Data Overload - Making Sense of Statistics in the News, Kristin Sainani

Explore the complexities of interpreting statistics in news media through this Stanford webinar led by Kristin Sainani. Delve into three key statistics lessons drawn from current events, examining concepts such as relative and absolute risk, risk ratios, and the importance of communicating these metrics accurately. Analyze the Niskanen Center's methodology and its application in simulations, while also addressing potential issues in this approach. Investigate the impact of correlated and uncorrelated polling errors, using the 2016 election as a case study. Critically evaluate the tendency to seek biological explanations for statistical findings, considering factors that influence vitamin D levels and VO2 max. Gain insights into the challenges of unmeasured and residual confounding in statistical analysis. Access additional resources and learn about Stanford's Medical Statistics Certificate Program to further enhance your understanding of data interpretation in various contexts.

Introduction.
Three statistics lessons from the news.
Statements made.
The numbers: relative risk (risk ratio).
The numbers: absolute risk difference.
Relative vs. Absolute Risk.
Communicating relative risks correctly.
Relative risks don't tell the whole story.
Niskanen Center Methodology.
Implement in a simulation.
But this approach has a problem!.
Correlated errors in the 2016 election.
Redo the simulation focused on polling errors, uncorrelated.
Then make the polling errors correlated.
Reaching for biological explanations....
Factors that affect vitamin D levels.
Factors that affect vo, max.
The problem of unmeasured and residual confounding.
Further resources.
Medical Statistics Certificate Program.