第四章 研究結果與討論
第二節 建議
已達相對有效率境界的單位有 A1、A2、A4、A13 等 4 個消防分隊;尚未達 相對有效率境界的單位,其中已達技術效率而未達規模效率的有 A3、A5、A7、
A9、A14、A15、A16 等 7 個消防分隊,另未達技術效率且未達規模效率的有 A6、A8、A10、A11、A12、A17、A18、A19、A20 等 9 個消防分隊。按上述消 防分隊先對照其所屬大隊及地理環境:4 個相對有效率消防分隊,有 3 個屬於臺 東大隊並位於臺東市;而 9 個相對無效率消防分隊有 4 個屬於大武大隊並位於南 迴線四鄉鎮;再者,依據投入項之對偶變數表,無效率分隊投入於教育訓練量均 偏高,而應按其幅度減少,所以未達規模效率的原因,是環境因素或大隊訓練要 求不夠精實,值得決策者參考。
A1 分隊有效率程度遠高於其他 19 個單位,以致將其抽出後,效率值較常態 平均。經調查其原因,A2 分隊與 A1 分隊共用值班臺,而受報案或受理單位皆 為 A1 分隊,造成 A1 分隊產出量較高,相形之下效率較好,為正確統計分析,
連同前統計資料不夠符合實務(應將支援單位一併計入出勤次數),應檢討其統 計機制,俾利正確績效管理。
經敏感度分析發現,A1 單位較其他單位效率值之異常突出,且未達技術效 率單位,規模報酬均處於遞增狀態(IRS),可能為資源分配不當或其無效率單位 教育訓練不夠精實所造成。另對偶變數分析,所計算出未達技術效率消防分隊的 投入減少目標值,可作為決策者未來配置教育訓練相關資源時之參考。
在提出結論前,適逢「八仙塵爆」事件,除感慨事件造成 4 百餘人的傷亡外,
對照本研究,也突顯了消防單位在產出項之變異性,如這次的「八仙塵爆」事件、
去年的高雄氣爆事件,皆短時間、特定範圍造成消防單位救護量、救災量暴增,
其暴增的產出量(分子)變大,相形之下投入(分母)就相對顯得比較小,而數 據上會認為有效率,雖本研究統計時間未有類此重大事件,但卻值得爾後研究消 防相關績效認定之參考。
參考文獻
內政部消防署(2015)。102 年消防統計年報。取自 http://goo.gl/adbFxF
內政部消防署全球資訊網(2015)。統計資料。取自 http://goo.gl/Mof8Jh
內政部消防署法令查詢系統(2015)。內政部消防署辦理教育訓練課程作業要 點 (2010 修正)。取自 http://goo.gl/P5cD9D
中華民國統計資料網(2014)。臺東縣面積、人口數、人口密度。取自
http://statdb.dgbas.gov.tw/pxweb/Dialog/Saveshow.asp
丘昌泰(2000)。公共政策:基礎篇。台北市:巨流圖書公司。
台東縣消防局網站(2014)。關於台東縣消防局。取自
http://www.ttfd.gov.tw/about.php
林奕傑(2007)。應用資料包絡分析法建立消防分隊最事人力配置模式-以高 雄市政府消防局為例(未出版碩士論文)。國立高雄應用科技大學,高雄 市。
莊亮倫(2010)。應用資料包絡分析法與系統動態模擬於消防機關績效評估及 資源配置策略之研究(未出版碩士論文)。南華大學,嘉義縣。
張潤書(1990)。行政學。台北市:三民書局。
溫逸菁(2005)。我國消防機關緊急救護資源運用效率之研究(未出版碩士論
文)。中央警察大學,桃園縣。
孫遜(2004)。資料包絡分析法理論與應用。台北市:揚智文化。
維基百科(2014)。2014 年臺灣高雄氣爆事故。取自 https://goo.gl/Cx1K0D
黃鏡如、傅祖壇、黃美瑛(2010)。績效評估-效率與生產力之理論與應用。台 北市:新陸書局股份有限公司。
劉春初(2009)。Data Envelopment Analysis 資料包絡分析。取自
http://goo.gl/5CrqpF
趙哲彰(2012)。資料包絡分析法評估消防分隊績效之研究-以新北市板橋區、
中和區、新莊區、三重區為例(未出版碩士論文)。中國科技大學,台北 市。
陳國嘉(2001)。績效管理-全面提升利潤的評估制度。台北市:青泉出版社。
薄喬萍(2007)。績效評估之資料包絡分析法。台北市:五南文化事業。
羅凱文(2003)。臺灣地區各縣市消防機關效率評估之研究(未出版碩士論文)。 交通大學,新竹市。
Golany, B., & Roll, Y.(1989). An application procedure for DEA. Omega,
17(3), 237-250.
