Print

Nuts and bolts of data analysis

 

Online course with panel: Members-in-Training ($49); Members ($79); Nonmembers ($109)

Day 1 | Jan. 19, 2022, 10am-noon EST
Day 2 | Jan. 20, 2022, 10am-noon EST

Audience:
Vision researchers and clinicians with some experience in basic data analysis using a statistical software package but without advanced training in statistics.  The target audience are those who plan to perform analyses with vision or eye data, or those who review and interpret output and results from eye and vision data analyses and want a deeper understanding of appropriate methods.

Course summary:
Have you ever wondered what to do with both eyes in your dataset?  Or how to obtain an average visual acuity or refractive error?  Data from vision research can stymie researchers trying to make the most out of the hard work of collecting it. This course will teach simple and accessible tools for understanding and appropriately analyzing data with a focus on data exploration, two sample comparisons and data presentation. We will discuss ways to include both eyes in studies/experiments and in analysis and show the harm when the correlation is not accommodated appropriately.  We will examine data from vision function tests to understand the statistical problems that arise and work through practical examples.  The course will provide an introduction to analytic tools and give researchers information on when more advanced methods are needed.  We will demonstrate statistical analysis approaches for evaluating the performance of screening/diagnostic ocular tests or machine learning techniques when evaluations are performed in both eyes. Lastly, we will show how data visualization methods and techniques can be leveraged to make research presentations more accessible and impactful. Following the course, attendees will understand the challenges and appropriate handling of vision research data, be able to choose appropriate statistical tools to make the most of the information, and be able to present the data and results in a meaningful and accessible way.

After this course, participants will be able to: 

  • Discuss the measurements most often used in vision research and how to analyze them appropriately.
  • Explain the problems that arise with improper handling of correlated ocular data. 
  • Perform simple statistical analysis for the various types of correlated ocular data including continuous data (normal or skewed distributed), binary data or ordinal data. 
  • Perform the simple statistical analysis for evaluating the performance of screening/diagnostic ocular tests using sensitivity, specificity, predictive values, and area under ROC curve.
  • Use graphical and statistical tools to understand vision data and chose the best vision metric for your question.
  • Use data visualization best practices to improve the impact and accessibility of data reports and presentations

A Certificate of Attendance will be available to attendees.

Speakers

Alison AbrahamAlison Abraham, PhD, MS, MHS
University of Colorado School of Medicine and School of Public Health

Xiangrong KongXiangrong Kong, PhD
Wilmer Eye Institute, Johns Hopkins University

Maureen MaguireMaureen Maguire, PhD, FARVO
School of Medicine, University of Pennsylvania

Bernard RosnerBernard Rosner, PhD
Harvard Medical School and School of Public Health

Jiangxia WangJiangxia Wang, MS, MA
Department of Biostatistics, Johns Hopkins University


Gui-shuang Ying, PhD
School of Medicine, University of Pennsylvania

 


Program Agenda

 

Day 1


Welcome: Course aims and introductions
Alison Abraham, PhD, MS, MHS

Review of commonly used measures in the study of eye disease and health and how to do simple arithmetic
Maureen Maguire, PhD, FARVO

The basics of correlated data - Continuous data
Bernard Rosner, PhD

Visualizing data and grabbing attention - How to broaden your audience and send them away without eye strain – Part 1
Jiangxia Wang, MS, MA

Day 2


Welcome to Day 2
Alison Abraham, PhD, MS, MHS

The basics of correlated data - Binary/categorical data
Xiangrong Kong, PhD

The basics of correlated data - Diagnostic data
Gui-shuang Ying, PhD

Visualizing data and grabbing attention - How to broaden your audience and send them away without eye strain – Part 2
Jiangxia Wang, MS, MA