台灣地區家戶特性與區域特性對於住宅竊盜被害風險之影響

Residential and Contextual Influences on Burglary Victimization in Taiwan

邱靖方(Jing-Fang Chiu)

指導教授 ： 黃蘭媖

國立臺北大學/社會科學學院/犯罪學研究所/碩士(2009年)

摘要

本研究結合「2000年台灣區域犯罪被害經驗調查研究」、內政部戶政司「70,75,80-96年底各鄉鎮市區人口數」與行政院主計處1996年「台閩工商及服務業普查報告」的資料的進行次級資料分析，茲將住宅竊盜的風險因素區分為微觀層次的「家戶特性」與鉅觀層次的「區域特性」，欲了解區域特性與家戶特性解釋住宅竊盜被害風險的情形；兩者結合之後，解釋住宅竊盜被害風險的模式是否會發生變化？
本研究係根據結構選擇被害理論的四個概念（系絡因素、標的吸引性、防衛與暴露），將研究架構分成「客觀系絡因素」、「主觀系絡因素」、「標的吸引性」、「社會防衛」、「物理防衛」與「暴露」等六個部分，從前揭被害調查資料與社會統計中選出所需的變項。原始被害調查共有10427個樣本，經研究者整理後共計使用10358個樣本。分析方式包括以多元迴歸(Multiple Regression)分析台灣地區各鄉鎮市區被害率的影響因素；以及Logistic迴歸(Logistic Regression)分析不同區域住宅竊盜風險因素，並比較結合區域特性與家戶特性前後，解釋各地區家戶住宅竊盜被害風險的模式為何。
分析結果相當程度地印證了結構選擇理論的命題，區域中的人口流動性與人口異質性是解釋區域住宅竊盜被害率的重要因素。關於影響各地的住宅竊盜被害風險因素，研究結果的發現如下：家戶特性中以「標的吸引性」與「暴露」較能解釋各地家戶的住宅竊盜被害風險，而且各地的情形不同。結合家戶特性與區域特性分析不同區域的被害風險因素，整個模式的解釋力大幅改善，而某些家戶特性的效果更凸顯；將家戶特性與區域特性結合之後發現：客觀系絡因素在各地區的影響方式不同、主觀系絡因素（居民對於周遭環境問題之感受）未必反映該地區的情形；標的吸引性對於住宅竊盜被害風險的影響，須結合防衛程度一併考量；社會防衛對於住宅竊盜被害的影響效果有限；物理防衛與住宅竊盜被害只有在工商市鎮有關聯；暴露程度與住宅竊盜被害之關連，不是純粹的線性正相關。
由於被害調查的資料的問項未能完全貼近理論中概念的測量，故本研究對於被害調查的項目提出相關的建議：系絡因素援引集體效能(collective efficacy)的概念，加入鄰里關係的測量項目；標的吸引性之測量，以竊賊較能觀察且可接近者為佳；家中防衛設備的測量，需要區分被害前後；暴露的測量加入其他的項目，例如「房屋類型」、「房屋週遭環境」與「家人的防竊意識」等；住宅竊盜被害經驗的問題應能區分「經同意進入」與「未經同意進入」以及「侵入既遂」與「侵入未遂」。

關鍵字

結構選擇理論 ； 住宅竊盜 ； 臺灣 ； 被害調查

並列摘要

This study combines data from 2000 Crime Victim Survey in Taiwan, Population for Township and District (Source: Dept. of Household Registration Affairs, MOI) and 1996 Industry, Commerce and Service Census (Source: Directorate-General of Budget, Accounting and Statistics, Executive Yuan) to investigate how residential and contextual characteristics influence household burglary in Taiwan.
The analytic framework is guided by “structural-choice” theory of victimization, which consists six sets of determinants. They are: objective proximity, subjective proximity, target attractiveness, social/interpersonal guardianship, physical guardianship and exposure. The variables are selected from 2000 Crime Victim Survey in Taiwan and created according to previous social statistics. Crime Victimization Survey in Taiwan 2000 concludes 10427 samples; among them, 10358 samples are selected and then analyzed via methods of multiple regression and logistic regression. I apply multiple regression to examine contextual influences on household burglary rate in township and district. Logistic regression is used to examine the combination of residential and contextual influences on household burglary risk in different areas.
The major findings are as follows:
First of all, household burglary rate of every township and district is explained by population transience and heterogeneity.
Secondly, “target attractiveness” and “exposure” account more for household burglary risk among all of the household traits. However, factors work differently among areas. When both residential and contextual factors are introduced, the original models are modified. The goodness of fit enhances in great degree and some household factors stand out. Objective proximity works differently among areas. Subjective proximity does not reflect conditions of areas. The effect of target attractiveness must be considered along with guardianship. The effect of social guardianship is poor. Physical guardianship and household burglary risk is related only in industrial and commercial towns. The relationship between exposure and burglary risk is not linear. In other words, only when the house is left empty for more than three hours, burglary risk increases.
Some suggestions are proposed for future victimization surveys in Taiwan. According to collective theory, measurements of proximity may include neighborhoods relationship. Measurements of target attractiveness are better when offenders’ decisions (the accessibility and value of target) are considered. Measurements of physical guardianship must distinguish pre victimization from post victimization. Measurements of exposure can add other items such as “location of house”, “type of house”, etc. The screen question should be able to differentiate break-in from entry under permission and burglary from intended burglary.

並列關鍵字

structural-choice theory ； burglary ； Taiwan ； victimization survey

參考文獻

許春金、邱淑蘋，2006，〈住宅竊盜被害者報案決定因素與被害反應之研究〉。《中央警察大學犯罪防治學報》7：65-98。

王文忠，2003，《住宅竊盜被害與預防之研究》，國立中正大學犯罪防治研究所碩士論文。

邱豐光，2008，《常業住宅竊盜犯罪歷程之研究》，國立台北大學犯罪學研究所碩士論文。

吳鄭重 譯著，Jane Jacobs 著，2007，《偉大城市的誕生與衰亡－美國大城市街道生活的啟發(The Death and Life of Great American Cities)》。台北：聯經。

Bernasco, Wim , and Floor Luykx. 2003. Effects of attractiveness, opportunity and accessibility to burglars on residential burglary rates of urban neighborhoods. Criminology 41 (3):981-1001.