities from which the reflection object is picked. (a) The object of an activity can also be the object of reflection. (b) Tools or concepts that mediate the linked activity can be the object of reflection. (c) The community of an activity can be the object of reflection. Any other constituent (subject, community rules, division of labour) or combination thereof could also be the object of reflection. (d) In particular, the reflection object may also fully encompass another activity. 4 FRAMING REFLECTION AS AN ACTIVITY In this work we conceptualize reflection as activity in the sense of activity theory, in that the subject of the reflection activity is the reflective practitioner or learner, and the object of the reflection activity is the reflection object. Consequently, the following understanding of reflection and the reflection object is derived from activity theory. Tools for reflection shape the reflection activity. Examples for tools are concepts about learning and reflection, as well as methods and structures for reflection, and of course also software tools that support reflection. In the present work, we focus on the design of software tools. Since tools are both enabling and limiting, a proper understanding of the reflection object such as that on which the tool should work is imperative in order to design in a way that enables adequate actions, and limits the wrong ones instead of vice versa. What is an adequate tool for a particular reflection activity hence also depends on the reflection object as the tool needs to help in manipulating and transforming the object in order to support reflection on this particular object. Reflection is linked to other activities in which a subject participates via the reflection object. In reflection, people pick as a reflection object other activities or important elements of other activities, which we call ‘linked activities’. This means that people reflect on the linked activity or focus specifically on one element of a linked activity, such as on one or more tools, the community, community rules or the division of labour in linked activities, etc. Note that following activity theory, elements of activities, such as one self, tools, the community, community rules, etc., can only be understood in the context of the full activity in which they are embedded. We now have two possibilities to conceptualize this: we could demand that the reflection object can only be a full activity since smaller elements cannot be understood on their own. Alternatively—and this is the conceptualization that we have chosen here—we allow different sizes of granularity for choosing a reflection object (full activity, tools, community, community rules, self, etc.), and subsequently seeing the reflection object as also appearing in a linked activity. In this conceptualization, the linked activity in which these elements are embedded is still necessary to understand the single elements; this is discussed further in the subsection on ‘Context: Additional information about the reflection object’ below. Examples that illustrate that reflection is linked to other activities via the reflection object are as follows: reflecting on a product that is the object of a company’s core value creation process (Fig. 1a—reflecting on the object of a linked activity), reflecting on knowledge or concepts that are used as tools in other activities (Fig 1b—reflecting on tools of a linked activity), reflecting on expectations on professional behaviour and appearance (Fig 1c—reflecting on community rules of a linked activity) or reflecting in an overarching manner about another activity such as a past project (Fig 1d—reflecting on another activity, the other, linked activity is the reflection object). We note that alternatively, reflection could be conceptualized as action that is subordinate to a wider activity. We do not use this conceptualization in the present work: viewing reflection as activity allows us to conceptualize separately tools and the community with its rules and division of labour of the reflection activity, and tools and communities, as well as other constituent entities, of connected activities. 5 CONCEPT DEVELOPMENT ALONG FOUR THEMES: OBJECT, CONTEXT, TIME AND REPRESENTATION In this section we develop the concept of the reflection object along four themes: the first three (identifying the object, identifying contextual information and its development over time) are derived as consequences from conceptualizing the reflection object in the sense of activity theory, and by bridging from such derived characteristics to conceptualizations of reflection and designing for it. The fourth theme discusses the role and desirable characteristics of the object’s representation. This fourth theme is tool oriented and goes back to the initial motivation for this paper, to develop a theoretical basis for what existing tools for reflection already do, namely represent the reflection object. While also the fourth theme builds upon activity theory, it adds perspectives from computer-mediated literature and wider HCI literature. Every theme is first developed theoretically, always in relationship to activity theory and reflection theory, as well as in relationship to literature. Secondly, every theme is explained by the use of example cases from the literature. In this elaboration, the examples serve to explain each theme and to show, by example, the applicability of the concept of reflection object and the four themes. We highlight here that activity theory is an analytical tool. This means that when describing example cases, there may be multiple ways to model the case using activity theory. The decision criterion of a good way of modelling the case is that it must be helpful for thinking about the case, and of course overall represent the case in a reasonable manner following the understanding of activity theory and the understanding of the case. In particular, for all example cases, the modelling and decision of what to understand as underlying reflection object is ours and is post-hoc. All examples are taken from peer-reviewed, published literature. Examples could therefore be framed as secondary analyses of published cases (cp. Yin, 1994). However, the example cases are not fully described and analysed as cases; rather elements of these cases are used to explain the four themes that we have developed. Moreover, while the list of examples in the introduction section has been sampled to cover a broad spectrum (purposeful sampling along the dimensions of published in different communities, supporting individual as well as collaborative reflection, along the spectrum of fully automatic to fully manual data collection, supporting reflection for working, learning or private life), the example cases used in this section are fewer; this has to do with the level of relevant details that appears in published papers (where published details certainly fulfil each publication’s goal, but sometimes do not include sufficient detail for our secondary analysis). Subsequently, we have overall preferred the looser term ‘examples’ over the more rigorous term ‘case studies’. Thirdly, in a discussion subsection per theme, we synthesize the theoretical theme development and theme explanation into prototypical, heuristic questions that could be asked at design time, and we discuss the additional complementary perspective that using the concept of reflection object brings with it. The role of the questions for design time in the overall line of argument is to serve as a tool to synthesize succinctly the lengthier descriptions made in the respective theoretical theme development and examples above the corresponding discussion subsection. In parallel, we hope that the questions for design time constitute easy-to-excerpt elements of this paper that will be relatively easy for practitioners to use, if they wish to do so. Following this synthesis, we discuss what we see as the additional value of the reflection object and each particular theme. This discussion serves to show that while of course designing for reflection is and has been possible without having this concept at hand, it adds a specific and complementary perspective. In summary, the development of the concept along four themes, and each theme’s description, has the following roles in the context of this paper’s line of argumentation: to develop the core concept of the reflection object further by going more in-depth in relationship to activity theory as well as other relevant theories; to explain the concept by making it more concrete via examples; to showcase the applicability of the concept via these examples; and to make more concrete what value this concept can bring. 