When reviewing a couple of my past articles, I noticed a trend: in several cases I have lambasted media that reports wrongly or without nuance on science. My recent post about “poop pills” covers a story that’s probably a magnet for less-accurate reporting, given the subject. And when websites need to sell an archaeological discovery of a buried mule, they probably need some exaggeration to attract attention. The most flagrant was the one about the “discovery” of the so-called ninth planet. It was actually only a calculation of a possible planet (a good calculation presumably), but no planet was actually observed.
It is my opinion that such shoddy reporting, even though it may attract attention, is in the end detrimental to science: people risk getting disillusioned when they finally realize nothing was discovered or that the actual discovery was interesting but overblown. This disillusion might lead to disinterest or distrust of science – even though it probably isn’t even the fault of the scientist, but of the reporter!
It felt good when I encountered the following headline: “Paid family leave could reduce rates of abusive head trauma in infants” (my emphasis) from Medical News Today. That one word, “could,” is very important. The study on which this article reports found a correlation (not causation) between head trauma in infants and the possibility of taking paid leave by the parents.
The researchers, Dr. Joanne Klevens and colleagues from the US’s National Center for Injury Prevention and Control at the Centers for Disease and Prevention (extra points to Medical News Today for mentioning the researchers by name, including their institution) did an observational study, comparing states where paid family leave is available with states where it is not, in relation to head trauma admissions in hospitals. Furthermore, the rate in California was compared before and after the institution of said policy in 2004. Other potential factors (unemployment apparently) were taken into account, but even if the result was nice (5.1 fewer hospital admissions per 100,000 children less than 1 years old), it is still a correlation. A third or fourth variable, influencing both, may be at work.
And luckily, both the title and the conclusion show this very clearly. Words like “could” and “may” are used when appropriate, and show that, even though there are hints for a causal relation due to earlier studies, the conclusion is not certain.
Even though the language is a bit more cumbersome, it is, in my humble opinion, the best way to do science reporting. Not only is there no overhasty conclusion (“more research needed”) but more importantly it shows how science works: it shows the uncertainty linked to any scientific endeavor, and shows that any conclusion is only temporary, barring confirmation and future research. More like this, please!