本文目的旨在了解台灣民眾運動頻率與運動樂趣間之關係，並探討影響運動頻率與運動樂趣的決定因素。由於個人真正運動頻率與運動樂趣是無法直接被觀察到，研究者所擁有的資訊是問卷受訪者對運動頻率與運動樂趣兩變數自行主觀認定和自我評估的結果；亦即，我們觀察到的是受訪者運動時所享受到的樂趣和多久做一次運動的等級選擇，針對此類問題研究，我們採用二變量順序多重probit模型（bivariate ordered probit model）。利用2007年《台灣社會變遷基本調查》休閒組資料，我們建立一個運動頻率與運動樂趣二變量關係的模型，並以充分訊息最大概似估計法（full-information maximum likelihood estimation）估計之。研究結果顯示，影響台灣民眾運動樂趣因素有教育程度、年齡、父母親籍貫、父母教育程度、自身健康狀況、居住地。影響台灣民眾運動頻率因素有勞動工時、所得、年齡、父母親籍貫和父母教育程度。兩者顯示具有因果關係，即運動樂趣愈高者，其運動頻率也愈高。 This thesis aimed to investigate the relationship between adults’ sports pleasure and the frequency of sports in Taiwan. Due to a person’s sports pleasure and the frequency of sports are not directly observed, what researchers have are respondents’ report about the categories of the subjective assessment of their exercise frequency and sport fun status, therefore, a bivariate ordered probit model is used to solve this problem. Using “Taiwan Social Change Survey 2007” leisure group data, we build a model on sports pleasure and the frequency of sports two-variable relationship, and to estimate the model by full-information maximum likelihood estimation. The results show that education level, age, parents’ ethnic group affiliation, health status, urban-rural area have significant effects on adults’ sport pleasure. Besides, adults’ sports frequency are significantly determined by working hours, income, age, parents’ ethnic group affiliation and parents’ education level. The result also supports that there is a causal relationship between sports pleasure and the frequency of sports, that is, the higher the sports pleasure, the higher the exercise frequency.
