The Practical Alternative to the p Value Is the Correctly Used p Value

Area of Science:

• Statistical inference
• Scientific methodology
• Research integrity

Background:

• Overreliance on p-values in scientific literature is a significant concern.
• Researchers often misinterpret p-values, impacting the quality of scientific research.
• Current discussions often focus on replacing p-values rather than improving statistical questioning.

Purpose of the Study:

• To advocate for a shift from solely focusing on p-values to improving the questions researchers ask.
• To highlight the "statistician's fallacy" in proposing alternatives to p-values.
• To emphasize the importance of understanding what researchers truly want to know.

Main Methods:

• Critique of the overreliance on p-values in scientific research.
• Discussion of the "statistician's fallacy" in statistical practice.
• Proposal for improved statistical education and user-centered software.

Main Results:

• The current focus on p-values and their alternatives distracts from fundamental issues in statistical questioning.
• Minimum-effect tests and equivalence tests can improve the questions researchers ask.
• Better statistical education and software are crucial for preventing p-value misinterpretation.

Conclusions:

• Improving statistical questions is a prerequisite for enhancing statistical inferences.
• Statisticians should educate researchers on the questions they can ask, not dictate what they want to know.
• Prioritizing evidence-based education and user-centered software can mitigate p-value misinterpretation and elevate research quality.