Analysis of the Effects of Cost Drivers on Software Project Cost Estimation in Nigeria

ABSTRACT

This research work investigates the effect of cost drivers on software cost estimation. Several models exist that provide one or more mathematical algorithms that compute cost as a function of a number of variables or cost drivers. There are also several other tools and techniques available to assist in creating software project cost estimates, they are still very inaccurate. Several authors in software development have given varied suggestions for these inaccuracies and ways to overcome some of them. All classes of technique are challenged by the rapid pace of change in software technology.

The purpose of this research is to determine the extent to which software cost variables affect cost of software development. Primary data collected using the Likert scale was analyzed using Multiple Regressions and Discriminant functions statistical tools. The first hypothesis in this work was tested using multiple regression analysis tools, while the other two were tested using discriminant analysis tool.

Results from this work led to the conclusion that the cost drivers as a whole were contributors in predicting cost of software development in Nigeria. Their contributions to specific software application types indicated that some cost drivers were critical to the development of some software types, while some were not. The study also revealed the implication of not using cost estimation tools and stressed those variables that are critical to meeting project dead line.