軟體流程改善是企業提升競爭力一項重要的活動，且這項活動所需耗費的成本及時間更是不容小覷，所以組織必須要有足夠的決心來從事軟體流程改善，因此瞭解CMMI導入對組織所帶來的效益，將是一個很重要的信心憑藉。從國內外文獻探討中皆肯定CMMI的導入確實能為組織在多方面帶來正面的效益，然而卻缺乏以CMMI成熟度等級來探討各效益構面(成本、時程、品質、客戶滿意度、投資報酬率及員工)之間是否有顯著的差異。

本研究主要探討ML2(低成熟度等級)、ML3(中成熟度等級)及ML4加ML5(高成熟度等級)這三個成熟度等級與ML3二種不同導入模式(直接導入ML2和ML3及ML2認證通過後再導入ML3)對於CMMI流程改善各個效益構面的影響。本研究分析中華民國資訊軟體協會所收集到的國內46家CMMI導入廠商之68筆專案量化績效指標資料，結果發現：在不同成熟度等級中僅有「品質構面」的效益有顯著的差異(高成熟度等級顯著優於中、低成熟度等級)。在ML3導入模式不同中，顯示循序漸進的導入模式(ML2認證通過後再導入ML3)是有助於品質效益的提升，此兩種不同的導入模式對各效益構面之影響不盡相同，廠商可以依據自身需求而決定CMMI ML3的導入模式。綜合本研究結果的分析，可以做為國內組織欲導入CMMI流程改善決策之參考。

Software process improvement (SPI) is particularly important for enterprises in enhancing their competition in the business market. However, enterprises need to spend much cost and time in the deployment of SPI, such as capability Maturity Model integration (CMMI). Therefore, enterprises must have plentiful determination to ensure the success of their SPI investment. It is also important to understanding the benefits of CMMI adoption, as it brings enterprises confidence to decide whether to conduct SPI activity. The reports in the literatures have affirmed that CMMI adoption can bring the positive benefits in various degrees for enterprises. However, so far there is no report on the effects of CMMI maturity levels to the various categories of CMMI performance, including cost, schedule, quality, customer satisfaction, return on investment, and employee.

The main objective of this research is to investigate the effects of the maturity levels: ML2 (Low Maturity), ML3 (Middle Maturity), and ML4 plus ML5 (High Maturity), and the two adoption methods of ML3 (ML1 to ML2 plus ML3 and ML1 to ML2, and then to ML3) on the performance of CMMI adoption. The results of analyzing 68 projects from 46 domestic CMMI deployment companies by the Information Service Industry Association of R. O. C. showed that only the “Quality Dimension” was significantly affected by three different CMMI maturity levels (High Maturity Level was significantly superior to that Low and Middle Maturity levels), and the other performance dimensions were not significantly affected by different maturity levels. As for two different adoption methods for CMMI ML3, the results showed that the quality dimesion of CMMI performance by gradually improveing process maturity to ML3 (ML1 to ML2, and then to ML3) was better than the adoption of ML1 to ML2 plus ML3. The results also showed that the two different adoption methods had different degrees of influence in CMMI performance dimensions, so companies need to decide their adoption method of CMMI ML3 based on their enterprise demands.