5.1 The reflection object: something that motivates reflection When we conceptualize ‘reflection as activity in the sense of activity theory […] then the object of the reflection activity is the reflection object’ (cp. ‘Using activity theory as a tool in order to conceptualise the reflection object’ above). In this section we ask what further characteristics of the reflection object can be derived from activity theory. These characteristics serve as criteria to decide on how to think about reflection in a concrete use case, and what to choose as reflection object. From framing reflection as activity in the sense of activity theory, we derive firstly that the reflection object is an overarching entity that motivates and directs reflection. It is meaningful and valuable as a whole. It is not a small, isolated something that one thinks about for a minute, but is larger and is more than what is reflected upon in a limited time frame. Secondly, we derive from activity theory that the reflection object is changed through the activity of reflection carried out by the subject (=learner, reflective practitioner, the group of reflection participants when we consider a collective subject). Thirdly, we derive that the reflection object can be of a conceptual as well as of a physical/material nature. Thirdly, we derive that the reflection object is both internal to the subject and external to the subject. 5.2 Examples In this section we give examples for what could be modelled as a reflection object. In care-oriented professions such as medical care, ‘treatment of patient <X>’ could be a meaningful reflection object. As a caveat: for brevity, and all medical practitioners may forgive us, we use ‘treatment’ as shorthand for ‘diagnosis and treatment’, and also do not strictly differentiate between medical treatments and treatments following other disciplines, nor between medical treatments and care. Treatment is meaningful to practitioners in medical care, as this is what they essentially do. Medical practitioners typically are highly motivated to reflect on treatment of patients as improving every single treatment is in line with their professional role, and as medical practitioners have significant power over single treatments. Understanding ‘treatment of patient <X>’ as a reflection object is particularly useful where care for every concrete patient is challenging and potentially needs continual monitoring and adjustments; where for every given patient in care, treatment is typically repeatedly thought about, discussed, re-iterated, maybe changed, evaluated, etc. In such cases, treatment is not something small, and not something that is reflected on within small time frames. This is the reflection object underlying the tool CaReflect (Müller et al., 2015). Of course, in some environments most or all cases are routine, and treatment is quickly decided. In such environments, ‘treatment of patient <X>’ is probably not a good way to model the case; and thinking about the ‘overall treatment process’ may be a more useful way to model reflection and the reflection object (which, of course, would change the focus of design, and impact all subsequent considerations). Another meaningful reflection object in medical care could be ‘communication with patients and their relatives’. This is helpful in cases where conversations with patients and their relatives tend to be difficult and important for decision-making, as for instance in emergency or palliative departments. This is the reflection object underlying the tool TalkReflect as evaluated in (Prilla, 2015). In other cases, ‘communication with patients and their relatives’ might not be experienced as an object that is valuable on its own, but rather as an action that is simply part of treating a particular patient, or part of the more administrative treatment process, and hence not especially motivating to reflect on for medical care staff in a department dealing with less critical medical conditions. Again, in such cases, ‘communication with patients and their relatives’ might not be the best way to frame the reflection object. In such a case, it may make more sense to frame the reflection object differently with the result that communication remains only one of multiple aspects in care that might or might not turn out to be relevant. On the other hand, a patient’s medical data do not constitute a reflection object, as data per se are not a meaningful object. Data are only valuable because they represent aspects of the patient that are relevant for diagnosing or treating the patient. 5.3 Discussion During design, identifying the reflection object in the sense of asking what should be reflected on and should be changed, scopes and focuses design, and moves the questions of reflection and outcome (=change) to the same level. Following activity theory, where the objects of activities are considered to also be motivating the activity, suggests to also immediately think about motivation for reflection. Prototypical questions to be asked at design time in order to identify the reflection object are as follows: What do target users want to reflect on? What do target users want to change? Why are users motivated to reflect on the object? What about the reflection object is valuable to the users and their environment? Some aspects of the reflection object may be more valuable in linked activities than others. As an overarching statement of complementarity that the reflection object brings, we go back to our analysis of the gap in literature from above: ‘[prior] works do not clearly bind together that which is reflected on with that which is changed’. Hence, existing discussions of prototypes mostly separate what is reflected on (proximity sensing data, proximity between carers and residents in Müller et al., 2015, for instance) and what is changed (self-knowledge about own professional behaviour in Müller et al., 2015, for instance). This is not necessarily wrong; however, it provides a fragmented view. For Müller et al. (2015) for instance, using the concept of reflection object, we would for instance suggest to use ‘care of resident X’ as reflection object (see Examples section above). This would better enable to think about the relevance of reflecting on proximity for caring about a specific resident (connecting one selected relevant aspect to the overall reflection object), and better enable to connect between reflecting for instance on medical treatment of a specific resident, this resident’s daily activities and this resident’s interactions with caregivers as multiple aspects of the reflection object ‘care of resident X’, and finally would systematically point out the relevance of reflecting on carers’ proximity to a specific resident in the context of what is a clear organizational goal of reasonable care homes, namely caring well for every single resident. For showing the potential of the reflection object to support focusing both on that which is reflected on and that which is changed at the same time, one particular case may be interesting as it describes a failed design study: the first author of this paper has been able to constructively comment on a design study on tracking menopause symptoms (commentary published in Homewood, 2019). Using the concept of reflection object, it became clear that a key open question in this design study is what would be the reflection object represented by the collected data? So, what would be something that is both meaningful and motivating to think about, and that women in menopause can change? This cannot be menopause itself, since this cannot be changed. 