統包工程特性及執行效益之研究

Abstract

日新月異的科技與公共工程規模日漸龐大，傳統工程採購方式己無法滿足現今業主與工程專案之需求，統包模式因應而生，根據工程會統計統包工程標案91年約93億元(104件)、92年約407億元(603件)、93年約557億元(541件)，合計約1047億元。 過去有相當多有關統包各項議題之研究報告，不乏探討其效益之研究，但多屬少量統包工程案例或是單一同種性工程來進行分析評估，更多研究提到統包採購之優點、適用性等，其內容缺乏客觀之依據證明，同時缺乏實際執行單位對統包之觀點做探討，因造成決策者或執行單位在評估工程是否適用統包時無具體資訊可參考，或者執行單位執行統包時造成與承包商之誤解。 故本研究首先進行工程案例資料收集，利用91~93年1248件的統包實際案例作分析。本研究分三個主軸，第一去瞭解執行單位對統包之觀點並討論之；第二利用統計分析來評估其效益，主要的方法為卡方檢定與G平方法，卡方獨立性檢定用來檢定兩因子間是否有關係性存在，檢定因子項目為採購特性、廠商特色、經費、進度與品質五個部份，G平方為交叉分析三者之間關聯性強度，並結合χ平方統計法輔助說明三者之間關係性，主要是依據部份的卡方檢定項目，加入經費增減或工期變化之第三因子；第三將工程會隨機抽樣所發函的問卷(發出約230件，回收107件)與本研究額外收集得到統包之作業範圍，94件作業範圍案例與第二主軸內容進行卡方獨立性檢定與對G平方法之統計方法評估其效益。然第二、三主軸因部份資料有限，並受卡方統計方法限制，則改使用精確檢定或百分比表示現象。其初步分析結果發現小型工程案件量大，因再針對預算1000萬以下之小型工程做更進一步之研究。 用現有數據資料，運用迴歸分析預測成本、工期和品質，結果卻因R平方值低，判斷他們之間無線性關係存在，但不表示他們無相關。

Because of the improving technology and the growing scale of Public Construction, the traditional design/bid/build delivery approach is not sufficient for the current Construction Project. The Turkey project delivery method can supplement this insufficiency. On the basis of the Public Construction commission, in the case of Turkey Project of procurement, there are 9.3 billion NT dollar projects in 2002, 40.7 billion NT dollar projects in 2003, and 55.7 billion NT dollar projects in 2004. Most existing researchs of the Turkey project explore the results and advantages of the Turkey project on the basis of few cases or of homogeneous constructions. Because most of them lack objective evidence and opinions from operational entitles, directors or operational entities don’t have sufficient information to evaluate their projects. Moreover, these researches provoked misunderstanding between operational entities and contractors. Therefore, we collect 1248 Turkey cases from 2002 to 2004 and dissect them. This research is dividend for three topics. First, we discuss/deliberate the point of views from operational entities. Second, we evaluate the results by stastistical analysis — the chi-square test and the Exact Test. The chi-square test is used to examine whether correlation exists between two factors. The categories of factors are identities of procurement, characteristics of contractors, the cost, the schedule and the quality. The exact test is used to analyze the intensity of relationship among three factors and combined with χ2 –test stastistic to explain this relationship. Mostly, exact test is based on the chi-square test and add the third factor which is either the change of budget or the change schedule. Third, we use chi-square test and exact test to evaluate the results from the content of the 2nd topic, questionnaires (random samples from the Public Construction commission) and our additional collections of Turkey projects. However, because of lacking partial information of the 2nd and 3rd topics and the limitation of chi-squrare test, we show the phenomena by using percentage and exact test.