What You'll Learn
Learning Path
- Understanding Study Design
- Randomized controlled trials versus observational studies, why this matters for causation claims. Practice identifying study types from abstracts and assessing their limitations.
- Effect Size and Significance
- What p-values actually mean, why statistical significance doesn't equal importance. Calculate absolute versus relative risk to understand real-world impact of findings.
- Survey Methodology
- Sampling techniques, margin of error, confidence levels. Analyze polling data and identify methodological problems that invalidate results.
- Common Statistical Errors
- Base rate neglect, Simpson's paradox, regression to the mean. Work through examples of each in published journalism and learn to spot them in press releases.
- Data Visualization
- Recognize misleading charts and graphs. Understand how axis manipulation, selective data ranges, and poor scale choices distort information presentation.
- Interviewing Researchers
- Questions that reveal study limitations, how to get scientists to explain findings in concrete terms, recognizing when sources are overselling their results.
Program Details
How many times have you seen a headline claiming coffee both causes and prevents cancer? The problem isn't the science—it's journalists who don't understand relative risk, sample size, or the difference between correlation and causation.
This course teaches you to read scientific papers and statistical claims with the same critical eye you'd apply to a politician's speech. You'll learn which numbers actually matter in a study, how to spot p-hacking and other forms of data manipulation, and when confidence intervals mean a finding is basically worthless.
Real Numbers, Real Context
We work extensively with health reporting, where statistical misinterpretation can genuinely harm people. You'll analyze actual published articles that misrepresented research findings, then learn to calculate the real-world impact of reported effects. A 50% increase sounds dramatic until you realize it means going from 2 cases to 3 cases per 100,000 people.
The program also covers surveys and polls—margin of error, sampling methodology, question wording bias. You'll learn to recognize when a poll is designed to produce a predetermined result and how to explain statistical uncertainty to readers without hedging every statement into meaninglessness. Throughout, the focus stays on practical application: giving you the skills to report numbers accurately without needing a statistics degree.