Sample Size Matters: Misconceptions about Graphs and Statistical Analyses in Lab and Clinical Research

Summary

Could you have misconceptions about data visualization and statistical analysis that may affect the outcomes of your research? Misconceptions and problems with data visualization and statistical analysis are common in published studies. In this course, learners identify and correct misconceptions about data visualization and statistical analysis that are common in the basic biomedical sciences and other disciplines using small sample size studies.

Unsure if this course is for you? Take this quiz to see what misconceptions you may have.

Mayo's Sample Size Matters course is approximately ten hours of instruction. It is brought to you in an asynchronous format, allowing you to engage with the materials at your own pace and on your timeline. Registering for the course will give you access for one year.

Learning Objectives

• Recognize problems with current standard practices in basic biomedical science research.
• Identify solutions for problems with current standard practices in basic biomedical science research.

Mayo Clinic's Sample Size Matters online self-paced course can be completed in approximately ten hours. It is delivered through a series of units and scenarios outlined below.

Course Outline:

• Data Visualization
  • Beyond Bar Graphs
    • The Problem with Using Bar Graphs for Continuous Data
    • What to Use Instead
    • Resources for Creating More Informative Figures
  • Line Graphs
    • Interpretation and Limitations
    • Creating Interactive Line Graphs
    • Alternatives to Static Line Graphs
  • Understanding Variability and Summary Statistics
    • What are Summary Statistics?
    • What Summary Statistics Should We Report?
    • Standard Deviation and Standard Error
• Statistical Analysis
  • Challenges with Small Sample Sizes
    • Simulator Activity
  • Implications of Power and Effect Size
    • Power and Effect Size
    • Simulator Activity: Power
    • Simulator Activity: Effect Size
    • False Discovery Rate
    • Exploratory and Confirmatory Studies
  • Clustered Data
    • Types of Non-Independent Study Designs
    • Advanced Concepts
    • Presenting Clustered Data
    • Why Account for Non-Independence?
    • Strategies for Analyzing Clustered Data
    • Reporting
  • Reporting
    • Understanding the Problem
    • Solutions
  • Selective Reporting, Publication Bias, and Multiple Comparisons
    • What is Publication Bias?
    • Simulator Activity
    • What have We Learned
    • Selective Reporting
    • Solutions
• Synthesis: Three scenario exercises with quizzes

Sample Size Matters is an online course for graduate students and research professionals. Participants learn to apply the knowledge and skills gained to improve rigor, transparency, and reproducibility in their research.

	Stacey J. Winham, Ph.D. Associate Professor of Biostatistics Dr. Winham is a statistical geneticist interested in the genetic etiology of common, complex diseases. She develops statistical methods to identify genetic risk factors for diseases in high-dimensional data and applies those analysis methods to studies of psychiatric genetics, breast cancer and ovarian cancer.
	Tracey L. Weissgerber, Ph.D. Dr. Weissgerber is a former Mayo Clinic faculty member in the Division of Nephrology and Hypertension and a group leader and metaresearcher at the Berlin Institute of Health Quality, Ethics, Open Science, Translation (QUEST) Center at Charité — Universitätsmedizin.