Here’s a research synopsis from the Science Daily website with the headline, “Running on little sleep? You’re twice as likely to get hurt.” The way this heading is phrased, it makes it sound like a causal claim. Phrases such as “twice as likely” don’t make it causal, but the “if…then…” nature of the headline does point to causality.

In the summary, the journalist warns readers with a causal-ish if…then…statement:

….if you haven’t slept well the night before, you could be putting yourself at greater risk of injury.

So in general, the author is saying that If you haven’t slept well, it leads you to be more likely to get injured. Let’s dive into the science behind this headline. We should have two reservations about it. First, let’s check on the source. Where was the original empirical article published?

a) Click on the Science Daily story and scroll to the end. You’ll find the reference for the original empirical article there. What is the journal name? (keep in mind the journal name is not the same as the article’s title).

b) You should be able to click on doi link in the reference and call up the original empirical article. Do that now!

Once you’re looking at the empirical article, you’ll see the journal name and, in the upper left corner, the initials MDPI. MDPI is a publisher–it publishes (and makes money from) over 400 different scientific journals.

c) Review the concept of a predatory journal in Chapter 2. What are some of the qualities of a predatory journal? You might mention fees, peer reviews, and article quality.

d) Using a search engine, ask if MDPI journals are considered predatory. What kinds of information do you get?

e) Many librarians trust a list called Clarivate’s Web of Science, which maintains a list of non-predatory, reputable journals. First, look for the journal name of a reputable journal on the list to make sure it works (I suggest searching for Psychological Science or Psychological Bulletin, both reputable journals in our field). Then, look for the journal name where the article on sleep and running appeared. Remember–you’ll want to plug in the journal name, not the authors or the article title. Is the journal on the Clarivate list?

In short, publication quality is the first potential problem with this sleep and running article. The second potential problem is its causal claim.

f) Review: What kind of study does it take to support a causal claim? (a correlational study or an experimental one?)

Here’s how Science Daily summarized the study. As you read, decide if the study was correlational or experimental:

In a survey of 425 recreational runners, the researchers discovered that participants who reported shorter sleep duration, lower sleep quality, or frequent sleep problems were almost twice as likely to experience an injury compared to those who slept well.

The sleep variable and the sports injury variable were both measured, and were measured at the same time in this survey. That makes the study correlational! Given those details, let’s work through the three criteria for causation.

g) The first criterion is covariance, which is about the study’s results. Does the study’s result show that runners with worse sleep (Variable A) have more injuries (Variable B)? That is, does the study show that A <–> B? (what part of the summary are you using to answer this question?)

h). The second criterion for causation is temporal precedence, which is about the study’s method. Does the study’s method make sure that sleep quality (A) was measured first in time, before injuries (B)? That is, does the method ensure that A –> B?
As you answer this question, think about whether it’s also plausible that injuries might have come first, and reduced people’s sleep quality (B –> A) .

i) The third criterion for causation is internal validity, which is also about the study’s method. Any correlational study is potentially susceptible to third variables (C variables) that might be responsible for the association. Here’s one example. It’s possible that a third variable (C) is Age; runners who are older (high C) might have worse sleep quality (low A) and odler runners may be more susceptible to injury (high B),
Now, come up with your own possible C variable, and explain how your C variable relates to both low A (low sleep quality) high B (high injury rate).

Thanks to Sue Franz for sharing this example!