6 CONTEXT: ADDITIONAL INFORMATION ABOUT THE REFLECTION OBJECT In activity theory, the activity is the context that is necessary to understand its constituent parts. In principle, activity theory claims that a human cannot be understood without understanding the activities (s)he carries out; that a tool cannot be understood outside the activities in which they are used, etc. In the above framing of reflection as activity, we have developed the notion that reflection is linked to other activities: ‘people reflect on a linked activity, or more specifically on tools, the community, community rules, or the division of labour in linked activities […] or […] about another activity’. Reflection is therefore linked to other activities via the reflection object. Subsequently, there is of course a problem in design that needs resolution: the reflection object can only be fully understood, following activity theory, by fully considering the linked activity. On the other hand, the linked activity cannot be fully represented in a computational tool, as every representation is a choice, a selection and hence intrinsically incomplete. In design, the solution is to make a trade-off between needing to represent the linked activity and needing to focus on the most important aspects of it to remain feasible. This trade-off appears as representing context, or contextual information. In the literature on reflection tools, context appears exactly in this sense: for instance, in (Fessl et al., 2017), the authors describe contextualization components, which are technical tool features that help to contextualize ‘main data captured within an application’ in the sense of further explanations. This was identified as a necessary functionality via user-centred design methods by the authors. Li et al. (2010) have found, when exploring what questions people ask of collected data, that context was something people were looking for; again, in the sense of context being something that explains data that are captured in quantified self-tools. This understanding aligns with that of Dourish (2004) in understanding context as additional information that is helpful and necessary in order to create meaning in relationship to a particular activity (Dourish, 2004, p. 4). Furthermore, such additional information can be understood as a communicative context, in the sense of being a prerequisite for reflection as inner conversation with oneself, or for reflection as a conversation with others in the case of collaborative reflection (Herrmann and Kienle, 2008). Hidden behind the discussion of context as additional information is the understanding that what is represented in tools only ever represents parts of all information relevant for reflection, and hence context constitutes a selection of what is thought to be most important. This selection is done both at design time by the design team, and at use time by users. Synthesizing, we see that any partial representation of the reflection object within a tool may need additional, contextual, explanatory information, which helps to explain and describe the reflection object. In tool design, contextual information therefore needs to become part of the representation of the reflection object. In related works (e.g. Li et al., 2011; Fessl et al., 2017), such contextual information is not always captured upfront but is sometimes created via the reflection activity. For instance, prompts were used to incite users to give contextual information in Fessl et al. (2017) after the respective events/experiences. This strengthens the point we made above of seeing this contextual information as constituting communicative context that enables reflection as communication (either of an inner type, or with others). Such contextual information therefore also constitutes visible evidence for reflection, in that it captures changing insights and perceptions. 6.1 Example In this section we discuss an example of what contextual information is that helps to explain the reflection object. The discussion is based on the example of KnowSelf as described in Pammer and Bratic (2013). Activity log data capture only a partial aspect of time management, albeit a relevant one—that of actual time use. In order to reflect on time management, however, fine granular time use needs to be related to overarching tasks, projects and activities. In order to reflect on scheduling, time use also needs to be understood in terms of urgency and importance of actions, and interruptions need to be understood in terms of the potential to remove them in the future. Such additional, contextual information helps to explain and understand the activity log data with respect to time management. In the process of explaining data, and questioning own behaviour as visible within such data, reflection may occur, or vice versa, via reflecting, the significance of data and own behaviour with respect to time management may become clearer. In tools, users may manually label time with relevant labels or take additional notes in which they describe and reflect on the relationship of actions performed and overarching activities. Such contextual information is not necessarily captured within the same tool or indeed in any tool, but it still needs to be available for reflection. 6.2 Discussion During design, the consideration of context is therefore one of making pragmatic trade-off decisions on necessity or usefulness, versus feasibility. The above theoretical development of what context is in relationship to the reflection object suggests to ask the following prototypical questions at design time: To which activities is the reflection object linked? What data and informative artefacts are created within these activities? This is to identify available data and information. What data formats represent contextual information? Text, photos, videos or audio files would be standard formats, but one does not always need to design for all of them. In parallel, contextual information may be available in other formats as well. The complementary perspective that is added through our development of the concept of reflection object is that the reflection object constitutes a link to linked activities: in all related works, contextual information relates to what we have labelled linked activities in the present paper. This means that contextual information refers to linked activities. In other words: one role of contextual information is explaining and relating the reflection object to linked activities. This understanding allows focusing in design, pragmatically, on understanding which data or other digital artefacts can be re-used for reflection. At a more conceptual level, the concept of reflection object allows to focus in design not only on reflection, but on its relevance and role with respect to those linked activities, and thereby extends existing discourse on the process, qualities of and conditions for reflection, as well as the roles of technology in reflection (see background section), with connecting reflection and change (attention: not only behaviour change), and with connecting reflection to linked activities. 