PHP 2517: Applied Multilevel Data Analysis

Spring 2024

Instructor: Anarina Murillo, PhD

Email: anarina murillo@brown.edu
Office: 121 South Main Street, Room 713
Office Hours: TBD, see Canvas

Teaching Assistant: Mavis Liang

Email: xinwen liang@brown.edu
Office Hours: TBD, see Canvas

Lectures: Monday and Wednesday, 9:00 AM - 10:20 AM

Class Room: 121 South Main Street, 3rd Floor, Room 375

Labs: During lecture on Wednesday, 9:00 AM - 10:20 AM

Course Overview
This course provides a survey of statistical methods for modeling and analyzing data that are col-
lected at multiple levels of sampling. Multilevel data structures are found in longitudinal studies,
clustered designs, and many other settings where data are structured at multiple levels of aggrega-
tion. Statistical methods for analyzing this type of data take into account their multilevel structure
to draw proper inferences and appropriately partition multiple sources of variation. This course
will approach the topic primarily from the point of regression modeling with focus on generalized
linear mixed effects specified in a hierarchical fashion. Real data examples will be used to motivate
the material and R software will be used to carry out most analyses.
This course is designed for graduate and advanced undergraduate students who will be analyz-
ing data and want to develop a practical hands on toolkit as well as understanding of the theoretical
underpinnings of models for multilevel data. Students can work using any statistical software they
feel comfortable with. Lab notes will be provided with examples of analyzing multilevel data using
R statistical software.

Prerequisites
This course is designed as a half credit course. It will begin mid of the semester and will meet
twice a week until the end of the Spring semester. All students who wish to enter this course
must have completed PHP 2511 or PHP 2514 prior to enrollment. Students who have completed
PHP 2507/2508 may be admitted at the discretion of the instructor, this may require additional
independent study to cover extra material. The course will move at a quick pace and will focus
more on the applications of multilevel data analysis as opposed to the theory behind them.

1
Course Competencies
The purpose of this course is to provide training necessary for a solid foundation in statistical
methods for multilevel data analysis with an emphasis on applications in order to be able to work
collaboratively in an environment that values skill in analysis of data for important problems in
biomedical research (biology, public health and clinical medicine).

1. Demonstrate a foundation in statistical theory and methods for standard designs and analyses
encountered with biomedical data.

2. Identify and implement statistical techniques and models for analysis of data.

3. Acquire knowledge and skills in research methodologies to collaborate with substantive in-
vestigators.

4. Recognize key research designs and be able to assist in developing plans for their implemen-
tation.

5. Apply programming skills to analyze data and develop simulation studies.

6. Attain proficiency in management, and documentation of study data for use in practical
statistical analysis.

7. Formulate a public health question in statistical terms.

8. Develop proficiency in making oral, written and poster presentations of work to statistical
and non-statistical colleagues.

9. Review and evaluate the use of biostatistical methods in public health or biomedical field of
study.

Course Objectives
By the end of this course, students should be able to do the following:

1. Understand and identify multilevel and/or hierarchical structure in a particular dataset or

sampling design.

2. Perform exploratory data analysis for multilevel data.

3. Specify an appropriate regression model that can capture both individual-and group-level
effects.

4. Understand the nature and consequences of within-and between-group variation for models
having two or more levels.

5. Understand and model covariance structures.

6. Interpret the output from multilevel models, with emphasis on distinguishing individual- and
group-level covariate effects.

7. Specify, fit and provide contextual interpretation of multilevel models for addressing open-
ended data analysis questions.

2
 8. Apply statistical software to perform exploratory data analysis for longitudinal data.

9. Apply statistical software to organize, summarize, and present data using graphical and
tabular representations.

10. Apply statistical software to fit multilevel models.

Books, Supplies, and Materials

Course Text and Other Readings
The course will cover material found in the textbooks below. Additional resources, such
as journal articles and notes, may be provided as needed.
• Required: Gelman, A., & Hill, J. (2007). Data Analysis using Regression and Multilevel
Hierarchical Models. (Analytical methods for social research). Cambridge University
Press. ISBN 978-0511790942. http://www.stat.columbia.edu/~gelman/arm/
• Recommended: Stroup (2013). Generalized Linear Mixed Models. CRC Press. This
book focuses on SAS implementation of GLMM.
• Recommended: Faraway JJ (2016). Extending the Linear Model with R, 2nd Edition.
CRC Press. Slightly more technical than Gelman and Hill, with focus on using R.
• Lab materials will be posted on Canvas.
• Additional article readings may be assigned during the semester.

Statistical Software This course uses R/RStudio for labs and homework assignments. Students
should install software prior to the first day of lab.

• R Statistical Software and RStudio is available for free at http://cran.r-project.

org/ and http://www.rstudio.com/products/rstudio/download/. You may down-
load the most recent version that is compatible with your device.
• In this course, we will use R/RStudio statistical software. The book’s website Data
Analysis Using Regression and Multilevel/Hierarchical Models provides code for all
major book examples in STATA, R, and SAS. Students should refer to this and other
online resources for individual coding help. Students can work using any statistical
software they feel comfortable with. You may use any programming language with
which you’re more comfortable with for assignments. However, the teaching team will
not provide support for other programming languages.

