目的探讨多变异分析(MVA)在工作中各环节全过程质量特征及方法。方法通过对变异的分析,可以发现哪种变异是显著的变异,从而为质量分析和控制变异指明了方向。结果通过设计MVA的试验,搜集数据,进行方差分析,可以清楚的了解程序质量的变异主要来源,从而增强标准的通用性和它们的相关性,强化了质量体系的持续改进意识。结论 MVA是一种十分重要的程序诊断方法,在工作的实践中,许多程序均可采用这一方法。 Objective To explore the quality characteristics and methods of multivariate analysis (MVA) in all aspects of work. Methods By analyzing variants, one can find out which variants are significant variations that point the way to quality analysis and control variation. Results Through the design of MVA experiments, data collection and analysis of variance were conducted to clearly understand the main sources of variation in program quality, thereby enhancing the universality of standards and their correlations and strengthening the awareness of continuous improvement of quality system. Conclusion MVA is a very important method of program diagnosis, which is used by many programs in practice.