附錄一 授權同意書
致 Tim Coelli 教授取得使用 DEAP 2.1 軟體授權信
Dear Prof. Tim Coelli,
I am a student from the Department of Education, National Taitung University, Taiwan. I am researching in Data Envelopment Analysis (DEA). Study on Efficiency Analysis of Firefighting Education and Training in Taitung County, my study is conducting the efficiency analysis in manpower, equipment, and the amount of education and training at each fire branch. In the Efficiency and Productivity, written by Prof. Huang Jin Ru and C.A.Knox Lovell.
The book mentions the software: DEAP Version 2.1. Although the software can be bought in the internet (http://www.uq.edu.au/economics/cepa/software.htm), Prof. Huang hopes that students still have to send the letter of thanks to you. In fact, we admire and appreciate your selfless dedication and contribution to the public a lot in the DEA research.
However, we have some doubts in the DEAP Version 2.1. Here is my following questions:
1.Student’s Master thesis topic: “The Study in Taitung Fire Education and Training
Effectiveness Analysis". My study has been using your software in years and I would like to publish the paper in the future. I want to ask your permission.
2.In the result of the program: SUMMARY OF OUTPUT(INPUT). Weather the SLACKS is the equation of value S± in DEA?
At the end, once again I want to appreciate your effort in research and great contribution to the public. I wish you the wonderful day!
Best regards, Long Sam
Department of Education, National Taitung University (MA) E-Mail:[email protected]
Tim Coelli 教授授權使用 DEAP 2.1 軟體信
Comments below>>
---Original Message---
From: 18-6-189 [mailto:[email protected]]
Sent: 29 May 2015 03:22 To: Timothy Coelli
Subject: For Prof. Tim Coelli
Dear Prof. Tim Coelli,
I am a student from the Department of Education, National Taitung University, Taiwan. I am researching in Data Envelopment Analysis (DEA). Study on Efficiency Analysis of Firefighting Education and Training in Taitung County, my study is conducting the efficiency analysis in manpower, equipment, and the amount of education and training at each fire branch. In the Efficiency and Productivity, written by Prof. Huang Jin Ru and C.A.Knox Lovell.
The book mentions the software: DEAP Version 2.1. Although the software can be bought in the internet (http://www.uq.edu.au/economics/cepa/software.htm), Prof. Huang hopes that students still have to send the letter of thanks to you. In fact, we admire and appreciate your selfless dedication and contribution to the public a lot in the DEA research.
However, we have some doubts in the DEAP Version 2.1. Here is my following questions:
1.Student’s Master thesis topic: “The Study in Taitung Fire Education and Training
Effectiveness Analysis". My study has been using your software in years and I would like to publish the paper in the future. I want to ask your permission.
>> no problem
2.In the result of the program: SUMMARY OF OUTPUT(INPUT). Weather the SLACKS is the equation of value S± in DEA?
>> I do not know that symbol
At the end, once again I want to appreciate your effort in research and great contribution to the public. I wish you the wonderful day!
Best regards, Long Sam
Department of Education, National Taitung University
信件確認為研究者資訊:
舊名(現為:廖蕭龍笙)
附錄二 結果分析報告
Results from DEAP Version 2.1
Instruction file = 0517-ins.txt Data file = 0517-dta.txt
Input orientated DEA
Scale assumption: VRS
Slacks calculated using multi-stage method
19 0.270 0.703 0.383 irs 20 0.243 0.546 0.446 irs
mean 0.744 0.886 0.821
Note: crste = technical efficiency from CRS DEA vrste = technical efficiency from VRS DEA scale = scale efficiency = crste/vrste
Note also that all subsequent tables refer to VRS results
SUMMARY OF OUTPUT SLACKS:
7 7
19 0.103 0.709 0.007 0.181
SUMMARY OF OUTPUT TARGETS:
5 65.000 799.000 13.000 254.000
SUMMARY OF INPUT TARGETS:
18 6.384 24.460 2662.304 336.099 1028.624
output 2 1.000 0.000 0.000 1.000
variable original radial slack projected
Results for firm: 6
input 4 212.000 0.000 0.000 212.000
output 1 64.000 0.000 0.000 64.000
Technical efficiency = 0.608
input 3 3753.000 -213.687 0.000 3539.313
value movement movement
Technical efficiency = 1.000
5 0.010
output 2 216.000 0.000 0.000 216.000