How Science Works

Expert Tips and Best Practices

By Apoorva Mandavilli / 2 minute read

The following advice comes from the science journalists Emily Anthes, Caroline Chen, Laura Helmuth, Roxanne Khamsi, Ivan Oransky, and Ed Yong. 

• Don’t rely on press releases for story ideas; they are intended to promote, not to inform.
• If you do use press releases, never presume that everything in them is accurate or complete.
• Read the paper, not just the press release. You’ll be amazed how often the two bear no resemblance to each other in content or claims. 
• Be wary of claims that the findings represent a “revolution,” “breakthrough,” or “paradigm shift.” Science rarely works like that.
• Read the paper multiple times because, with every pass, you’re going to deepen your understanding and catch new things. 
• Read the authors’ description of funding sources and conflicts of interest. Ask them about their funding when you interview them.
• Google it! Have other studies been published on this topic? Have these scientists ever made bold or unwarranted claims to the press? Is this field controversial?
• Poke around in PubMed with keywords from the paper. The most recent similar study can help determine whether the new study is too incremental to cover.
• Know the difference between correlation and causation, and whether the study’s design supports its findings.
• Narrate your findings back to your sources by saying, for example, “Here’s how I’d summarize your study: …. Does this sound right to you?”
• Mine the references in the paper to find sources for comment or background reading. 
• For medical studies, know that the best trials have a control group, and that neither the scientists nor the participants know who received the drug and who received the control.
• Take a statistics class. Cultivate statisticians as sources you can turn to in a pinch.
• Do your own peer review. Build relationships with scientists who will give you quick takes on whether a paper is worth covering and what it adds to the field.
• No question is too small or dumb.