7 THE OBJECT IN TIME The notion of change is central to reflection: in reflection theory, reflection is understood to lead to learning, i.e. a change in the learner’s knowledge, behaviour and perception (cp. Boud et al., 1985). Both Boud et al. (1985) and Schön (1983) understand that future experience(s) and work practice are impacted through reflection. Also, all the design-oriented works discussed above share the understanding that reflection leads to change, or transformation. From the perspective of activity theory, all constituents of an activity develop over time. In particular, it is understood that through any activity, the subject changes and learns (Kuutti, 1996, p. 32; Kaptelinin and Nardi, 2009, p. 110). In parallel, it is understood that ‘the subject is transforming the object’ (Kuutti, 1996, p. 32). Both the internal and the external object can change. When the internal object changes, this change is at first unobservable from the outside. This perspective from activity theory captures how subjects change their understanding of and perspective on the reflection object via reflection. This aligns with the perspective of reflection theory: reflection means to change knowledge, behaviour and values. In particular also, subjects may change their understanding about what are aspects of the reflection object that are valuable to reflect on. These changes are because the subject engages in the activity of reflection. As a consequence for design, over time, new and different kinds of contextual information may become necessary to understand and reflect on the reflection object. When the external object changes, this change is immediately observable. Observable changes can for instance be the learner’s behaviour, or changes made to artefacts that document the reflection activity, or that document linked activities. 7.1 Examples The example we discuss here is that of understanding ‘communication with patients and their relatives’ as a reflection object in a case described in (Prilla, 2015). Here, the activity of conversing with relatives and patients is the reflection object. Such a reflection object then includes a physician’s understanding of conversation situations as well as the situation of patients and relatives, the locations of the conversation and the way in which the conversations were conducted. This understanding may be questioned through reflection and may change over time. Internal, subjective aspects on the one hand and observable aspects of further conversations on the other hand may be different. However, both a change in understanding and a change in further conversations may be observable in the documentation of the reflection activity (e.g. in a discussion of conversations with a particular person). 7.2 Discussion During design, the consideration of the object in time goes back to re-considering what exactly should change, and how this change would be noticeable, both by the subject (=reflective practitioner) and by others (others participating or connecting in the subject’s activities; researchers). Prototypical questions to be asked at design time are as follows: What should change? Overall, the answer of course needs to be the reflection object and its understanding. Firstly, it makes sense to be as clear as possible in identifying what should change/how the reflection object should become different, and secondly, whether this is feasible. Which parts of the change are external and observable, and could help make the reflection activity and outcomes noticeable? How is transformation achieved? Is there a particular procedure or are there specific activities such as engaging in conversation, data analysis, etc. How can the results of this transformation be preserved and made accessible? At a concrete level, many tools intended to support reflection by collecting data and visualizing it, do of course visualize collected data over time (cp. Li et al., 2011; Müller et al., 2015; Kay and Kummerfield, 2011; Pammer and Bratic, 2013). Of course, this shows the development of that particular aspect of the reflection object in time, but does not refer back to the change of the overall reflection object in time. At a conceptual level, considering the development of the object in time constitutes a unique perspective: others who have thought about reflection have, for simplicity’s sake, focused on consecutive phases (e.g. Boud et al., 1985; Li et al., 2010) or rejected explicating sequentiality (Baumer, 2015), or made explicit iterations of the same steps (Krogstie et al., 2013; Pammer et al., 2017). Such perspectives do not make it easy to see the sameness of that which is reflected on, and changed, while in parallel seeing the change. By this we mean, that while one’s lifestyle, or the activity of caring for a specific patient, are throughout the activity of reflecting on it, in some way the same reflection object (the object is not suddenly switched to another object—that would be another reflection activity), of course the object changes, in the sense that ideally one’s lifestyle becomes healthier, and ideally the care for a specific patient changes in response to the health status of the patient for instance. In the same manner that a person stays himself or herself throughout life, while in parallel of course changing; the reflection object stays itself, while in parallel also changing. Finally, the theme of the object’s development in time is particularly challenging and provides directions for further research, especially on how to measure the change, and how to bring about this change. There are guidelines and tools for reflection that are more specific and actionable, and often also more domain-specific, than the foundational literature discussed in this paper. An example is Kerth’s guideline to project retrospectives (Kerth, 2001). Overall, however, there is by no means a sufficient body of literature on how to effectively structure reflection using computational tools, or within computational environments. Existing literature in human–computer interaction has focused both on tools used in interactions with professionals whose responsibility it was to structure reflection (e.g. Mamykina et al., 2016), and on tools where no such human facilitator was available, i.e. where reflection prompts and reflection guidance worked to support reflection in self-regulated learning scenarios (e.g. Shin et al., 2018—reflection prompts in the context of video lectures; Fessl et al., 2017—successful field studies of adaptive reflection guidance in workplace settings; Renner et al., 2016 and Ifenthaler, 2012—experimental indications that prompts are more effective when they are more directed). 