資訊科技的發展日新月異，企業的經營依賴科技作為競爭力提昇的手段是日益明顯的趨勢。能善用資訊科技者，多能將之產生競爭力優勢。是以軟體品質的重要性從傳統的效率性構面，質變為競爭策略的構面。亦即，將資訊科技視為企業競爭武器。 能力成熟度模式 (Capability Maturity Model) 主要是針對資訊軟體生產流程發展出，作為軟體流程改善的模式。是近年最常被提及的軟體流程管理模式。在模式與實作之間，欲將模式的要求具體化，實有一段執行上的挑戰。 本研究主要是以能力成熟度模式第二階段之軟體專案管理 (Software Project Management) 的品質確保 (Quality Assurance) 流程為主軸，來研究其模式如何將之具體化，並以國內某電信公司的專案管理流程改善為個案實作之。展示該企業如何運用其中的精神及要求重點，實際導入於軟體專案管理品質確保流程的建立。並就實際導入經驗及檢討修正，提出對能力成熟度模式強化的建議。

為提升國內軟體專案生產力和品質，建立度量與分析的機制，本研究參考能力成熟度整合模式(Capability Maturity Model Integration, CMMI)之度量與分析流程領域作為指引，結合GQM (Goal-Question-Metric)及PSM (Practical Software & Systems Measurement)的度量方法論，發展一套提問式驅動度量與分析流程規劃法，稱為QMAP (Questions driven- Measurement & Analysis Process Planning)—運用5W1H (Why, Who, When, What, Where & How) 提問方式，有系統地依據CMMI度量與分析流程的要求，找出軟體專案的成功因素，逐步建立組織內部的度量與分析流程的基準。首先，以某公司之個案情境為例進行背景分析、目標分析和流程分析，接著以使用分析—使用案例及其描述之資訊需求，建議整合運用PSM Insight, MS Project及IBM Rational ClearQuest等工具來支援軟體專案度量與分析流程之自動化。經工具系統之實作展示，再以自我評鑑來檢視QMAP於CMMI度量與分析流程之目標及執行方法的符合程度，而部分未達成的項目則作為未來研究之方向。 / This research proposes a method called QMAP (Questions driven- Measurement and Analysis Process Planning) in order to meet goals and practices in Capability Maturity Model Integration (CMMI). QMAP is about 5W1H questions which begin with why, who, when, what, where and how. Before answering the 5W1H questions, we start with background analysis, goal analysis, and process analysis based on a sample company’s management context. Use Cases are then described as the formal information requirements of the MA process support system. In the implementation, we suggest to integrate some tools, such as MS Project, PSM Insight, and IBM Rational ClearQuest, to facilitate the measurement data collection and analysis activities during the software project development. After illustrating the implementation prototype, a checklist for self-appraisal is presented to identify the compliances of CMMI goals and practices, and the incomplete parts for future research extension.

軟體專案承諾升級現象發生時，可能會造成更多資源成本的投入，若投入更多的資源，而專案依然失敗，則會造成更多的浪費，是企業界最不希望看到的情形。本研究利用推敲可能性理論探討此現象，了解不同的說服方式如何影響受測者的決策過程。根據遇到專案的狀況，依照訊息訴求分成理性訴求與感性訴求的敘述方式，加入框架效應 (正向／負向) 以及訊息強度 (低／中／高) 兩變數，以自我責任作為調節變項進行探討，研究在不同情況敘述下，決策制定者接收專案訊息描述後其決策過程中的推敲可能性，以及推敲可能性與決策之間的關係。 研究發現如下： 1. 單一效果影響下，訊息強度中比起其他兩個強度較容易引起訊息接收者運用中央路徑思考。 2. 訊息訴求、訊息框架以及訊息強度會交互影響受測者推敲可能性，在正向框架下，訊息訴求與訊息強度對訊息接收者運用中央路徑做決策有顯著的交互作用。 3. 在訊息強度中與強的情況下訊息訴求與訊息框架對訊息接收者運用中央路徑做決策有顯著的交互作用。 4. 在感性訴求、正向框架與訊息強度強的訊息描述下，訊息接收者運用中央路徑思考時，較不容易做出承諾升級的決定。 5. 在自我責任調節下，則是感性訴求、負向框架與訊息強度弱的訊息描述，會引起訊息接收者運用中央路徑思考時，較不容易做出專案繼續的決策。 / Escalation of commitment is common in software project development. There are a few theories that have been used to explain this behavior, including the framing effect and self-responsibility. This study investigates the issue from the dual-path elaboration likelihood model (ELM) to examine how different persuasion routes may play roles in the decision process. An experiment was designed to study the effect of different descriptions of project status that may lead to different decision routes (central versus peripheral routes). The experiment design includes message appealing (rational vs emotional appealing), message strength (strong, medium and weak), and framing (positive vs. negative) as main variables and the responsibility as a moderator. The subject was asked to decide whether s/he would continue the project under a given scenario. Our results includes the following: 1. Message appealing, message framing, and message strength have significant interaction effect on the subject’s decision routes; 2. In positive framing, message appealing and strength has significant interaction effect on the use of the central route; 3. When message strength is medium or strong, message appealing and framing has significant interaction effect on the use of the central route; 4. Regarding to decision escalation, the likelihood of escalation is lower when the decision route is central (thinking) under the emotional appealing, positive framing, and strong message; 5. The likelihood of escalation is lower when the decision route is central under the emotional appealing, negative framing, and weak message description.