This study deals with long distance migration together with the importance of social networks and geographical preferences for the choice of destination when moving. Thus, the main purpose of this study is to analyze and specify the importance of the mover's social contacts when choosing a new place-of residence. In this connection, the social contents of the destination are chiefly focused. It has also been of interest to estimate in what degree these social contents have influenced this choice of destination by the individual. The differences in quality, intensity, dispersion and direction of social contacts is essential for explaining destination decisions. The basic hypothesis claims that the existence of social networks influences the individual's inclination to move and the choice ofplace-of-residence. In this study, empirical material is used, and its base is 2500 "movers" and a group of "stayers" of equal size. Data from public registers as well as data from inquiries have been used as empirical material. Far away moving is defined as moving across the communal boundaries. The examination of the empirical facts provides input paramétrés for a model, where a number of factors might explain the destination. The model work is implemented in three stages. In the multiple regression model, the impact of a number of variables is estimated. In the logistical regression model the probability of moving into a 'network commune' is estimated, i.e., moving to a commune where there are social contacts. In a third stage, the same variables are tested in a multi-nominal model. Strictly empirically, this study reveals, that there is evidence that the social networks influence the choice of destination when moving. Equally important, perhaps is to find out how strong an impact these networks have. Moving to a geographically and socially recognizable place is possibly regarded as having a more positive impact than the structure of the labour market. It can be stated that, in spite of the widespread migration within post-war Sweden, most people maintain some sort of social and geographical relation to the place where they grew up. The increased mobility per se has also implied that the individual, spatial and social contacts have become more widely spread. The place where one grew up is central and could be said to represent the place where there are many social contacts and the place you always refer to. Social contacts connected to relatives constitute one of the most important single factors. Previous place-of-residence is undoubtedly the most important single factor. Migration to a place of previous residence also implies that there already exists a social network. The access to a holiday cottage and the location of this cottage proves to have a great impact on destination- decisions. Many people simply move to the commune, where they own, or have access to, a holiday cottage. Of several plausible explanations two seem to be relevant. One concerns moving into former cottages, for year round residence, the other implies that the cottage could be regarded as the tangible link to the former home area. If there is no possibility of staying on in one's home area, the cottage represents a link to the past. / digitalisering@umu

本研究以經歷過921大地震以及多次土石流災害的雲林縣古坑鄉為個案，透過參與觀察法、質性訪談以及文獻檔案分析的方式，以了解當那些受到環境劇變的人們在經歷自然災害後，如何在有選的情況下做出遷移與否的決策，而哪些社會因素會在決策過程中影響人們的決定及行動。 本研究的重要發現包括以下幾點：第一，在自然災害發生後，多數的居民會選擇留下，這是基於本身的災害識覺排除了立即性、短時間內的生命威脅。同時，大部分的居民也會因為在地性的資源（如社會關係、在地知識、與在地的不動產）較多，而傾向選擇留在原地。同時，有一些社經弱勢家庭會因為非在地性資源的缺乏，也使他們不得不做出留下來的選擇。從生命史的觀點，本研究發現這些遷移決策往往是鑲嵌在長期累積的地方經驗與生活脈絡中的。第二，家庭的遷移往往是家庭成員整體的決定，或考量家庭整體需要後做出的決定。不同的家庭可能採取不同的方式來完成他們在地重建或者是遷移的目標。第三，個人和家庭的慣習往往在選擇的傾向或行動的方式中成為核心。多數人希望能在自然災害發生後恢復既有的生活慣習，但慣習的維持不僅在於個人能力，也與他所在環境中是否有足夠的社會支持有關。當社區中的社會安全網可以提供足夠的支持時，個人與家庭的慣習則得以被維持，而社區社會安全網又和地方產業型態具有緊密關聯。以觀光為主要發展的草嶺村，和在921大地震後轉型為觀光農業的華山村就是兩個極佳的對比案例。研究資料也顯示當地居民改變社會環境的可能，證明了人不是單向地受到社會環境的影響，也可以反向影響社會環境。 / This study aims at understanding how people make migration decisions after drastic environmental changes brought by natural disasters and exploring the social factors influencing these decisions. Gukeng was chosen as a case for this study due to its encounters of the Chichi earthquake and several landslides. The study utilizes qualitative research methods and collects field data, mainly through participant observation, in-depth interviewing techniques and text analysis. The main findings of this study are summarized as follows. First, most residents tend to stay after natural disasters, and such choice is based on their hazard perception that excludes immediate and short-term threats to their lives. The majority of residents prefer to stay due to more local resources（including social relationships, local knowledge, and ownership of real estate property）. Those families in lower social-economic status also tend to stay for the lack of non-local resources. From the perspective of life history, these decisions are embedded in local contexts contructed and experiences accumulated throughout the years. Second, family’s migration decision is usually made by all family members or based on the needs of all members. Different families may use different ways to reach their goals of migration or rebuilding on the same ground. Third, the habitus of individuals and families often become the core element in the modes of choice and action. Most people want to be able to recover their habitus after the natural disasters. However, the maintenance of the habitus is based not only on personal capacity, but also on the social support drawn from the his environment. People can maintain their habitus when the local community safety net can provide enough support, and the local community safety net is highly related to the development of local industries. The comparative study of Caoling Villege and Huashan Villege provides an illustrative example, in which the former is based on tourism and the latter was transformed into eco-tourism and tourist agriculture after the Chichi earthquake. The findings also show the possibility of people changing the environment, proving the mutual influence between individuals and the environment.