8 REPRESENTING THE OBJECT Finally, we delve into the theme that was the starting point of the present paper: the reflection object is represented in all tools that we know of, and certainly in all tools that we cited in the introduction. The reflection object is often represented via automatically or manually tracked data. In addition, textual or multimedia descriptions, metadata such as notes, tags, ratings, and all kinds of communications about the reflection object are used. Indeed, anything that explains or describes the reflection object within a tool can be part of its representation. Of course, as we understand the reflection object to be an overarching entity that exists outside tools for reflection, any such representation is only partial. From the perspective of activity theory, we derive that the reflection object has a dual nature in that it has internal (to the human actor, the reflective practitioner) and external aspects. The internal reflection object is the representation of the reflection object in the subject’s mind, as a mental model of that which is reflected on. The external reflection object is the reflection object as it exists outside the subject in its own right. The representation of the reflection object is now again something external, but it is only a representation. It has a specific purpose, namely to support reflection. The overall goal for the representation of the reflection object is to come as close to the internal and external reflection object as necessary. 8.1 The role of the representation From activity theory we derive that tools shape the interaction of the subject with the object. This means that the representation of the object—whether and how it is represented, how the object can be interacted with—acts as modifier to the reflection activity itself. This understanding has also clearly influenced the development of the Ghajargar and Wiberg (2018)’s understanding of smart objects as being useful to enable reflection; albeit with a further focus in the authors’ conceptualization of how the user (=subject of activity, reflective practitioner) interacts with and relates to the artefact (Ghajargar et al., 2018). From the perspective of reflective learning, we can in addition understand the representation of the reflection object to serve as external trigger (cp. Krogstie et al., 2013; Pammer et al., 2017) for reflection. This means, that one role and function in reflection is that the representation creates awareness of something that is worth being reflected on, and points towards a salient aspect of the reflection object. From the perspective of communication, the object’s representation serves to create communicative context both for inner conversations with oneself and also for conversations with others (Herrmann and Kienle, 2008). Thereby, the representation aims to mediate and stimulate critically reflective dialogue. We found it particularly useful to consider the similarity of the role of the reflection object’s representation to create communicative context with the role of boundary objects (Star, 1989). Boundary objects are artefacts that are known by different groups of people and for which each group has a different conceptualization or makes use of a different practice for dealing with it. Thereby, boundary objects raise different interpretations and thoughts among different people and groups, and overall support articulation and communication in the group. Examples put forward in the literature are maps used and interpreted differently by different people, or work requirements affecting people differently but enabling them to talk about work (Star, 2010). In collaborative reflection there may be no, or only a weak boundary between people in the group, as the reflection object typically constitutes shared concepts, shared practice, etc. Nonetheless, the role of the representation of the reflection object is similar: it needs to stimulate the explication of interpretations, thoughts and ideas for dealing with the issues reflected on within the group in order to stimulate collaborative reflection. The representation of the reflection objects thus needs to include different facets and aspects, and it also needs to allow the addition of thoughts and interpretations, which in turn documents changes in the shared reflection object. Finally, from prior discussions on contextual information about the object and the object in time, we know that the reflection object’s representation in time also changes. A change of this kind documents both the ongoing reflection activity and its outcomes. Tools can therefore take on the role of making changes visible. This role is not taken on by existing tools in related work, however, as no existing tools can analyse and emphasize the development of the object in time. 9 DESIRABLE CHARACTERISTICS OF THE REPRESENTATION The different roles of the representation of the reflection object, to mediate and enable reflection, to trigger reflection and to support the creation of communicative context, lead to a variety of desirable characteristics for the representation. Firstly, relevant aspects of the reflection object need to be represented. Existing works on data-driven tools for reflection highlight the importance of choosing relevant data that represent multiple facets of the reflection object. In the case of CaReflect for instance, the authors discuss critically that measuring the time spent with residents is only one aspect of care quality: ‘by putting the focus on one aspect, there is a risk that the others are neglected’ (Müller et al., 2015, p. 121). We note that the authors refer to ‘aspects of the reflection object’ without explicitly conceptualizing the reflection object as such in their paper. Also Li et al. (2011) note that tools for data collection tend to be specialized on a very limited number of data types. Designers then need to be aware of the risk that users may choose to switch tools in case they decide that another aspect of the reflection object is more relevant for reflection, and hence need another representation. Choosing relevant aspects also means taking into consideration that the representation serves as trigger for reflection: The representation needs to be able to focus users on something that is meaningful, interesting, relevant to reflection and not immediately obvious to users without the tool. Secondly, the role of the representation is to support articulation and communication and to raise and document different interpretations and thoughts. This requires the representation to be elastic (Star, 1989) and offer interpretative flexibility (Star, 2010). By this it is meant that as the reflection object is in the process of being critically re-considered, tool design should allow multiple interpretations, and tool design should leave space for users to annotate and adapt the reflection object’s representation. Thirdly, the role of the representation to support communication in collaborative reflection settings of course requires designers to think of issues like sharing, anonymity versus author identification, etc. We highlight in particular that the group of people who reflect can sometimes change. Such a change in reflecting subjects could happen for instance because new people ‘bring in new expertise or executive power into the group’ (Pammer et al., 2017, p. 10). In such cases, it may even be desirable to consider a complete ownership transfer for the representation of the object. Finally, representations of the reflection object could take up the role of making the temporal development of the reflection object visible in time by visualizing and emphasizing changes to the representation. This is not done in existing tools. 