Abstract
The traditional method for construct development in information systems (IS) is based on classical test theory (CTT). While the Rasch approach is more theoretically sound than CTT, it has not been widely adopted in the IS literature. This is probably because CTT and Rash often produce similar results but Rash it is more difficult to perform. However, test equating, an important aspect of Rasch analysis, provides IS researchers with an effective way to develop complex IS constructs such as those that are formative in nature. The current wisdom is that formative constructs should be determined as part of a nomological network (see Petter, Straub and Rai, 2007), but this approach has been shown to be problematic (Edwards, 2010). The approach taken in this study is to treat formative constructs as second order reflective with composites based on the component dimensions (Treiblmaier, Bentler, and Mair, 2010). These notions are illustrated by teasing out the components of the formative construct Interactive Project Control (IPC) in software development projects using a Rasch equating approach.
Recommended Citation
Fraser, Steven, "A Rasch Approach to Empirical Field Research in Information Systems" (2012). AMCIS 2012 Proceedings. 7.
https://aisel.aisnet.org/amcis2012/proceedings/ResearchMethods/7
A Rasch Approach to Empirical Field Research in Information Systems
The traditional method for construct development in information systems (IS) is based on classical test theory (CTT). While the Rasch approach is more theoretically sound than CTT, it has not been widely adopted in the IS literature. This is probably because CTT and Rash often produce similar results but Rash it is more difficult to perform. However, test equating, an important aspect of Rasch analysis, provides IS researchers with an effective way to develop complex IS constructs such as those that are formative in nature. The current wisdom is that formative constructs should be determined as part of a nomological network (see Petter, Straub and Rai, 2007), but this approach has been shown to be problematic (Edwards, 2010). The approach taken in this study is to treat formative constructs as second order reflective with composites based on the component dimensions (Treiblmaier, Bentler, and Mair, 2010). These notions are illustrated by teasing out the components of the formative construct Interactive Project Control (IPC) in software development projects using a Rasch equating approach.