Live 2020-03-02!!! Virus Models and p-hacking

TL;DR

The livestream discusses topics such as models for predicting the new coronavirus epidemic growth, P-hacking, and power analysis in statistics.

Transcript

a quest livestream is cool BAM hey everybody I'm Josh Dharma and welcome to a stat quest a live stream I hope everyone's doing okay I'm doing fine and check this out I'm wearing my static quest t-shirt double boom I love it um anyways let's just get started there's not a whole lot of introductory announcements we need to do today so we're just gonn... Read More

Key Insights

• ❓ Understanding and utilizing differential equations is crucial in modeling infection spread.
• 🥺 P-hacking can lead to false positives and misleading results in statistical analysis.
• ✊ Proper planning and conducting power analysis are important aspects of robust experimental design.

Questions & Answers

Q: What is the best model for predicting the new coronavirus epidemic growth?

The most commonly used models for infection spread are based on differential equations, taking into account factors like social networks and disease transmission characteristics.

Q: What is P-hacking?

P-hacking refers to the misuse and abuse of analysis techniques, specifically the cherry-picking of tests and the manipulation of data to achieve desired results.

Q: Why is it important to plan out tests in advance?

Planning tests in advance helps prevent false positives and ensures that proper statistical analysis is conducted, avoiding the pitfalls of P-hacking.

Q: How does power analysis play a role in statistical experiments?

Power analysis determines the sample size needed to correctly reject the null hypothesis, providing insights into experimental design and the probability of obtaining reliable results.

Summary & Key Takeaways

• The livestream begins with a thank you to viewers who used Super Chat and provides an update on upcoming topics.

• The first question addressed is about the best model for predicting the coronavirus epidemic growth, which leads to a discussion on differential equations and their use in infection spread modeling.

• The second question focuses on P-hacking, explaining the dangers of cherry-picking tests and the importance of proper statistical analysis.

• The livestream concludes with a discussion on power analysis and the significance of planning experiments in advance.