在社會科學的分析中，我們雖然想方設法將許多現象予以數值化，以方便我們使用多元線性迴歸來進行統計分析。然而，更多社會現象是以類別的姿態呈現於真實世界，這種質性變項難以數值化，但卻是我們日常生活的常態。例如，你有多快樂？非常不快樂、不快樂、普通、快樂、非常快樂，許多研究基於方便可能將其視為連續變項，但正確來說，這是一個具有順序層級的類別變項，當它成為研究的依變項時，我們無法使用一般線性迴歸來估計。因此，處理類別資料的統計模型成為社會科學研究中重要的一個部份。本課程即在此基礎上，將帶領同學熟悉類別資料的分析。本課程除了進行統計理論的討論之外，也將透過揣摩期刊論文的操作、詮釋來幫助同學熟悉這些分析模型，更強調實際的資料分析，以期同學具有操作、分析資料、與詮釋資料的能力。
此課程的核心目標包含兩個部分，第一部分是引導學生具有閱讀與評論量化研究論文的能力；第二部分著重於量化資料的實際分析，以培養學生利用統計方法與軟體進行資料整理與比較的能力。台灣學界已經累積豐富具全國代表性的「次級資料」，這些大型資料庫將提供我們這門課程實作的分析素材，其中橫斷面資料是較容易掌握與入門的資料庫，因此本課程將以Stata統計軟體來示範分析「台灣社會變遷基本調查計畫七期二次家庭組」的資料。最後，期許修課同學完成一份具有質量的學期報告。
In the analysis of social sciences, we numericalize many phenomena to facilitate our use of diversified line replication for statistical analysis. However, more social phenomena appear in the real world in a genre way. This kind of change in nature is difficult to numerical, but it is a common situation in our daily lives. For example, how happy are you? Very unhappy, unhappy, ordinary, happy, very happy, and many studies may view it as a continuous change based on convenience, but correctly, this is a category change with a sequential level, and when it becomes a dependent on research, we cannot use general linear rebirth to estimate. Therefore, the statistical model for processing category data has become an important part of social science research. This course will be based on this course, and will lead students to familiarize themselves with the analysis of category data. In addition to conducting theoretical discussions, this course will also help students familiarize themselves with these analytical models by exploring the operation and commenting of journal articles, and further emphasize actual data analysis, so as to achieve the ability to operate, analyze data, and comment data.
The core objective of this course includes two parts. The first part is to guide students to have the ability to read and evaluate quantitative research papers; the second part focuses on the actual analysis of quantitative data to cultivate students' ability to use statistical methods and software to organize and compare data. The Taiwanese academic community has accumulated a large number of representative "secondary data" in the country. These large databases will provide analysis materials for our course implementation, among which the silence data is a database that is easier to master and enter. Therefore, this course will use Stata statistical software to analyze the data of the "Seventh Secondary Family Group of Taiwan Social Change Basic Survey Plan for the Seventh Period of Secondary Family Group". Finally, students in the courses will complete a quality academic report.
Categorical data analysis that deals with qualitative or discrete quantitative data is one of the most important statistical tools nowadays. In recent years, this tool plays a fundamental role on analyzing polychotomous data, particularly in the social and health sciences. This course introduces statistical theories and models for analyzing categorical data. The main topics cover :
(1) likelihood-based inferences on measures of association for two-dimensional and three-dimensional contingency tables under different assumptions. (2) generalized linear (mixed) models with emphasis on binary (Poisson) regression and logit models. (3) Repeated categorical data modeling, such as generalized estimating equation approaches and quasi-likelihood methods. (4) Asymptotic results and other advanced topics.
Categorical data analysis that deals with quality or discrete quantitative data is one of the most important statistical tools nowadays. In recent years, this tool plays a fundamental role on analyzing polychotomous data, particularly in the social and health sciences. This course introduces statistical theories and models for analyzing category data. The main topics cover :
(1) likelihood-based inferences on measures of association for two-dimensional and three-dimensional contingency tables under different assumptions. (2) generalized linear (mixed) models with emphasis on binary (Poisson) regression and logit models. (3) Repeated category data modeling, such as generalized estimating equation approaches and quasi-likelihood methods. (4) Asymptotic results and other advanced topics.
Long, John Scott. 1997. Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks: Sage Publications.
［2002，鄭旭智、張育哲、潘倩玉、林克明譯，《類別與受限依變項的迴歸統計模式》。台北：弘智。]
Long, John Scott, and Jeremy Freese. 2014. Regression Models for Categorical Dependent Variables Using Stata, third edition. College Station, Tex.: Stata Press.
Long, John Scott. 1997. Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks: Sage Publications.
[2002, Zheng Xuzhi, Zhang Yuzhe, Pan Qianyu, and Lin Keming, "Revolutionary Statistical Model of Category and Restricted Dependencies". Taipei: Hongzhi. ]
Long, John Scott, and Jeremy Freese. 2014. Regression Models for Categorical Dependent Variables Using Stata, third edition. College Station, Tex.: Stata Press.
| 評分項目 Grading Method | 配分比例 Grading percentage | 說明 Description |
|---|---|---|
| 1、 平日作業1、 平日作業 1. Weekday work |
60 | (1) 四份實作作業（40%）；(2) 導讀（依修課人數調整）：（20%） |
| 2、 學期報告2、 學期報告 2. Study period report |
40 |
