Skip to main content
. 2014 Dec 10;349:g7015. doi: 10.1136/bmj.g7015

Summary of results for analyses of advice, primary claims from correlational data (causal claims), and human inference from non-human studies (human inference)

Variables No PR with news No with news Odds news uptake Odds ratio (95% CI) Odds news exaggerated Odds ratio (95% CI)
Advice:
PR not exaggerated 128 66 188 1.1 1.3 (0.8 to 2.4) 0.2 6.5 (3.5 to 12.4)
PRs exaggerated 85 50 172 1.4 1.4
Total 213 116 360
Causal claims:
PR not exaggerated 122 61 169 1 1.3 (0.6 to 3.0) 0.2 19.7 (7.6 to 51.4)
PRs exaggerated 60 34 92 1.3 4.3
Total 182 95 261
Human inference
PR not exaggerated 67 29 68 0.8 1.3 (0.7 to 2.5) 0.1 56.1 (14.9 to 211)
PRs exaggerated 38 19 47 1 5.9
Total 105 48 115

PR=press release.

The key results are that odds ratios for the dependence of news uptake on PR exaggeration are indistinguishable from 1, whereas odds ratios for the dependence of news exaggeration on PR exaggeration are much larger and clearly distinguished from 1. See the results section and figures 2 and 3 for further information, including percentages and 95% confidence intervals.