A real research case study on press releases, news coverage and the words that turn an association into a promise.
Study published 18 Nov 2019.
The COMPare project followed trial publications back to their plans. Use its findings to inspect an outcome change without jumping to accusations.
A historical audit of large trials shows why a literature search needs a results registry—and why an absent paper is not the same as an absent result.
Learn to separate adverse events from adverse reactions, count people rather than episodes and interpret zero events without claiming zero risk.
A combined outcome can tell a useful story while leaving its most important chapter unresolved. Work through a first-event example and learn what to look for in the component results.
A worked guide to precision, clinical importance and the difference between an uncertain result and evidence of little effect.
Use two fictional populations to see how sensitivity, specificity and the frequency of a condition shape what a positive result means.
Follow a simple time-to-event model to see why hazard ratios, risks at a fixed time and survival duration answer different questions.
A pooled estimate is the end of a chain of decisions. Trace the question, search, independent studies, bias and uncertainty before deciding what the headline deserves.
Find the missing denominator, connect results and ask better questions. Our evidence toolkit gives you a place to start.
Open the evidence toolkitOne question. A connected reading path. Institutional announcements and earlier coverage, preserved at their original addresses. The collection begins in 2010.
Browse the press release archive