本文試著以空間統計分析方法探討台灣地區教育成就之城鄉差距。空間自我相關性可估計資料區域聚集的關聯及強度，並可反映資料空間分佈的特性。分析資料以96年大學學科能力測驗成績為主，並利用學生戶籍地郵遞區號的資料，輔以台灣地理資訊系統的應用，以地理區域圖形描繪出在台灣地區學測成績的城鄉差距及其分佈聚集的型態，除了可進一步了解358個鄉鎮之間在學測成績空間分布，並以空間聚集統計值呈現出城鄉教育發展失衡的程度。本文的貢獻在於以空間群聚指標分析學習成就之城鄉差別，避免以人為的行政區域劃分資料，因而可降低估計上的加總偏差；其次，實證結果發現部分鄉鎮的學測表現並不亞於大都會地區，然而部分都會區之教育成就並不如預期中表現優異；最後，從空間例外區之分佈特性隱含改善教育資源分配的問題與契機。

In this paper we take a spatial statistics approach to examine the Taiwan's urban-rural gap in educational attainment. The spatial autocorrelation statistics measure the degree of dependency among observations in a geo-space so that the spatial clustering association and intensity of the data at proximal locations can be analyzed. Based on the data on the subject competency test in 2007, along with the zip code of students' permanent residence, we integrate spatial statistics and Geographic Information Systems (GIS) to graphically map the clustering patterns of test scores across the 358 townships of Taiwan. In that sense, the spatial clustering index provides a better view into the spatial distribution of test performance, and highlights the magnitude of the unbalanced education development over the Taiwan region. The contributions of this paper are threefold: First, we spatially depict the urban-rural disparity in educational attainment so as to avoid artificial data division imposed by administrative districts. As a result, the problem of the aggregation bias can be eliminated. Second, our empirical finding indicates that the academic performance in a part of the townships is significantly greater than that in metropolitan areas. But contrary to common belief, test outcomes in several urban areas are not as exceptional as observed by the previous studies. Finally, the scattering of spatial outliers reveals the inherent problems and the promising solutions for improvement of educational resources allocation.

初婚夫妻的居住安排，不僅是探討家庭形成的起始點，更是瞭解家庭型態與結構發展的重要事件，然而國內家庭研究卻甚少關注。故本研究以第一次結婚的受訪者作為分析對象，採用華人家庭動態資料庫中兩個主樣本各兩波的追蹤調查資料，經篩選出有效初婚夫妻樣本1176對。分析方法採用兩個邏輯迴歸模型，分別探討先生人口與社經資源變項對於初婚夫妻居住安排的影響，夫妻相對權力關係與代間同住選擇的因果關係，並釐清上述因素是否具有社會變遷的影響效果。
本研究結果大致符合資源論與生命歷程的理論預期，先生樣本的資源特徵對於選擇自立門戶或代間同住有顯著的影響，亦即結婚年齡愈早、結婚年輪愈晚與曾因工作而離家者，愈可能傾向選擇自立門戶。再就夫妻樣本的分析結果，亦發現夫妻間佔有資源優勢的一方，會傾向有利於自身原生家庭的代間居住型態。值得注意的是，教育程度變項的影響效果，卻與研究預期有所出入，特別是中、高教育程度的丈夫，反而會強化以與父母同住的方式來實踐傳統孝道倫理，且當先生教育程度大於太太時，相對會破除傳統父系的思想限制，接受與太太娘家同住的事實。
由於兩個邏輯迴歸模型中的收入變項，均因缺漏值過高而無法精確估計出其對初婚夫妻居住安排選擇的影響效果。雖經原始樣本與篩選樣本基本特徵的比對，初步排除選擇性偏誤的問題，但是兩個樣本中從妻居佔代間同住比例甚高的現象，發現本研究分析受訪者在剛結婚時的居住安排，與前人調查受訪當時的居住安排有明顯差異。故新婚夫妻先選擇從妻居，再改變至其他的居住安排，此新家戶組成與後續轉變的過程，值得未來深入研究。

Although the living arrangements of first married couples (FMC) are influenced by the original shape of Chinese family composition, patriarchal tradition, and the development of family structure, little attention has been given to this topic. In order to clarify the phenomenon of new household formation, this study introduced the resource theory and life course perspective. The data used combined two-waves and two panels (RI 1999, RII 2000, RI 2000 and RII 2001) of the ”Panel Study of Family Dynamics” (PSFD), with 1,176 valid FMCs selected in total. Two logistic regression models were employed to figure out the living arrangements decision of newlywed couples, including neolocal, patrilocal or matrilocal residence. The results find that those who married at a younger age, the later marriage cohort, and those experienced leaving home for work are more likely to choose to live independently. The couple relative power model estimations are similar to husband's findings. The predominant party of the FMC tend to decide to reside with their own family. Most research results supported the life course and resource theory, except for the education variable's effects on the living arrangements. Contrary to hypotheses, the newly married husbands with higher education achievement are more likely to choose intergenerational co-residence or even matrilocal residence.

