Skip to content Skip to navigation

Register today! Workshop on Power and Sample Size for Multilevel and Longitudinal Study Designs

Tuesday, November 28, 2017 - 09:00 to Thursday, November 30, 2017 - 17:00

10-28 R25 Web Banner

Course Overview


 

The Harvard Catalyst, the Harvard Medical School Office for Diversity Inclusion and Community Partnership andthe Disparities Research Unit at the Massachusetts General Hospital are delighted to sponsor and support a three-day power and sample size workshop. The short course is offered in cooperation with Margarita Alegría, Ph.D., chief of the Disparities Research Unit at the Massachusetts General Hospital and professor in the departments of medicine and psychiatry at Harvard. The short course, taught by Keith E. Muller, Ph.D., professor in the University of Florida College of Medicine, and Deborah H. Glueck, Ph.D., associate professor in the University of Colorado School of Public Health, will give scientists training for selecting a valid sample size for balanced longitudinal and multilevel study designs with normally distributed outcomes. The workshop is accessible to all, from graduate students to senior researchers. Enrollees must have had introductory statistical training and be comfortable with basic hypothesis testing.

 

The course covers studies with response variables that follow a bell-shaped (normal) distribution, for:

  • multilevel designs
  • longitudinal designs
  • multivariate designs
  • combinations of multilevel, longitudinal or multivariate designs
  • some kinds of missing data
  • We do not cover binary or Poisson outcomes or time-varying covariates

Date


HostLocationDateRegistration Deadline
Harvard Medical SchoolBallard Room at Countway Library (HMS)Tues., Nov. 28 – Thurs., Nov. 30Tues., Nov. 21

Cost


Eligible Harvard Students and Fellows selected for scholarship*Free
Harvard and Harvard-affiliated Hospitals and Research Institutes  (click here for list of eligible affiliates).$250
Other Registrants (Not affiliated with Harvard)$1,000

* Students and fellows interested in attending the course may receive a scholarship with the approval of Dr. Margarita Alegría. Email Sandra Salloum, academic specialist, at SandraSalloum@ufl.edu to begin registration.

Registration


Harvard Students & Fellows Applying for ScholarshipEmail Sandra Salloum, academic specialist, at SandraSalloum@ufl.edu to begin registration. Please include a 1-2 paragraph description of your work and relevance of the program. Scholarships are limited.
Harvard and Harvard AffiliatesClick Here (select $250 option)
Non-Harvard AffiliatesClick Here (select $1,000 option)

For questions, please contact Sandra Salloum, academic specialist, at SandraSalloum@ufl.edu.

Workshop Objectives


At the end of the workshop, attendees will be able to:

  • Understand a framework and strategy for study planning
  • Write study aims as testable hypotheses
  • Describe a longitudinal and multilevel study design
  • Write a statistical analysis plan
  • Ensure sufficient sample size for racial or ethnic subgroups for studies of disparities
  • Demonstrate the feasibility of recruitment
  • Describe expected missing data and dropout
  • Write a power and sample size analysis that is aligned with the planned statistical analysis
  • Write the design, analysis, power and sample size sections for a grant
  • Use expert feedback to improve the product

Workshop Schedule


While the exact content of each workshop will vary, please see the model schedule below for an idea of what is covered each day.

Day 1: Study Design
  • 9 a.m.-12 p.m.: Understanding four common study designs and factors in experimental design
  • 12-1 p.m.: Instructional workshop on the ethics of power and sample size
  • 1-2 p.m.: Optional accessible walk and talk with faculty or break time
  • 2-5 p.m.: Understanding the hypothesis, linking study designs to statistical models and understanding types of study outcomes
Day 2: Power and Sample Size Analysis
  • 9 a.m.-12 p.m.: Choosing the appropriate statistical test, understanding correlation structure in multilevel and longitudinal studies, understanding power and Type I error and aligning power analysis and data analysis
  • 12-1 p.m.: Working lunch with faculty on class projects or optional break
  • 1-2 p.m.: Optional accessible walk and talk with faculty or break time
  • 2-5 p.m.: Summarizing a study design, predicting missing data and dropout and accounting for missing data and dropout
Day 3: Writing the Grant
  • 9 a.m.-12 p.m.: Collecting inputs for power analysis, studying power via statistical simulation, demonstrating recruitment feasibility and handling multiple study aims
  • 12- 1 p.m.: Working lunch with faculty on class projects or optional break
  • 1-2 p.m.: Optional accessible walk and talk with faculty or break time
  • 2-5 p.m.: Writing the sample size section for your grant, graphics for power and sample size, power for subgroup analysis and getting funded
Post-Class Assignments
  • Four homework assignments
  • One two-hour in-person or online lab to receive guidance on assignments 1-4
  • Final assignment
  • Consisting of a one-page statistical analysis plan and power or sample size analysis
  • Three hours of in-person or online lab to receive help on final assignment.
  • One-hour interactive video conference via GoToMeeting with students and workshop instructors