All theories comprise a set of assumptions from which empirical generalizations are derived (Merton, 1949). In the social sciences, a satisfactory explanation must ultimately be anchored in hypotheses or assumptions about individual behavior (Elster, 2007). Constructing a set of explicit, behavioral assumptions enables us to underline the point that all macro-level, societal phenomena are inherently derived from human beliefs and actions. Based upon these assumptions, we thereupon apply the method of retroduction to hypothesize about the mechanisms that are capable of explaining the process by which initial conditions work together to produce human actions, which collectively result in a (potentially) observed outcomes at the societal level. Since the core behavioral assumptions of a theory often form the foundation of its mechanistic
explanations, it is crucial to define these assumptions explicitly, during the early stages of empirical research (Tsang, 2006).
4.1 Basic Assumptions
The classical economic model of strategic interaction assumes that people are: a) largely concerned about bettering their own material situation; and b) in order to achieve this, they determine an optimal strategy through perfectly rational judgments.
However, new results from behavioral economics discredit these assumptions, since new evidence discloses that people are, in fact, concerned with unobservable outcomes such as reputation, fairness or the well-being of others, as well as the ways in which their own immediate actions might affect their and others future well-being (Shogren, 2012). A comprehensive review of empirical studies of behavioral assumptions pointed to hundreds of studies that discredit the validity of the assumption of self-interest (Rabin, 1998). Individual preferences may undoubtedly be at odds with public welfare.
This is characterized as a social choice problem, which may be alleviated via constructive social choice (Sen, 1999). Constructive social choice stipulates that individuals do not simply act on their planned personal and known preferences, but interact with each other to rearrange and shape those preferences (York et al., 2013).
Information dissemination through social networks may, as an example, act as a medium for constructive social choice.
The rationality of individuals is limited by the information they possess, the cognitive limitations of their minds, and the finite amount of time they have to make a decision (Shogren, 2012). People typically do not apply sufficient cognitive effort to calculate an optimal strategy, but rather resort to heuristics, which may be influenced by context (Simon, 1957). Bounded rationality is “behavior that is intendedly rational but only limitedly so” (Simon, 1957, xxiv). Limited quality and quantity of information available to firms and individuals introduces a “boundary” to rational behavior, as it reduces an individual’s ability to assess present and future environmental states (Hitt et al., 2011). We must therefore assume that people face a multi-dimensional value function consisting of complex relationships and even contradictions between individual and social well-being. Moreover, we as individuals have limited ability to choose rationally between different options, especially when lacking information.
Proceeding from the above, we should clarify the principal distinction between data and information. The same meaning seems to be frequently attached to these two important concepts in the open data/big data discourse, in some measure due to the novelty of these phenomena. Nevertheless, the data-information-knowledge
relationship is already well established, and constitutes the foundation of the IS discipline (Kettinger & Li, 2010). Data are recorded (captured and stored) symbols and signal readings, whereas information is a message that contains relevant meaning, implication, or input for decision and/or action (Liew, 2007). Knowledge, on the other hand, is reflected in (1) cognition or recognition (know-what); (2) capacity to act (know-how); and (3) understanding (know-why) that is contained within the mind or in the brain (Liew, 2007). If we are to understand the relationship between raw data and value, we must have a basic understanding of the underlying value chain, or perhaps more accurately, value network. It has been suggested that the ultimate purpose of knowledge is for value creation (Liew, 2007).
Kettinger and Li (2010) note that information represents a status of conditional readiness for action, which is generated from the interaction between the states measured in data and their relationship with future states predicted in knowledge (Kettinger & Li, 2010). “Information is the meaning produced from data based on a knowledge framework that is associated with the selection of the state of conditional readiness for goal-directed activities.” (Kettinger & Li, 2010, 415). Thus, new information forms the basis for action through its potential influence on people’s decisions, and consequently, their behavior. An important role of technology is therefore to augment people’s capabilities in dealing with multiple sources of data, in order to generate information as a basis for better operations or decisions (Kettinger &
Li, 2010). Following this interpretation, with the aim of facilitating value generation from data, the data should be transformed to information that serves as a catalyst for enhanced decision-making, applicable behavioral change and the ability to act swiftly, more accurately and with optimal use of resources.
As a foundation for the proposed theory, I advance two basic assumptions.
The first assumption is that individuals in general aspire to go beyond increasing their own material wealth in their efforts at value generation. In other words, most individuals possess the necessary intrinsic motivation for sustainable value generation.
Accordingly, I assume that individuals will strive to generate sustainable value, including the economic, social and environmental dimensions, for all stakeholders and future generations (van Osch & Avital, 2010). However, while intrinsic motivation is a necessary condition for sustainable level value generation at the individual level, it is not sufficient. A class of behavioral theories, based on the Motivation- Opportunity-Ability (MOA) framework, further explored in Paper V, indicates how extrinsic motivation combined with the opportunity and ability to perform certain tasks influences the behavior of individuals (Blumberg & Pringle, 1982; Reinholt et al.
2011). Consistent with this framework, motivation is defined as goal-directed arousal (MacInnis et al. 1992; Rothschild, 1999). In the context of this study, motivation implies that individuals are provided with an incentive to allocate resources to generate sustainable value from data; opportunity refers to the environmental or contextual factors that enable action; and ability represents the power or capacity to act.