Book/Course Material Support Pilot Program If your Brown undergraduate financial aid
package includes the Book/Course Material Support Pilot Program (BCMS), concerns or
questions about the cost of books and course materials for this or any other Brown course
(including RISD courses via cross- registration) can be addressed to bcms@brown.edu. For
all other concerns related to non-tuition course-related expenses, whether or not your Brown
undergraduate financial aid package includes BCMS, please visit the Academic Emergency
Fund in E-GAP (within the umbrella of “E-Gap Funds” in UFunds) to determine options
for financing these costs, while ensuring your privacy.

3
 Course Website (Canvas): The canvas site will contain all the information for this course
including this syllabus, office hours, weekly course info, and posted grades. Make sure that
you are enrolled on Canvas and that you check the site regularly.

Zoom: Office hours may be offered using Zoom. Please check Canvas for details.

Ed Discussion: This course will use Ed Discussion to allow the class to ask questions about
lecture, labs, or homework. The teaching team will check the discussion posts regularly and
respond to questions. Students in the course can answer each other’s questions. Please note
that while Ed Discussion will be used to answer questions about course content, all announce-
ments about class meeting times and exam or assignment due dates will be communicated
using Canvas announcements.

Method of Evaluation
Your course grade will be based on:

Item Percentage
Homework Assignments 40%
Participation 10%
Final Presentation 30%
Final Paper 20%

Homework assignments (HW)

There will be 3 assignments each worth 35 points each. The assignments will require statisti-
cal software for data handling, comprehensive data analysis based on techniques we learn in
class, and interpretation of the results. Assignments must be submitted through Canvas no
later than 11:59 PM on the due date. All assignments should be completed individually. You
may discuss homework with your classmates. However, you must work on your assignments
individually, submit your own assignment, and write your own analysis and conclusions. It
is not acceptable to send or copy the homework solutions or codes for assignments to other
classmates. If you have any questions about what constitutes plagiarism, please consult the
Brown University Code of Conduct.

Final Project
The final project will be a group project and will be comprehensive in nature requiring
students to plan and carry out an in depth analysis of multilevel (or longitudinal) data. The
final grade will be determined by a written report, oral presentation, and a peer evaluation.

Participation
Participation in-class is strongly encouraged. There will be a grade for participation.

Grading Policy
In accordance with Brown’s grading system, ABC/NC or S/NC grades will be tentatively
assigned based on the following cutoffs:
ABC/NC: A: 90%-100%, B: 80%-89%, C: 70%-79%, NC: 0%-69%
S/NC: S: 70%-100%, S with Distinction: 90%-100%, NC: 0%-69%
These ranges are tentative and the cutoffs for final grades will be determined at the end of
the semester.

4
Tentative Schedule
Week Date Topic Readings Assignment
1 3/18 Lecture 1: Course Overview, Introductions GH CH 3.1-3.6
3/20 Lecture 2: Review of Linear Regression Models GH CH 4.1-4.5
Lab: No Lab

2 3/25 No Class, Spring Break

3/27 No Class, Spring Break

3 4/1 Lecture 3: Multilevel Linear Models (MLM) GH CH 1.1-1.3

4/3 Lecture 4: Model Fitting, Interpretation (MLM) GH CH 12.1-12.9
Lab 1: Exploring Multilevel Data

4 4/8 Lecture 5: Varying Intercepts and Slopes (MLM) GH CH 13.1-13.4 HW 1 Due

4/10 Lecture 6: Non-Nested Models (MLM) CH 8, 9
Lab 2: Multilevel Linear Models

5 4/15 Lecture 7: Review of Generalized Linear Models GH CH 14.1-14.2 HW 2 Due

4/17 Lecture 8: Multilevel Logistic Regression GH CH 14.1-14.2
Lab 3: Multilevel Logistic Regression

6 4/22 Lecture 9: Multilevel Poisson Regression GH CH 15.1-15.3 HW 3 Due

4/24 Lecture 10: Multilevel Poisson Regression GH CH 15.1-15.3
Lab 4: Multilevel Poisson Regression

7 4/29 Lecture 11: Sample Size and Power Calculation GH CH 20.1-20.4

5/1 Lecture 12: Case Studies

8 5/6 Final Presentations, In-Class

5/8 Final Presentations, In-Class

9 5/15 Final Paper and Presentation Due on Canvas

General Expectations
Overall Course Expectations
Students in this course will be expected to do the following:

1. Attend all lectures and actively participate in discussion.

2. Read material provided prior to each lecture and actively participate in class discussions.
3. Complete and turn in all assignments on time. Solutions to homework must be clearly
written with appropriate tables and figures included.
4. Demonstrate an understanding on material on final project.

5
 5. Respect each other, each others questions and each others discussion.

Readings
Please complete the required reading in advance of the lecture. This course will require
dedicated time for full understanding of concepts.

