Two Structural Equation Modelling Approaches for Cloud Use in Software Development

Creative Commons License

Pişirir E., Chouseinoglou O., Sevgi C., Uçar E.

Young Business and Industrial Statisticians Workshop (y-BIS) 2019, İstanbul, Turkey, 25 - 28 September 2019, pp.20

  • Publication Type: Conference Paper / Summary Text
  • City: İstanbul
  • Country: Turkey
  • Page Numbers: pp.20
  • Hacettepe University Affiliated: Yes


Structural equation modelling (SEM) is a statistical analysis method that can be used to
calculate relationships between variables in complex models. SEM is a preferred method for complex social models because of its ability to calculate effects of normally unmeasurable or unobservable variables on each other. This is made possible by using measurable indicators to understand the effect of unmeasurable (latent) variables. There are two different general approaches to SEM analysis, covariance based (CB)-SEM and Partial Least Squares (PLS)-SEM methods. Technology adoption models are complex social models. They are used to estimate factors that affect users’ intention to adopt or use innovations over the previous alternatives. These factors might include personal reasons, environmental factors, or business-related depending on the context of the technology and population in the study.
This study is a technology adoption study specifically focusing on the adoption of cloud
computing services in software development projects. To model the use intention, a hybrid
conceptual model is developed with the inclusion of a novel variable structure (called Person-Organisation-Project, POP) to two existing theories in literature, namely Technology Acceptance Model (TAM) and Technology-Organisation-Environment (TOE). A questionnaire is designed and personally administered questionnaire sessions are conducted in 30 different software development organisations. 268 valid observations are collected for statistical analysis. Hypotheses based on the conceptual model are tested with CB-SEM and PLS-SEM methods. Thus, the suggested hybrid model is evaluated and also two SEM approaches are compared in the context of a cloud adoption in software development study. After modifications, the final model is reached. The novel hybrid technology adoption model is validated and several practical conclusions are drawn about the population and cloud use in software development.