When reporters write garbage science
How do we learn about health? Science? Medicine? Risk?
Most of us still learn from our schooling or from the news.
Even though traditional journalism has transformed ink to pixels, newspapers and television news get loads of readers and viewers.
I still read The New York Times which remains the agenda-setting publication of North America.
What that means is law-makers, other news outlets and citizens like us are influenced by what appears on the pages of The Times.
So when a headline claims that walking changes the brain, we should believe it, right?
Sometimes the news presents a shiny package and sometimes it’s just plain garbage.
Health reporter Gretchen Reynolds writes about peer-reviewed science studies for lay-folk in The New York Times.
This week a popular story (the most emailed piece today for The Times) announced that “Walking in nature changes the brain.”
Other media picked up on the story: The Huffington Post, The Atlantic and The Washington Post, for example.
And none of the reporters reveals the flaws in the study. And they should.
In fact, no one proved walking in nature changes the brain.
Reynolds reports on the study where researchers claimed that walking in the park is more soothing than walking near a highway.
But most of us already know this from personal experience, right?
Sometimes we just don’t need the empirical evidence to shore up our beliefs that really make common sense.
Wouldn’t you rather walk by a riverbank than a truck route?
But researchers nevertheless wanted to put their common sense to the test and see whether they could wring out evidence by studying people’s attitudes and brains.
They recruited 38 participants and divided them into two groups. All answered a battery of questions about their attitudes.
Individuals also had their heads examined—literally—with neuro-imaging that measures blood flow.
One group was assigned to walk along a busy road and another group walked in a tree-lined park.
You should know that a 19-member group (2 groups totaling 38) is very small by scientific standards.
Such small samples make it impossible to generalize your findings to the population that interests you.
But the numbers dropped even more.
Volunteers left and equipment failed, leaving just 31 volunteers.
It comes as no surprise that the nature group reported feeling more relaxed after the walk (than before) compared to the urban group.
But if you read the small print in the study, you’ll see that their result fail to meet scientific muster–the p value is .07, meaning, the result in this instance is statistically insignificant.
Reporters should have read the small type.
Next, researchers wondered whether walking would impact cognitive function, so researchers recorded blood flow in their brains.
They found significant decreases in blood flow when they compared the two walking groups.
But what does that mean?
I asked the chief of neurology at one of our local hospitals to help interpret the blood flow data.
“It’s a black box,” he said.
That means we don’t know the impact of less or more blood flow.
Some scientists say blood flow is a surrogate measure for other phenomena—such as engagement or arousal.
But we really don’t know.
“Decreased (or increased) blood flow doesn’t tell us what drives the neurological system.”
In this case, interpreting the findings isn’t straight-forward.
And because the groups are so small and the subjects young (average age was 26½ years), you can’t extrapolate the results to the larger population.
In other words, the findings aren’t robust.
And walking? Yes, it’s good for you.
But there’s no evidence that it changes your brain.
Note: the original study isn’t available free online. You can pay for a copy or you may be able to access the study by contacting your library. Here’s the citation:
Gregory N. Bratman, J. Paul Hamilton, Kevin S. Hahn, Gretchen C. Daily, and James J. Gross. Nature experience reduces rumination and subgenual prefrontal cortex activation. PNAS 2015 112 (28) 8567-8572; June 29, 2015
Image from the National Science Foundation website
Thank you for the post, I really enjoyed reading it!
On a different note, what should be the sample size for you to take the study seriously?
That’s a good question: To estimate a good sample size, you would need to know the population size of the folks you want to study. I don’t know what that is for this group. Usually you try for a size that will give you 95% confidence that your results can be generalized to the population–that’s why pollsters interview about 1200 people (in a randomly selected sample) to predict voting preferences. Good resource in Earl Babbie’s Social Science Research Methods
LikeLiked by 1 person