中文部份
[1] 台灣軟體工程學會，「CMMI 系列研究暨產業應用研究計畫」，經濟部工業局，民國九十四年。
[2] 吳美芳，「軟體程序成熟度、使用者參與度與軟體品質之相關研究」，碩士論文，國立成功大學，民國九十一年。
[3] 吳宗穆，「軟體流程改善量化績效指標制訂之研究」，碩士論文，國立臺灣科技大學，民國九十六年。
[4] 吳純慧，「軟體程序成熟度與專案績效」，碩士論文，國立中正大學，民國八十九年。
[5] 李金泉，「精通SPSS統計分析實務與應用」，全華圖書股份有限公司，民國九十六年。
[6] 林佑達，「政府資訊委外廠商服務品質之研究」，碩士論文，國立台北大學，民國九十年。
[7] 郭怡岑，「組織特性與投入資源對CMMI流程改善效益影響之研究」，碩士論文，國立臺灣科技大學，民國九十五年。
[8] 陳政雄，「軟體能力成熟度整合模式下的專案管理流程領域對軟體品質成本影響之研究」，碩士論文，國立成功大學，民國九十三年。
[9] 陳燿昇，「組織特性、資訊化成熟度、資訊部門特性與企業軟體程序成熟度關係的探討」，碩士論文，國立中正大學，民國八十七年。
[10] 國立臺灣科技大學，「國內組織CMMI 流程改善績效資料基準比較儲藏所」，經濟部工業局，民國九十六年。
[11] 馮王貞美、魏秋建、林育德，「品質達成率應用在專案品質的評估與控制」，Journal of the Chinese Institute of Industrial Engineers，第22期，No. 2, pp.155-170，民國九十六年。
[12] 楊茲檢，「導入CMMI對軟體專案管理績效之影響」，碩士論文，輔仁大學，民國九十四年。
[13] 經濟部工業局提升資訊軟體品質(CMMI)計畫網站，http://www.cmmi-taiwan.org.tw/。
[14] 謝承鑫，「專案團隊、軟體能力成熟度與專案績效之探討」，碩士論文，中正大學，民國九十年。
[15] 謝明峰，「運用能力成熟度模式改善軟體專案管理品質績效之探討-以國內某電信公司為例」，碩士論文，政治大學，民國九十三年。
[16] 謝菊蓁，「CMMI及知識管理導入與軟體專案組織成效之關係」，碩士論文，輔仁大學，民國九十三年。
[17] 藍昭雄，「軟體組織導入CMMI對組織績效影響之研究-企業程序再造之調節觀點」，碩士論文，輔仁大學，民國九十六年。
[18] 蕭怡祺，「使用者因素與成熟度層級對組織績效的影響：軟體能力成熟度為例」，碩士論文，清華大學，民國九十一年。
英文部份
[19] Ashrafi. N. (2003). “The impact of software process improvement on quality: in theory and practice.” Information and anagement 40(7), 677-690.
[20] Brodman, J.G. & Johnson, D.L. (1995). “Return on investment from software process improvement measured by U.S. industry.” Crosstalk, 9(4), 23-29.
[21] Damian, D., Zowghi, D., Vaidyanathasamy, L. & Pal, Y. (2002). “An industrial experience in process improvement: An early assessment at the Australian Center for Unisys software.” International Symposium on Empirical Software Engineering (ISESE’02), 111-123.
[22] Dennis R. Goldenson, “Software Tech News,” Vol. 10, No. 1, March 2007.
[23] Diane L. Gibson, Dennis R. Goldenson, Keith Kost, “Performance Results of CMMI-Based Process Improvement”, CMU/SEI-2006-TR-004.
[24] Dion, R. (1993). “Process improvement and the corporate balance sheet.” IEEE Software 10(4), 28-35.
[25] Emam, K & Briand, L. (1997). “Costs and benefits of software process improvement.” Institute for Experimental Software Engineering (IESE) Report No 047.97/E.
[26] George, E.S. (1996). “Measurements for Managing Software Maintenance.” IEEE Press 4(8), 152-161.
[27] Goldenson, D.R. & Gibson, D.L. (2003). “Demonsrating the impact and benefits of CMMI: an update and preliminary results.” Special Report, CMU/SEI-2003-SR-009.
[28] Goodman, L.A. & Kruskal, W. (1954). “Measures of Association for Cross-Classification.” Journal of the American Statistical Association, 49, 732-764.
[29] Harter, D.E., Krishnan, M.S., Slaughter, S.A. (2000). “Effects of process maturity on quality, cycle time, and effort in software product development.” Management Science 46(4), 451-466.
[30] Herbsleb J, Carlton A, Rozum J, Siegel J, Zubrow D..(1994). “Benefits of CMM-based software process improvement: initial returns.” Technical Report, CMU/SEI-94-TR-013.
[31] Hyde, K. & Wilson, D. (2004). “Intangible benefits of CMM-based software process improvement.” Software Process Improvement and Practice 9(4), 217-228.
[32] Jiang, J.J., Klein, G. Hwang, H.G. Huang, J & Hung, S.Y.(2004). “An exploration of the relationship between software development process maturity and project performance.” Information & Management 41(3), 279-288.
[33] Li-Wei Chen, Sun-Jen Huang (2009). “Accuracy and Efficiency Comparisons of Single- and Multi-cycled Software Classification Models,” Information and Software Technology , Vol. 51, pp. 173-181.
[34] McGibbon, T. (1999). “A business case for software process improvement revised.” DoD Data Analysis Center for Software(DACS).
[35] Mehner, T., Messer, T., Paul, P., Paulisch, F, Schless, P. & Volker, A. (1998). “Siemens process assessment and improvement approaches: experiences and benefits.” 22nd International Conference on Software Engineering (ICSE), 186-195.
[36] O’Hara, F. (2000). “European experiences with software process improvement.” Proceedings of the 22nd international conference on Software engineering, 635-640.
[37] Pitterman, B. (2000). “Telcordia Technologies; the journey to high maturity.” IEEE Software 17(4), 89-96.
[38] Ryu, H. and Hong, G.Y. et al (2004). “A Research on Quality Assessment Technique for Ubiquitous Software and Middleware.” For the Final Deliverable (Year1) to TTA.
[39] Sandra, A.S. and Donald, E.H. et al (1998). “Evaluating the Cost of Software Quality.” Communications of the ACM 41(8), 67-73.
[40] Wohlwend, H. & Rosenbaum, R. (1993). Software improvement in an international company. 15th international conference on Software Engineering, 212-220.
[41] Yamamura, G. (1999). “Process improvement satisfies employees.” IEEE Software 16(5), 83-85.