The second assumption I advance is based on the notion of bounded rationality. People will be able to make better decisions if provided with information that is relevant to the situation they face. Access to relevant information therefore pushes the boundaries of our ability to choose rationally and contribute to the generation of sustainable value, which is what we wish to do. Recent research on online shopping behavior has supported the notion that increased transparency of information can significantly influence user behavior, consequently contributing to customers being more satisfied with choices they make (Xu et al., 2014). While having access to too much information can have a paralyzing effect, the right amount of information sufficiently contextualized can really help us make better decisions. Why is this assumption relevant for this study? The underlying motive is my proposal that a large part of value generation from use of open data essentially comes about through the creation and sharing of information over networks. Access to this information subtly influences individual behavior, utilizing the forces of constructive social choice to influence collective behavior, and eventually creating value through means such as improved decisions making.
4.2 Social Mechanisms
Let us assume that we have observed a systematic relationship between two types of events, I and O. The way in which the two sets are linked to one another is expressed through the mechanism M: I → M → O. This high-level model represents what can be described as a structural equation, in which the mechanism mediates the effect of I on O. A mechanism can be explained by opening up the so-called black box and thereby making “…explicit the causal cogs and wheels through which effects are brought about.” (Hedström & Ylikoski, 2010, 54). The mechanisms approach thus develops a causal reconstruction of a phenomenon by identifying the overarching processes through which an observed outcome was generated (Bunge, 2004; Hedström &
Swedberg, 1998; Machamer et al., 2000; Mayntz, 2004). A majority of the accepted mechanism definitions across various disciplines share several general concepts (Hedström & Ylikoski, 2010). First, a mechanism is identified by the type of effect or phenomenon it produces. A mechanism is always a mechanism for something (Darden,
2006). In this study, the mechanisms I aim to uncover are for value generation. Second, a mechanism is an irreducibly causal notion. It refers to the entities of a causal process, which produce the effect of interest. Third, the mechanism has a structure. When a mechanism-based explanation opens the so-called black box, it discloses this structure.
Mechanisms are a key construct is many disciplines of science and are defined as frequently occurring and easily recognizable causal patterns (Elster, 2007). “In the natural sciences, no event or process is regarded as having been satisfactorily understood unless its actual or possible mechanism has been unveiled” (Bunge, 1999, p. 63). Mechanisms are not like the deterministic laws of physics, in which certain inputs positively lead to certain outputs (Elster, 2007; Hedström & Swedberg, 1998).
Instead, mechanisms allow us to address the probabilistic nature of social life. In social theory IS research, collective entities and processes unfolding in a social context are understood to typify and represent the generation of aggregate effects from the actions of individuals (Avgerou, 2013). Therefore, a methodological strategy to cope with theoretical complexity is the acceptance of mechanisms in terms of collective actors (ibid). Avgerou (2013) identifies two main spheres wherein social mechanism based explanations augment social theory IS research: First, social mechanisms make explicit the causal paths that produce outcomes of IS phenomena, and thus, lead to a more constructive explanatory theory. Second, being empirically driven, the tracing of social mechanisms is likely to produce new insights beyond those implied by the theories that frame existing research, thus contributing to a more complete, multi-causal explanation (Avgerou, 2013).
4.3 Coleman´s Framework – The Micro-Macro Conundrum
Melville (2010) identified three distinct classes of sustainability phenomena: 1) how cognitive states about sustainability (beliefs, opportunities, motives etc.) emerge; 2) actions of organizations and individuals on the topic of sustainability practices and processes; and 3) environmental and financial performance outcomes. These classes encompass the act of generating sustainable value from data, ranging from factors that influence individual cognitive states to the individual actions that collectively contribute to the generation of value. Viewed as a whole, the three classes of phenomena comprise micro and macro constructs (Melville, 2010). Coleman’s theoretical foundation underscores the mediating role of individuals in linking macro-level structural variables to social macro-level impacts as shown in Figure 5.
Figure 5: Meta-theory based on Coleman´s Framework, adapted from Hedström and Ylikoski (2010)
Following Coleman (1986, 1990), we deduce that the goal of the social system is to maximize utility or what we have defined as sustainable value. We use Coleman’s framework as a meta-theory to explain the micro to macro level relationship between use of open data and the generation of sustainable value. Coleman (1990) proposes that a theory which can generate macro-level empirical generalizations as specific propositions may be thought of as “a theory of individual action, together with a theory of how these actions combine, under specific rules, to produce systemic behavior”
(Coleman, 1990, p. 20).
Three types of relations are included: 1) macro-level variables that explain the societal context which affects individual capacity through situational macro-micro mechanisms (by influencing the motivation, opportunity and ability of individuals); 2) individual capacity affects individual action through action-formation micro-micro mechanisms;
and 3) collective individual action affects macro-level constructs such as sustainable value through transformational micro-macro mechanisms (Hedström & Swedberg, 1998). In Paper V, I propose that the opportunity for value generation through use of data is created through the act of making data open. By constructing a robust regulatory data protection and privacy framework that produce trust and stability in information markets, we gain the means to influence the extrinsic motivation of stakeholders to use data. Furthermore, governments can motivate individuals if they demonstrating strong IT leadership, for instance by organizing events like hackathons, by promoting use of data or by being forerunners in utilization of digital solutions.
Finally, individual ability and expertise can be influenced by the supply of relevant skills and tools in the open data ecosystem.