Lecture and lab attendance

It is expected that students attend lectures and labs. This is important for understanding
concepts and remaining engaged in the learning process throughout the semester. Students
do not need to inform the instructor and/or TA if they will miss a lecture or lab. Students are
encouraged to make arrangements with other students in the classroom to acquire missed
lecture notes. Students with excused and/or long-term absences should discuss this with
their academic advisor and instructor.

Late or missed homework assignments

Homework (HW) assignments must be turned in online at or before the posted due date
which is Monday of the specified week. Should you turn your homework in late, you will lose
10% points for every additional late day up to two days total or Wednesday of the specified
week. This means assignments will not be accepted after Friday of the specified week.

Participation
This course will move in a fast pace and active participation in class discussions is a crucial
component for success in the course. Unexcused absences are not allowed and may result in
a loss of percentage points or even an incomplete (INC) at the end of the semester.

Credit hours and time expectations

Over the course of a half semester students are expected to spend a total of 38 hours in class,
42 hours on class preparation, 70 hours completing homework assignments, 20 hours on the
final project for an overall 170 hours. These amounts of time will vary by students.

Academic integrity Plagiarism will not be tolerated in this course.

• Review the Academic Code: (https://college.brown.edu/design-your-education/

academic-policies).
• Plagiarism occurs when you deliberately use someone else’s original material without ac-
knowledging source (i.e., carelessly or inadequately citing ideas and words from another
source). Other forms of plagiarism includes submitting an assignment or exam where
some or all of the work is written by another person. You must properly document
the sources used even if you are not directly quoting from another person’s work. This
includes citing online sources (e.g., Wikipedia, online books, Stack Exchange, Wolfram,
ChatGPT and related tools, etc). If you have any questions about properly citing your
work, please speak to the professor.
• Additional comments on the use of artificial intelligence (AI) language mod-
els (e.g., ChatGPT): Students may not use ChatGPT tools to complete homework
assignments. ChatGPT may be used as a study guide for the course but not for complet-
ing homework or the take-home quiz. If you have any questions about using ChatGPT
or other AI language models, please speak to the professor.

6
 • All situations of suspected academic dishonesty will be handled in the manner as de-
scribed in the Academic Code.

Communication about the course

Please check your Canvas regularly so that you can receive emergency communications and
periodic updates about the class.

Communication with the teaching team

The preferred method of communication with the instructor is e-mail. Please use your
Brown email account to send correspondences. Your emails should be professional in nature.
However, students should not wait until after Friday to send an email if a response is needed
before the following Monday. Weekly office hours are the best time to ask the instructor
questions regarding course materials.
Direct communication with the TA’s will take place during office hours only. Please do not
email the TA’s with questions about class materials and do not request individual meetings
with the TA’s. Questions about class materials should be asked during office hours.

Class recording and distribution of materials

All lectures and course materials are copyrighted. Students are prohibited from reproducing,
making copies, publicly displaying, selling, or otherwise distributing the recordings or tran-
scripts of the materials. Disregard of the University’s copyright policy and federal copyright
law is a Student Code of Conduct violation.
Students with accommodations approved by Student Accessibility Services should speak with
the instructor and TA’s before recording lectures and/or meetings for their private use.

Course syllabus subject to change

The syllabus content and course schedule may change as the class progresses. Please check
Canvas regularly for updated information.

Academic Support
Diversity and inclusion
A diversity of backgrounds, skills, and perspectives is critical to addressing public health
issues. We all benefit from diversity in the classroom. It is the responsibility of all of us to
maintain an environment that is respectful and supportive. Brown University is committed to
full inclusion of all students. The instructor and TAs intend to present materials and activities
that are respectful of diversity: gender identity, sexual orientation, disability age, socio-
economic status, ethnicity, race, culture, perspective, and other background characteristics.
Please keep in mind that some studies that will be discussed in class that may have not
considered diversity when the data were collected. In those cases, the instructor and TAs
will do our best to provide context.

Accessibility and accommodations

Brown University is committed to full inclusion of all students. Please inform the instructor
early in the semester if you may require accommodations or modification of any of course
procedures. You may speak with me after class, during office hours, or by appointment. If you
need accommodations around online learning or in classroom accommodations, please be sure
to reach out to Student Accessibility Services (SAS) for their assistance (seas@brown.edu,

7
 401-863-9588). Students in need of short-term academic advice or support can contact one
of the academic deans in the College.

Health and safety statement for course meetings

The class will be held in-person and office hours may be in person over Zoom. During both
lecture and lab, a face mask is not required in university facilities. However, for in-person
meetings with the instructor or TAs, I will ask students to wear a face mask when meeting
with me if in close proximity.

Use of technology to support student learning in your course

This course will use the following technological platforms: Canvas, R/RStudio software, Top
Hat, and Zoom. If you have any concerns or questions about access or the privacy of any of
these platforms, please reach out to me. The IT Service Center (https://it.brown.edu/
get-help) provides many IT Services including remote assistance, phones, tickets, and chat.
Please also see the Online and Hybrid Learning Student Guide.