9.1 Examples In this section, we discuss an example of how a reflection object can be represented for the case of reflecting on time management based on Pammer and Bratic (2013). Activity log data about application and document usage are automatically collected. Activity log data capture time management only partially, such that additional contextual information is definitely needed for reflection. In particular, one representation within the research prototype (ibid) focuses on representing worktime fragmentation, which is a source of inefficiency and stress. The representation thereby takes the role to trigger reflection on a particular aspect of the reflection object by design. Secondly, the representation within the used research prototype focuses on automatically captured data. Interpretative flexibility and elasticity were implemented via free-text labels, and free text notes in the form of a running diary. In a follow-up study, the authors have also investigated user-directed data analytics to provide additional flexibility (cp. Luzhnica et al., 2016). Communicative context is created both within KnowSelf, in labels and notes, and outside KnowSelf in discussions with peers. 9.2 Discussion During design, the question of how to represent the reflection object can be expected to occur before considering context or development over time. It is a pragmatic as well as a creative question. It is pragmatic because it focuses on what is possible and creative because the representation does not naturally derive from having developed an understanding of the reflection object, but really is a design task. Having said that, prototypical questions to be asked at design time are: Which aspects of the reflection object are meaningful, interesting, relevant and not easily obvious to users without the tool? What data and digital artefacts can represent relevant aspects of the reflection object? Should the change in the reflection object, or the change in the representation of the reflection object, be emphasized in the reflection tool, and how? The concept of reflection object adds a definitely complementary rather than superior perspective to understanding the role of (computational) tools for reflection: activity theory in general sees the role of tools for an activity as both enabling and constraining; this is of course also true for the reflection object’s representation. Krogstie et al. (2012) discuss in a fine-granular manner the role of computational tools for reflection, among those to capture data and provide relevant information, albeit without saying about what (we would argue: data and information that represents and explains relevant aspects of the reflection object). The authors focus on the role of tools in the reflection process, but do not differentiate between process support, and functionality that serves to represent the reflection object. Ghajargar et al. (2018) discuss the interaction between a computational tool that aims to make the user think (=reflect; borrowing on the authors’ phrase ‘make me think’). In relationship to activity theory, the authors focus on the connection between the subject (=user, reflective practitioner) and computational tools in the activity of reflection, and thereby on characteristics of the tool, and design choices that enable the subject to reflect. In this work, the overall focus on the concept of the reflection object adds the perspective that a substantial role of (computational) tools for reflection is to represent the object of reflection. 10 10 DISCUSSION AND CONCLUSION In this section, we summarize and discuss the contributions that are made in this paper. In particular, the discussion centres on the concept of the reflection object in terms of its complementarity for thinking about reflection, and in terms of methodological issues. 11 CONTRIBUTION Our work extends current work on designing for reflection by developing the concept of reflection object, using activity theory as a grounding theory. The reflection object is the object of the reflection activity. This emphasizes that there are mutual dependencies between the reflection activity and what it is about. This fills a gap in existing work and enables designers to weave together threads when designing for reflection. In particular, we showed that activity theory constitutes a framework within which to think about reflection and designing for reflection, and we inherit broad insights about reflection and reflection tools from using this framework. Firstly, our work emphasizes that through the reflection activity, both the subject (the person who reflects) and the reflection object change, but do not lose their identity. The reflection object is therefore at the same time that which is reflected on, and that which is changed through reflection. The second inheritance from activity theory is that the tools we design for reflection influence, impact and shape how reflection is carried out. This also includes the representation of the reflection object, which we have discussed at length in this paper. The object’s representation is a tool for reflection; it triggers reflection and creates communicative context. Therefore, designers need to carefully think about how to represent the reflection object. Thirdly, by combining activity theory and reflection theory, we emphasize that reflection is connected to other activities in which the subject participates via the reflection object. It is for these linked activities for which reflection is valuable, and additional information that explains the reflection object in within these linked activities needs to be available at the time of reflection as context. Overall, we understand the reflection object as a conceptual tool that creates a focus and structure to think about reflection and designing for reflection. This concept as a tool is our contribution to literature on designing for reflection. Designing for reflection, however, still contains substantial creative elements that do not logically or necessarily follow from given pre-requisites but are designed, invented, claimed as being a good solution (cp. also Fallman, 2003; Löwgren, 1995; Löwgren and Stolterman, 2007). Designers who use the concept of ‘reflection object’ therefore do not have at their hand immediately a finished design. On the other hand, ‘nearly every design situation requires analytical skill from the designer’ (Löwgren and Stolterman, 2009; Section 3.1.3). Designers who use the concept of ‘reflection object’ have at hand exactly this: a concept that helps to focus and structure to think about reflection when designing for reflection. In particular, given the above key insights about the reflection object, the reflection object requires the designer to make the following connections by requiring to think in parallel: i) about that which is reflected on, and that which changes ii) about the sameness of and change in the reflection object (the object keeps its identity, even though it is changed) iii) about reflection and the linked activity, which is either as a whole or in parts the reflection object, and for which reflection is important. These connections constitute a novel perspective when thinking about and designing for reflection. Re (i): This differs substantially from the way of thinking communicated in most learning analytics, quantified self and similar literature: to review the past, or to analyse data in order to understand something or to change some behaviour. Such communication leads to an artificial separation between that which is reflected on and that which is changed. When people reflect on their blood sugar values (Mamykina et al., 2016) they are looking to change their food intake. Such perspectives aren’t wrong of course, but the connection is more hidden. Re (ii): Li et al. (2011) report that people move between tools for data collection, and Krogstie et al. (2013) and Pammer et al. (2017) highlight how iterative reflection is. This can be understood differently if we see reflection as an overarching activity with a reflection object that stays the same—in terms of identity—over time, but is of course changed through reflection (and through the linked activity as well). In addition, through the