Towards a Theory of Web Effort Estimation

Reliable Web effort estimation is one of the cornerstones of good Web project management. Hence the need to fully understand which factors affect, and how they affect, a project’s outcome, and their causal relationships. The aim of this paper is to propose a mechanism to obtain an empirical generalization of four different causal process form type of theories for Web effort estimation. Each one of these theories was constructed using a hybrid approach to theory building, based on existing knowledge elicited from several domain experts, data on past completed Web projects, and a technique that enables the modeling of causal relationships and their uncertainty. The aggregation methodology used to combine these theories was the same one we employed previously with causal maps, which extends previous work by adding a mapping scheme to handle complex domains (e.g. effort estimation), and in addition also uses an aggregation process that preserves all the causal relations from the original theories. The resultant empirical generalization contains 70 concepts (variables), and highlights several patterns amongst the four theories used as input, thus providing a starting point to a more general theory of Web effort estimation.