Article Preview
Top1. Introduction
Conceptual modeling is an activity performed during information systems (IS) development and maintenance to represent certain semantics of real-world domains (Weber, 2003). It is motivated by a single goal: providing an accurate and complete representation of someone’s or some group’s understanding of a domain (Bodart et al., 2001). It has become a conventional research theme in the IS discipline (Bera et al., 2014; Siau and Rossi, 2011) and industry (Fettke, 2009). Owing to the increasing importance of databases in the business domain and the unique characteristics of modeling that require both theoretical and empirical examination, it has captured the interest of many researchers. Although database management inside and outside an organization is becoming critical, the grammars and methods used for conceptual modeling vary depending on the user or project (Topi and Ramesh, 2002), and there are some controversial issues such as construct overload and construct excess that remain to be resolved (Wand and Weber, 2002). In addition, previous research on conceptual modeling seems to have focused on the individual effects of syntax (i.e., constructs in the grammar), semantics (i.e., meaning of the construct), and pragmatics (i.e., context or domain in which a grammar is used) independently (Burton-Jones et al., 2009).
In the context of grammars and methods, construct overload, especially the part-whole relationship, remains a problem because alternative representations in conceptual models have been proposed. Based on the theory of ontological clarity provided by Wand and Weber (1993), Shanks et al. (2008)1 concluded that the ontologically clear model allows users to better understand a domain, which indicates that a distinction needs to be made between an entity and a relationship. However, Allen and March (2012) came to the opposite conclusion of Shanks et al. (2008) and argued that no distinction is needed between an entity and a relationship. The representation of the part-whole relation as a relationship or an entity remains an issue to be resolved, and these conflicting viewpoints were published in the same issue of MIS Quarterly in September 2012. To reconcile the inconsistent results of the conceptual modeling research described above, this study sought a more convincing and clear experiment to be performed.
In light of previous research on conceptual modeling, most studies have focused on the semantics (i.e., meaning) of models to present the best way to convey meaning clearly and completely (Bera et al., 2014). However, this can lead to problems because it is hard to exclude user domain knowledge (i.e., pragmatic factor) when a user interprets the conceptual model in a different domain. Some research has examined the processing aspects of domain knowledge, but far fewer studies have emphasized data aspects such as conceptual modeling (Vessey, 2006) and the interaction effect between syntax, semantics, and pragmatics (Bera et al., 2014; Burton-Jones et al., 2009). Even if they did, most studies were performed within familiar domains, such as the project planning domain for industry workers (Skanks et al., 2008) and the business domain for university students majoring in management information systems (MIS) (Allen and March, 2012). In such cases, it is difficult to measure the exact effect of a domain because of the lack of comparisons where the model domain is unfamiliar to the user.
This study examined research that combined these two issues: construct overload and the lack of interaction effect between syntax, semantics and pragmatics. Specifically, the motivation of this study was to answer the following research question: Does domain familiarity affect the user’s performance in construct overload?