activity of reflection, also the subject, and the overall activity changes such that over time, different aspects of the reflection object could be focused on. This change in focus could explain some of the switches between tools (Li et al., 2011), and both the change in focus and the change of the object can explain the cyclic iterations (Krogstie et al., 2013; Pammer et al., 2017). Regarding the design of reflection tools, this provides two directions. First, it may be beneficial to track the changes the reflection object undergoes over time to understand and reflect on these changes. Second, to deal with switches in tools, designers need to find ways how to maintain the identity of the reflection object across these tools. Again, the existing perspectives in others’ works are not wrong; understanding reflection as an overarching activity simply allows focusing on something different, in this case overall stability (same reflection object over time) in parallel to ongoing change (reflection object as changed through reflection and linked activities). We argue that this perspective in particular is necessary for tool designers and researchers to better support learning, change and transformation; something that is perceived as an open challenge (cp. e.g. Baumer, 2015). Re (iii): Finally, contextual information appears in concrete designs, and the concept of context is well known (even though difficult) in human–computer interaction. Understanding the reflection object as a linked activity, or important elements of this activity, creates a conceptual connection between reflection and those activities for which reflection is valuable. Through establishing these connections—especially seeing that which is reflected on and that which changes as one thing, namely the reflection object—the concept of the reflection object provides a frame to focus on how to support change and transformation, which is a major challenge when designing technologies for reflection. 12 METHODOLOGY This work constitutes a theory development work. In a field dominated by empirical research such as human–computer interaction, such a work warrants reflection in terms of methodology. Firstly, with activity theory, we have used a mature theory that is accepted in HCI as a basis for concept development. The key step in this was to frame reflection as an activity. Next, from framing reflection as an activity, key characteristics of the reflection object were derived, such that it is in itself valuable and motivating to the reflective practitioner (theme: the reflection object), that in order to understand the reflection object—when it is an element of a linked activity rather than a complete linked activity, understanding of the complete linked activity in which the object exists would be necessary (theme: context), that the reflection object develops in time and that this is true both for those aspects that are internal to the subject and those aspects that are external to the subject (theme: the object in time) and that the representation of the object is a tool that influences the reflection activity (theme: representing the object). These derivations are direct, nearly verbatim transferals from activity theory to reflection. We have shown applicability by example cases in the above section ‘Concept development along four themes: object, context, time and representation’. Such a concept development is an inductive conceptual process, in this case inductive through discussion among authors, grounded in the authors’ design experience. Most of the related works discussed here are not empirical. However, all of them are informed by and grounded in the respective authors’ practical experiences in research and design. Accordingly, all of them regularly make use of selected examples to illustrate the developed conceptual structure (Baumer, 2015; Boud et al., 1985; Fleck and Fitzpatrick, 2010; Ghajargar et al., 2018; Pammer et al., 2017; Schön, 1983). Li et al. (2010) and Li et al. (2011) are exceptions in that they directly ground their theory development in mixed methods empirical data, and Slovák et al. (2017) in a case study. The usefulness of these works has been shown by other researchers and designers making use of the concepts and ‘ways of thinking’ laid out in these papers as tools that helped them think about reflection during design. The usefulness of the concept of reflection object is therefore something that time will have to show. We certainly hope that future design works make use of our concept, both to show its applicability and to inform theory. Beyond this, we would of course be happy for further theoretical papers to emerge on designing for reflection that take up, complement or critically discuss the concept of a reflection object. FUNDING This work has been funded by the European Commission (25761), and by the Austrian Research Promotion Agency under the COMET Programme—Competence Centers for Excellent Technologies under the auspices of the Austrian Federal Ministry of Transport, Innovation and Technology, the Austrian Federal Ministry of Economy, Family and Youth and by the State of Styria. References Baumer , E. ( 2015 ) Reflective informatics: conceptual dimensions for designing technologies of reflection. In Proc. 33rd annual ACM conf. human factors in computing systems (CHI ‘15) , pp. 585 – 594 . Baumer , E. , Khovanskaya , V., Matthews , M., Reynolds , L., Sosik , V. S. and Gay , G. ( 2014 ) Reviewing reflection: on the use of reflection in interactive system design. In Proc. 2014 conf. designing interactive systems (DIS ‘14) , pp. 93 – 102 . Boud , D. , Keogh , R. and Walker , D. ( 1985 ) Promoting reflection in learning: a model. In Reflection: Turning Experience Into Learning , pp. 18 – 40 . Routledge Falmer , New York . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Choe , E. K. , Lee , N. B., Lee , B., Pratt , W. and Kientz , J. A. ( 2014 ) Understanding quantified selfers’ practices in collecting and exploring personal data. In Proc. SIGCHI conf. human factors in computing systems (CHI 2014) , pp. 1143 – 1152 . Crowston , K. and Kammerer , E. E. ( 1998 ) Coordination and collective mind in software requirements development . IBM Syst. J. , 37/2 , 227 – 245 . Google Scholar OpenURL Placeholder Text WorldCat Daudelin , M. W. ( 1996 ) Learning from experience through reflection . Organ. Dyn. , 24/3 , 36 – 48 . Google Scholar OpenURL Placeholder Text WorldCat Dourish , P. ( 2004 ) What we talk about when we talk about context . Pers. Ubiquit. Comput. , 8/1 , 19 – 30 . Google Scholar OpenURL Placeholder Text WorldCat Fallman , D. ( 2003 ) Design-oriented human–computer interaction. In Proc. SIGCHI conf. human factors in computing systems (CHI '03) , pp. 225 – 232 . ACM . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Fessl , A. , Wesiak , G., Rivera-Pelayo , V., Feyertag , S. and Pammer , V. ( 2017 ) In-app reflection guidance: lessons learned across four field trials at the workplace . IEEE Trans. Learn. Technol. , 10/4 , 488 – 501 . Google Scholar OpenURL Placeholder Text WorldCat Fleck , R. and Fitzpatrick , G. ( 2010 ) Reflecting on reflection: framing a design landscape. Proc. 22nd conf. computer–human interaction special interest group of Australia on computer–human interaction (OZCHI ‘10) , pp. 216 – 223 . Ghajargar , M. and Wiberg , M. ( 2018 ) Thinking with interactive artifacts: reflection as a concept for design outcomes. In Design Issues , vol. 34/2 , pp. 48 – 63 . MIT Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Ghajargar , M. , Wiberg , M. and Stolterman , E. ( 2018 ) Designing IoT systems that support reflective thinking: a relational approach . Int. J. Des. , 12 , 21 – 35 . Google Scholar OpenURL Placeholder Text WorldCat Herrmann , T. and Kienle , A. ( 2008 ) Context-oriented communication and the design of computer-supported discursive learning . Int. J. Comput.