A one-hour seminar on null hypothesis significance testing and understanding the use and interpretation of p-values.
What the seminar offers
Hypothesis testing is the foundation of statistical inference procedures yet the meaning of the results (e.g., p-values and confidence intervals) is often not fully understood or appreciated by investigators, potentially leading to a misinterpretation of the results.
This seminar provides an understanding of the objectives and limitations of classical inferential testing, thereby supporting a more critical examination and interpretation of statistical analysis results. Specifically, this seminar:
- reviews the history and intent of inferential testing,
- discusses what Type I and II errors are, how to interpret results of a statistical test in light of these errors, and how they relate to the power of a statistical test,
- illustrates p-values and confidence intervals with visual simulations,
- explains concerns about multiple testing and how to address them, and
- provides guidance on how to report results of statistical tests.
Finally, the seminar examines current criticisms of the use and interpretations of results based on p-values and provides recommendations for addressing the limitations of the classic null hypothesis significance testing framework.
Who created it
This seminar was given by Sandra Taylor, PhD, Principal Statistician at the University of California-Davis’ Biostatistics, Epidemiology and Research Design Program.