An Empirical Examination of Dynamic Quality-Based Learning Models
利用一家财富500强公司12家工厂的缺陷率和质量成本数据，首次直接检验两类动态质量学习模型的预测，发现同时考虑主动投资和自主学习的模型更优，并揭示了预防活动与过去不合格支出对缺陷率的影响。
Using detailed data on defect rates and quality costs from twelve plants of a Fortune 500 company, we provide the first direct tests of predictions arising from two sets of dynamic quality-based learning models. We find greater support for quality-based learning models that assume learning is a function of both proactive investments in quality improvement and autonomous learning-by-doing, than for models that assume learning is a function of reactive investments in quality improvement alone. We then extend these two sets of models to examine the impact of individual prevention activities and past nonconformance expenditures on defect rates. We find that benefits from different types of prevention expenditures vary, and that past nonconformance expenditures provide learning opportunities that allow the organization to more efficiently cope with future failures, thereby reducing subsequent nonconformance costs. These important implications are absent in current quality-based learning models, providing an opportunity for future theoretical development.