-Support. Collab. Learn. , 3/3 . Google Scholar OpenURL Placeholder Text WorldCat Homewood , S. ( 2019 ) Inaction as a design decision: reflections on not designing self-tracking tools for menopause. Proc. SIGCHI conf. human factors in computing systems , paper No.: alt17 . Ifenthaler , D. ( 2012 ) Determining the effectiveness of prompts for self-regulated learning in problem-solving scenarios . Educ. Technol. Soc. , 15/1 , 38 – 52 . Google Scholar OpenURL Placeholder Text WorldCat Isaacs , E. , Konrad , A., Walendowski , A., Lennig , T., Hollis , V. and Whittaker , S. ( 2013 ) Echoes from the past: how technology mediated reflection improves well-being. Proc. SIGCHI conf. human factors in computing systems (CHI ‘13) , pp. 1071 – 1080 . Kaptelinin , V. and Nardi , B. A. ( 2009 ) Acting With Technology. Activity Theory and Interaction Design . MIT Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Karanasios , S. ( 2018 ) Toward a unified view of technology and activity: the contribution of activity theory to information systems research . Inf. Technol. People , 31/1 . Google Scholar OpenURL Placeholder Text WorldCat Kay , J. and Kummerfield , B. ( 2011 ) Lifelong learner modeling. In Adaptive Technologies for Training and Education . Cambridge University Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Kerth , N. L. ( 2001 ) Project retrospectives: A handbook for team reviews. Dorset House Publishing Co . Krogstie , B. , Prilla , M., Wessel , D., Knipfer , K. and Pammer , V. ( 2012 ) Computer support for reflective learning in the workplace: a model. In 12th IEEE int. conf. advanced learning technologies (ICALT 2012) . Krogstie , B. , Prilla , M. and Pammer , V. ( 2013 ) Understanding and supporting reflective learning processes in the workplace: the CSRL model. Proc. 8th European conf. technology enhanced learning, (EC-TEL 2013) , pp. 151 – 164 . Kuutti , K. ( 1996 ) Activity theory as potential framework for human–computer interaction research. In Context and Consciousness—Activity Theory and Human–Computer Interaction . Press , The MIT Chapter 2 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Leontiev , A. ( 1981 ) Problemy Razvitija Psihiki. Izdatelstvo MGU , Moscow . Li , I. , Dey , A. and Forlizzi , J. ( 2010 ) A stage-based model of personal informatics systems. Proc. SIGCHI conf. human factors in computing systems (CHI ‘10) , pp. 557 – 566 . Li , I. , Dey , A. and Forlizzi , J. ( 2011 ) Understanding my data, myself: supporting self-reflection with Ubicomp Technologies. Proc. 13th int. conf. ubiquitous computing (UbiComp ‘11) , pp. 405 – 414 . Löwgren , J. ( 1995 ). Applying design methodology to software development. Proc. 1st conf. designing interactive systems: processes, practices, methods, & techniques (DIS '95) . ACM , pp. 87 – 95 . Löwgren , J. and Stolterman , E. ( 2007 ) Thoughtful Interaction Design: A Design Perspective on Information Technology . The MIT Press . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC Luzhnica , G. , Fessl , A., Veas , E., Mutlu , B. and Pammer , V. ( 2016 ) Designing generic visualisations for activity log data. Proc. 6th workshop on awareness and reflection in technology enhanced learning, CEUR WS Proceedings , vol. 1736 , pp. 11 – 25 . Malacria , S. , Scarr , J., Cockburn , A., Gutwin , C. and Gossmann , T. ( 2013 ) Skillometers: reflective widgets that motivate and help users to improve performance. In Proc. 26th annual ACM symposium on user interface software and technology (UIST ‘13) , pp. 321 – 330 . Mamykina , L. , Heitkemper , E. M., Smaldone , A. M., Kukafka , R., Cole-Lewis , H. and Davidson , P. G. ( 2016 ) Structured scaffolding for reflection and problem solving in diabetes self-management: qualitative study of mobile diabetes detective . J. Am. Med. Inform. Assoc. , 23/1 , 129 – 136 . Google Scholar OpenURL Placeholder Text WorldCat Müller , L. , Divitini , M., Mora , S., Rivera-Pelayo , V. and Stork , W. ( 2015 ) Context becomes content: sensor data for computer-supported reflective learning . IEEE Trans. Learn. Technol. , 8/1 , 111 – 123 . Google Scholar OpenURL Placeholder Text WorldCat Pammer , V. and Bratic , M. ( 2013 ) Surprise, surprise: activity log based time analytics for time management . Proc. ACM SIGCHI conf. human factors in computing systems, extended abstracts , pp. 211 – 216 . Pammer , V. , Krogstie , B. and Prilla , M. ( 2017 ) Let’s talk about reflection at work . Int. J. Technol. Enhanc. Learn. , 9/2–3 . Google Scholar OpenURL Placeholder Text WorldCat Prilla , M. ( 2015 ) Supporting collaborative reflection at work: a socio-technical analysis . AIS Trans. Hum.–Comput. Interact. , 7/1 , 1 – 17 . Google Scholar OpenURL Placeholder Text WorldCat Prilla , M. and Pammer , V. ( 2012 ) The push and pull of reflection in workplace learning: designing to support transitions between individual, collaborative and organisational learning. In Proc. 7th European conf. technology enhanced learning (EC-TEL 2012) , pp. 278 – 291 . Renner , B. , Prilla , M., Cress , U. and Kimmerle , J. ( 2016 ) Effects of prompting in reflective learning tools: findings from experimental field, lab, and online studies . Front. Psychol. , 7 . Google Scholar OpenURL Placeholder Text WorldCat Sas , C. and Dix , A. ( 2011 ) Designing for reflection on personal experience . Int. J. Hum.–Comput. Stud. , 69/5 , 281 – 282 . Google Scholar OpenURL Placeholder Text WorldCat Siemens , G. ( 2013 ) Learning analytics: the emergence of a discipline . Am. Behav. Sci. , 57/10 , 1380 – 1400 . Google Scholar OpenURL Placeholder Text WorldCat Schön , D. A. ( 1983 ) The Reflective Practitioner—How Professionals Think in Action (1st). Ashgate Reprint 2006 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Shin , H. , Ko , E.-Y., Williams , J. J. and Kim , J. ( 2018 ) Understanding the effect of in-video prompting on learners and instructors. In Proc. 2018 CHI conf. human factors in computing systems (CHI '18) . ACM . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Slovák , P. , Frauenberger , C. and Fitzpatrick , G. ( 2017 ) Reflective practicum: a framework of sensitising concepts to design for transformative reflection. Proc. 2017 CHI conf. human factors in computing systems (CHI ‘17) , pp. 2696 – 2707 . Star , S. L. ( 1989 ) The structure of ill-structured solutions: boundary objects and heterogeneous distributed problem solving . Distrib. Artif. Intell. , 2 , 37 – 54 . Google Scholar OpenURL Placeholder Text WorldCat Star , S. L. ( 2010 ) This is not a boundary object: reflections on the origin of a concept. In Science, Technology, & Human Values , vol. 35/5 , pp. 601 – 617 . Weick , K. E. , Sutcliffe , K. H. and Obstfeld , D. ( 2005 ) Organizing and the process of sensemaking . Organ. Sci. , 16/4 , 409 – 421 . Google Scholar OpenURL Placeholder Text WorldCat Yin , R. K. ( 1994 ) Case Study Research. Design and Methods (2nd, edn). Applied Social Research Methods Series, vol. 5 ). SAGE Publications . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC © The Author(s) 2021. Published by Oxford University Press on behalf of The British Computer Society. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - The Reflection Object: An Activity-Theory Informed Concept for Designing for Reflection
JF - Interacting with Computers
DO - 10.1093/iwc/iwab027
DA - 2021-10-05
UR - https://www.deepdyve.com/lp/oxford-university-press/the-reflection-object-an-activity-theory-informed-concept-for-YDpqoHCvKu
SP - 1
EP - 1
VL - Advance Article
IS - 
DP - DeepDyve
ER -