本文引入設計矩陣的概念，使用幾個簡化的例子說明傳統的迴歸分析與變異數分析乃是同一個分析方法，也指出次數分配的分析可以參數化，適用同一套最小平方法的分析，打破傳統方法與計量教科書有關計量尺度與分析方法的迷思。本文同時指出，設計矩陣的安排對應著模型與參數設定，儘管矩陣的階序相同，不同的矩陣內容代表不同的模型與參數設定及研究設計，則進行資料分析以前，一但有了分析模型與參數設定的構想，研究者宜乎審慎考量其設計矩陣與模型設立的對應關係，確認設計矩陣正確代表其研究設計。正交的設計矩陣保證參數估計「分別」為之，互不相干，不會產生解釋變項間牽扯不清的問題，乃研究者需要多加注意的設計。在討論中，本文也指出傳統方法與計量教科書對於分析單元的主張未必能適應學術的進展，尤其在多層次分析與機率或次數分析的架構下，此一主張有需要修飾之處。

Introducing the concept of a ”design matrix”, this paper relates the regression analysis, the analysis of variance, and the analysis of frequency counts to the use of the least squares method. It is demonstrated with some simplified examples showing that, given a properly arranged design matrix, the analyses of means, probabilities, and frequencies share the same least squares approach as that of the regression analysis. Special attention is called to the ”orthogonal” design matrix which results in independent parameter estimations jointly. Researchers are urged to review the possible arrangements of the design matrix before actually doing on the estimation.

本研究主要依據集體記憶、跨國主義和社會資本等理論，從不同的空間尺度，探討加拿大臺灣移民在居住地選擇時的思維與作用力。
透過2001年加拿大普查資料與文獻回顧，了解溫哥華及多倫多兩地臺灣移民概況及基本特徵，並於兩地選取40位臺灣出生移民進行深度訪談，以探討其居住地選擇過程與想法，並與理論做呼應。
研究結果發現，加拿大臺灣出生移民因居住環境佳及天氣因素而選擇溫哥華，且集中在Vancouver、Burnaby和Richmond等地區。因家人親戚與求學環境而選擇多倫多，並聚居在Toronto、Mississauga、Markham和Richmond Hill等地區。一直以來，他們在臺、加二地間，跨國界的將文化進行傳播、保留與複製，建立了一個跨國界的社會領域。同時，也在此領域中建立起特有的社會網絡，累積了社會資本。此資本不但可以幫助原居地文化跨國傳播到移居地，建立起族裔的集體記憶，也可以提供最新的資訊，讓移民在移居地找到好的居所，並取得新的社會認同。這樣的氛圍吸引了臺灣移民特別喜歡居住在某些特定的城市與地區。

The purpose of this research is to explore how Taiwanese migrants select their residential locations when they move overseas using the concepts of collective memory, transnationalism and social capital.
The methods employed in this research included a review of the 2001 Canadian Census data, followed by an extensive literature survey on the socio-demographic characteristics of Taiwanese migrants in Vancouver and Toronto before a total of 40 in-depth interviews were carried out in the two cities, investigating how their decisions were made in choosing their residential location.
This research finds that most Taiwanese migrants chose to live in Vancouver and Toronto. The mild climate and pleasant environment were the main determinants they cited for deciding to live in Vancouver, Burnaby and Richmond were the specific suburbs they congregated in. For those who chose to settle in Toronto, they did so mainly for family and education reasons; and Toronto, Mississauga, Markham and Richmond Hill were the locations in which they preferred to settle.
Having spent time living between Canada and Taiwan, these migrants had constructed a transnational space that allows them not only to fuse the cultures between the two places, but also to build social networks and social capital that enable collective memory to develop, further encouraging information on residential suburbs to spread. In turn, these locations acquired an identity as preferred choices for urban living among Taiwanese immigrants.