Modular system design offers several benefits such as flexibility, upgrade, extension and inter-operability across multiple vendors. The level of system modularity is strongly driven by design inter-dependencies. Several academic models have been proposed for assessing these inter-dependencies. However, there is lack of consistency across these models in terms of evaluation metric and scope of application. In this paper, by measuring system pleiotropy, we compare three existing models to investigate essential elements of reliable assessment methodology. The three models are applied on a computer mouse architecture. It is identified that by including module composition, inter-module interface complexity and functional dependencies, a credible measure of system design dependencies can be obtained.