The Replication Crisis - Crash Course Statistics

The video discusses the importance of replication in scientific research, highlighting the reproducibility crisis where many studies fail to replicate. It explores causes such as data analysis variability and p-value misunderstandings. Solutions include more replication studies, sharing data, and reconsidering p-value thresholds. The power posing study is used as a case study to illustrate these issues.

Why is replication considered essential in scientific research?

• It ensures all studies are published.

• It reduces the cost of research.

• It allows for faster publication of studies.

• It helps confirm the validity of results.

What percentage of cancer treatment studies were successfully replicated by Amgen scientists?

• 11%

• 25%

• 50%

• 75%

What is one unethical practice contributing to the replication crisis?

• Intentional p-hacking

• Sharing data openly

• Publishing null results

• Using large sample sizes

What is a common misunderstanding about p-values?

• They measure the size of an effect.

• They indicate the importance of a result.

• They are the only factor in scientific conclusions.

• All of the above

Why might published studies overestimate effects?

• Because they are always replicated.

• Because of large sample sizes.

• Due to a bias towards significant results.

• As a result of thorough peer review.

What is one suggestion to improve the reproducibility of scientific research?

• Reduce sample sizes.

• Limit data sharing.

• Publish more null results.

• Increase the p-value threshold.

What change in p-value threshold has been suggested to improve reproducibility?

• From 0.05 to 0.005

• From 0.01 to 0.1

• From 0.1 to 0.01

• From 0.005 to 0.05

How can sharing data more widely help with reproducibility?

• It allows others to verify results.

• It reduces the need for replication.

• It ensures all studies are significant.

• It speeds up the publication process.

What role do academic journals play in addressing the reproducibility crisis?

• They adopt guidelines emphasizing transparency.

• They publish only significant results.

• They discourage data sharing.

• They focus on new discoveries only.

What is the impact of the reproducibility crisis on public trust in science?

• It is irrelevant to trust.

• It increases trust.

• It has no impact.

• It can undermine trust.