Counter to previous findings, researchers surprisingly found that tackling (simple) mental tasks while padaling bike could help people improve cycling workout. As said in the podcast, "The findings could point toward new programs in which we get better workout simpling by using our heads."
The counterintuitive result made me both excited and a little confused. I looked up the related paper (http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0125470). The subjects of the experiment were HOAs (healthy older adults) and people with PD (Parkinson's disease), not general population. That's fair, since each experiment has its own goals on which participant prescreen is based. And it would be less realistic to achieve real subject diversity in such complicated human-involved experiment. However, popular science readers tend to have an assumption that "people" in the article means everybody on earth. Other examples could be "apple" to "fruit", "cycling" to "physical exercises", etc. This neglection of the possible limitation of scientific study could to some extent lead to the misunderstanding of the reported findings.
Another confusion is about the conditions in the experiment. It's mentioned in the podcast that "(the subjects) tackled these (mental) tasks once while sitting in a quiet room and again while on the bike. " The two conditions here are 1-Mental Tasks, and 2-(Mental Tasks + Cycling). However in the paper, "(the subjects) completed a baseline cycling task with no secondary tasks and then completed dual task cycling while performing 12 tasks from six cognitive domains representing a wide range of difficulty." The conditions mentioned in the paper are 1-Cycling, and 2-(Cycling + Mental Tasks). This is a mistake, which is different from the tendency of the specific-to-general misunderstanding mentioned above. But some people may argue that popular science readers just don't care about what conditions the experiment has. Well, who knows. Maybe they do, maybe they don't. It may also depend on what kind of mistake that has been made. However, it is somewhat similar to the typo in newspapers, which is hard to predict the effect, while it is fairly easy to avoid by paying a bit more attention.
People will get closer to those findings from popular science readings with the awareness of scientific limitations and the avoidance of low-level mistakes. However, we all know that there are still gaps between those findings and the "truth". Say, even in the paper, some of the critical points are not very clearly stated. According to the paper, every participant seems to have experienced Cycling first, then Cycling+Mental Tasks. The order of the conditions might influence the results: people may cycle faster even without other tasks during the second trial. What will the result be if some of them experience cycling+tasks first? (Counterbalance Unclearness)
We may even find the previous findings were totally wrong when we go ahead and explore more... Well, stop. that's too much... We've already known that we can only "make gradual 'approach' to the infinite truth". But what we can do is to understand, state, and make our process to these findings clearly.