Assessing the Effectiveness of a Distributed Method for Code Inspection: A Controlled Experiment