Identify the kind of research
Was the work a laboratory experiment, a computer model, an observational study or a trial? Each design can answer some questions better than others. A relationship between two measurements does not automatically show that one caused the other. Begin with the researchers’ actual question, not the most dramatic possible application.
Look at the size and scope
Find out what was studied and how many observations were included. A finding in one setting may not carry over to every population or environment. Check whether the result is preliminary and whether the paper has been reviewed. Peer review is a useful checkpoint, but it is not a guarantee that a result will never be revised.
Read the caveats as part of the story
Limitations are information. They can tell you whether a result is uncertain, how an effect was measured and what further work is needed. Look for independent specialists who can place the finding in context without overstating it. When possible, follow the link to the paper’s abstract and discussion. You do not need to understand every technical detail to notice the difference between a measured result and speculation about its future use.
Put these questions into practice: UK backs biotechnology projects for wildlife adaptation
Join the conversation