Perceived Control of Software Developers and Its Impact on the Successful Diffusion of Information Technology

Gina Green (Baylor University)
Alan R. Hevner (University of South Florida)

Special Report
CMU/SEI-98-SR-013

Why are beneficial software engineering practices not being used effectively in the development of software systems? This question has intrigued researchers in software engineering for many years. Billions of dollars per year are spent, and a large proportion wasted, on building and maintaining software systems that are either never completed or, if completed, are of poor quality. This state of software development has led to the introduction of innovative tools and techniques to support the software development process. Initial evidence from use of these tools and techniques shows significant improvements in development productivity and software quality. However, many of these potentially beneficial tools and techniques have not been widely adopted or diffused. This research seeks to examine the reason for why this is so: What factors explain the successful diffusion of new software development techniques into practice?

Software development techniques are viewed as a subset of the broader category of information technology (IT). A research framework is developed that explains the complex relationship between developer involvement in the IT adoption process, characteristics of the environment into which the IT is introduced, and IT diffusion success. The framework posits that the effects of developer involvement and IT diffusion environment characteristics on IT diffusion success are mediated by (1) the developer’s perceived control over his work when using the IT, (2) the developer’s perceptions of the IT, and (3) the developer’s perceptions of the impacts of IT use. Using the Personal Software Process^SM (PSP^SM) approach as an example of an innovative IT, we develop a survey instrument for the collection of study data, which are analyzed using path analysis. The survey was distributed to software developers who have used PSP on software development projects in industry. Results find support for mediating effects of Perceived Control on the relationships between Developer Involvement, Training, Voluntariness, Champion Support, and IT Diffusion Success. Results also find direct impacts of Developer Involvement, Training, and Voluntariness on IT Diffusion Success.

This research contributes to research by providing a model that integrates knowledge from the research fields of information systems (IS), diffusion of innovations, software engineering, and social psychology to better understand why user involvement and diffusion environment impact diffusion success. Further, it provides a measure for perceived control in an IS context. This research contributes to practice by stressing the importance of software developer perceptions in determining IT diffusion success. The research underscores the importance of creating a perception of greater control over the software development process in order to positively influence diffusion success. Guidance is provided to software development practice by stressing the importance of software developer involvement, training, voluntariness, and champion support in the introduction of new IT. These variables can be influenced by management to enhance perceived control and, therefore, better ensure successful diffusion of new software development practices in organizations.

Editor’s Note: This SEI Special Report also includes information from a presentation delivered by the authors at the Software Engineering Institute. The information, which appears in Appendix B, is material collected by the authors from their questionnaire and specifically concerns the Personal Software Process^SM (PSP^SM).