Quality improvement in terms of lower costs, shorter development times and increased reliability are not only important to most organisations, but also demanded by the customers. To enable management to early identify problems, and subsequently to support planning and scheduling of development processes, methods for identifying fault–prone modules are desirable. This thesis demonstrates how software metrics can form the basis for reducing development costs by early identification, at the completion of design, of the most fault–prone software modules. Based on empirical data, i.e. design metrics and fault data, that have been collected from two successive releases of switching systems developed at Ericsson Telecom AB, models for predicting the most fault–prone modules were successfully developed. Apart from reporting the successful analysis, this thesis outlines a quality framework for evaluation of quality efforts, provides a guide for quantitative studies, introduces a new approach to evaluating the accuracy of prediction models, Alberg diagrams, suggests a strategy for how variables can be combined, and evaluates and improves strategies by replicating analyses suggested by others.