Evolution of a process model

The notion of process model in information systems research was first instigated by Newman and Robey in 1992. In their research, they are looking at the relationship between the user and the analyst during an information system (IS) development project. They have made an analytical comparison between the factor- or variance- based research and also process-based models in understanding IS development projects. They have concluded that the factor model and the process model are complementary but should be combined into a single model due to their differing forms (Newman and Robey, 1992). Mohr (1982) further deliberates on the incompatibility of combining the two models.

They further deliberate the benefits of process model and proceed from there. The nucleus of the model is its difference with the factor model that treats the process as unknown and unknowable (Newman and Robey, 1992). With process research, focus is placed on the sequence of events (Mohr, 1982) within the process. This approach enables a better understanding on the dynamics of social change and also provides an in-depth explanation of how and why results are achieved (Van de Ven and Huber, 1990; Mohr, 1982). According to Kling (Kling, 1987) and Markus and Robey (1988), this type of model provides a faithful account of actual experiences of what really happens especially during a IS development project.

Figure 3: Factor and process model of system development (adapted from Newman and Robey (1992) and Lyytinen and Newman (2008)

As the name suggests, the development of the model follows the process theory approach (Van de Ven and Poole, 1990). It starts with the notion of events derived from observation of incidents (Newman and Robey, 1992). In this model, there are two types of events: encounters and episodes. Encounters are at a specific point of time at the beginning and the end of an episode, whereby an episode is a set of events which travels across time and space (Newman and Robey, 1992; Robey and Newman, 1996).

To understand the nature of the IS change, this model further elaborates on the concept of punctuated equilibrium (Gersick, 1991). According to Gersick (1991), there are two levels of IS change which are described as first- and second-order change. First-order change occurs when the change is continuous and incremental over time or during periods of stable infrastructure with incremental adaptations (Gersick, 1991). Second- order change involves episodic punctuations or brief periods of revolutionary upheaval (Gersick, 1991). Independent Variables •Resources •User involvement •Technical quality •Top management support Dependent Variables •Use of system •Economic benefit •Better decision •User satisfaction Inferred processes of development (Closed-box) Factor model Process model Critical Event 2 Critical Event 1 Critical Event N Critical Event N+1 Time Environment Antecedent conditions Outcomes

Figure 4: Mapping events in a social process (Newman and Robey, 1992)

Although it is not specifically stated in the article, since this model tries to understand the relationship between users and analyst, it indirectly elaborates the different process of work between the users and the analyst and shows how one‟s activities affect others within the project (Figure 5).

Figure 5: Interpretation of Newman and Robey (1992) user and analyst relationship

Viewing the model in a more general form (Figure 4), also involves an understanding of the antecedent condition or the history of the IS development project and how it will affect outcomes in relation to the users and the analyst sequence of work. Antecedent conditions usually encompass prior projects outcomes and their relative nature towards the new existing projects (Newman and Robey, 1992). In other words, users and analysts who are involved in the current project transfer their experience and expectations from prior projects which in turn affect how they perceive their current project.

Event 1 Event 2 Event n Event n +1

Event 1 Event 2 Event n Event n +1 User‟s activity Analyst‟s activity Antecedent conditions No significant pattern of development Analyst-led development User-led development Joint development Time Acceptance Equivocati on Rejection en 1 en2 en3 en4 en5 en6 ep6 (outcome) ep1 ep2 ep3 ep4 ep5

Legend: encounter (en) episode (ep)

Any episodes of user-led, analyst-led or joint development will be considered as first order change where the development is incremental and ongoing. Changes from user- led to analyst-led or even joint development will still be considered as first order change if it does not create punctuation or upheaval, e.g. resistance by either users or analyst towards the project. This brief period of punctuation will result in either the project continuing incrementally or even equivocation and further possible abandonment.

In relation to the outcomes, this model restrains itself from viewing it through success or failure dichotomy due to the inconsistency of the term itself. Rather, the outcome is conceptualised as state of relationships either user-led, analyst-led or joint development (Newman and Robey, 1992). When state of equivocation occurs, the future project will be surrounded by high level of risk or uncertainty due to lack of commitment from project team (Newman and Robey, 1992).

It is observed that there is no specific method or approach identified as a means of events identification rather than through observation of incidents. It is up to researchers‟ own perceptions and interpretive judgements of what an event should be (Newman and Robey, 1992; Robey and Newman, 1996). The authors added that since this model represents a simplification of reality, the identification of these events is a critical process.

Robey and Newman (1996) further elaborate on the model by focusing on the capability of the project trajectory built from the model to support theoretical interpretations. Based on Kling‟s (1980), theoretical perspectives on social analysis of computing, Robey and Newman (1996) discuss the findings from the model. There are two main perspectives introduced by Kling (1980); these are the rational perspective and segmented institutionalist perspective which has no superiority over the other. The rational perspective was further divided into three main approaches. In the rational approach, technologies are seen as tools to achieve goals, and the failure to attain these goals is due to technology inefficiencies rather the users. The structural approach includes the social context of the technology and evaluates its fit within the environment. The human relations perspective incorporates the social and technical

criteria into the analysis, where goals are jointly achieved and optimised (Robey and Newman, 1996).

The segmented institutionalist perspective covers the integrationist and organisational politics approach. The integrationist approach looks at the symbolic meaning of technology. Within this approach, system development is considered as a social process where social meanings are created and preserved through the interaction between sub- cultures. In organisational politics, the identification of the stakeholders and their interests are crucial, therefore focus is given to the conflicts between subgroups. In this approach of resolution of conflict, only an unstable compromise is achieved (Robey and Newman, 1996).

The application of these perspectives within the process model extends its applicability in making sense of the complex social process of IS development (Robey and Newman, 1996). According to Sabherwal and Robey (1995), the process model suggested by Robey and Newman (1996) “enables preservation of detailed information about specific events and their temporal order.” This observation is made in comparison to the stage model which is argued restricts the details of the unfolding events replacing them with prescribed events (Sabherwal and Robey, 1995).

The application of the social process model illuminates its benefits. The social process model was methodologically used as a „lens‟ for understanding the relationship structure between project team members (Holmstrom and Henfridsson, 2006; Holmberg et al., 2008) and assisting in viewing large datasets by capturing it in a project trajectory (Heiskanen et al., 2008; Holmberg et al., 2008).

3.3 The punctuated socio-technical information systems change (PSIC) model