Several modularity problems in software designs are related to the inadequate modularization of key broadly-scoped concerns, such as exception handling, distribution, and persistence. However, most of the current quantitative assessment aproaches are not sensitive to concerns that drive the design, thereby leading to a number of shortcomings in the modularity evaluation process. Therefore, there is a need for measurement approaches that support a more effective identification of modularity anomalies related to crosscutting concerns. Also, this necessity becomes more apparent in an age that a number of different fomrs of design decompositions, such as aspect-oriented software development, are emerging. In this context, this thesis aims at investigating a novel approach for quantittative modularity assessment of software design by promoting the concept of concern as a measurement abstraction. Our concern-driven measurement approach encompasses a set of mechanisms for assessing software modularity from architectural to detailed design. The prposed concern-sensitive approach includes:(i) a suite of architectural metrics, (ii) a suite of detailed design metrics, (iii) a suite of design heuristic rules for supporting the interpretation of metrics in meaningful ways, and (iv) a tool, called COMET, that supports bots concern-drive notation and measurement of architectual designs. We evaluated the usefulness of our concern-oriented measurement techinique in a series of emprical studies, comparing the modularity of conventional and aspect-oriented software design.