多階層模式建構與應用
Applied Multilevel Data Analysis
| 節 | 週四 |
|---|---|
5 13:20–14:10 | 多階層模式建構與應用 HA216 3 節連堂 |
6 14:20–15:10 | |
7 15:30–16:20 |
* 根據陽明交大上課時間表所列
In this course, the methods for analyzing multilevel data will be introduced with the emphasis on practical, hands-on analysis and interpretation of hierarchical linear models. The following topics will be covered in this course: conceptual development of multilevel models, estimation, model evaluation, centering and heterogeneity treatment, power, longitudinal data, and an introduction to multilevel models with non-normal outcomes (e.g., dichotomous and ordinal data).
Students taking this course should have a strong background in multiple regression as well as ANOVA models (at a level equivalent HSIE30047), and be comfortable analyzing data in R. In this course we will learn and mainly use R for running the MLM analysis, as well as the Optimal Design software for power, and we will learn how to fit some of the models in Mplus and SPSS as well.
•In this course, we will mainly use R and RStudio. Please bring your own laptop and have R and RStudio installed. If you find yourself in need of a computer, please reach out to me. •In addition, please download SPSS and demo version of Mplus. •Late submission of assignments or lab projects will result in score deduction.
Assessment a)Participation/Attendance (Attendance and participation for course discussion is expected.) b)Article gap analysis (10%): Find a published article and identify personal gaps and improvement in knowledge about MLM and a short presentation of the article in the class. c)Lab projects (20%): Students will be working on guided lab projects (hands-on exercise) throughout the course. d)Assignments (40%): Students will be working on four problem sets (guided data analysis and write-up). e)Midterm (10%) f)Final in-class presentation (20%): Use existing data to build MLM models and prepare a presentation to share the findings. Students are encouraged to work as a team (two person maximum). The final project should include the following four sections: introduction, method, results, and discussion. Every student (or every team) will have 20 minutes to present the project.
| 週次 | 主題 |
|---|---|
| 第 1 週 | Introduction and course overview |
| 第 2 週 | Brief review of regression and software options (R review); Clustering and ICC |
| 第 3 週 | Teachers' Day ( working day but no class) |
| 第 4 週 | Two-level models: Part I (random-intercepts model, means-as-outcomes model) |
| 第 5 週 | Two-level models: Part II (random-coefficients model, and intercepts- and slopes-as-outcomes model) |
| 第 6 週 | Estimation and model fit evaluation |
| 第 7 週 | Centering predictor variables |
| 第 8 週 | Midterm exam |
| 第 9 週 | Modeling Heterogeneity and Residuals |
| 第 10 週 | MLM with longitudinal data I |
| 第 11 週 | MLM with longitudinal data II |
| 第 12 週 | Advanced topics I: Multilevel logistic model |
| 第 13 週 | Advanced topics II: Sample size and power, missing data |
| 第 14 週 | Extensions: Alternative to MLM such as generalized estimating equations and cluster-robust standard errors; MLM best practices |
| 第 15 週 | Project Presentation |
| 第 16 週 | Project Presentation |
Raudenbush, S.W., and Bryk, A.S. (2002). Hierarchical Linear Models (2nd edition). Newbury Park: Sage. O'Connell, A.A. & McCoach, D.B. (2008). Multilevel Modeling of Educational Data. Charlotte, NC: Information Age Publishing
- 地點
- Office: HA324
- 時間
- Office Hour: Mondays 10:00am-12:00pm by appointment
- 聯絡方式
- Email: mtlo@nycu.edu.tw Email is the best way to reach me outside of class for a prompt response. When you email, please include your name and class name in the title so that I can respond to your emails more efficiently.