Really? "Cutting out meat might help prevent snoring"

Does the study support the causal claim the journalist attached to it? Photo: Dusan Petkovic/Shutterstock

Guest post by Carli Fine, University of Delaware

A recent news article for the general public was headlined, “Cutting Out Meat Might Help Prevent
Snoring: Study”. The journalistic source was Health Day: News for Healthier Living, which makes recommendations for the general public and has news for medical professionals. The original empirical journal article  (open access) was entitled “Plant-based and vegetarian diets are associated with reduced obstructive sleep apnoea risk”.

This study offers a great opportunity to discuss how association claims differ from causal claims and how empirical journals differ from popular journalism.

In the empirical study (Melaku et al., 2024), researchers investigated the variables of adherence level to a general plant-based diet (PDI) and risk level for obstructive sleep apnoea (OSA). PDI was measured with a 24-hour recall method, as participants reported the foods that they had
consumed in the past day.  OSA was measured using a tool called the STOP-BANG (which combines data about people's Snoring, Tiredness, Observed snorting, Pressure (blood pressure), Body mass index (BMI), Age, Neck size, and Gender). The researchers found a negative relationship between PDI adherence and OSA risk (STOP-BANG scores), such that higher PDI adherence went with lower OSA risk.

a) The journalist's headline reads, "Cutting Out Meat Might Help Prevent
 Snoring". What's the verb in this headline? To what category of claim does this verb belong: Frequency, association, or cause?  (Use Table 3.4 below for a hint.)

Table 3.4

b) What were the study's variables?  

Name of the variable (Conceptual level)	What are the variable's possible levels?	Is this manipulated or measured?

c) How do you know this was a correlational study (and not an experiment)?

d)  Sketch a well-labeled graph of the result they describe. It’s okay if you don’t know the exact
data for each group. Before you start: Will your graph be a bar graph or a scatterplot? 

 

e) Does the study support a causal claim/headline? Name, and then apply, each of the three causal criteria to this situation. When you get to the third one, you should specify a plausible third variable that could explain the relationship.

Name of causal criterion	Application to this example

f) At the end of the popular article you'll see a “What This Means For You” box, copied below. Comment on this takeaway: Does the study support the claim in the box? 

g) Scholarly sources are sometimes referred to as primary source literature, while journalistic sources may be referred to as secondary source literature. Fill out the table below identifying who the authors are for each type of article. Which type of source should you use for a literature review?  You'll find Table 2.3 helpful. 

Type of Source	“Degree” of Source	Authors of Source	Intended Audience of Source	Examples of this Type of Source
Scholarly	Primary
Journalistic	Secondary

h) Journalistic sources tend to analyze, interpret, or summarize the content of primary source literature. Besides the headline, here are a few more examples of quotes that differ between the two articles. What are some of the differences in meaning? Why do you think that the journalist made these choices?

Journalistic Source	Scholarly Source
“People with diets highest in plant-based foods were 19% less likely to have sleep apnea compared to those with the lowest amounts of plant-based foods in their diet.”	“Participants in the fifth (highest) quintile of PDI had 19% lower odds (OR=0.81; 95% CI: 0.66–1.00) of sleep apnea risk compared to those in the first quintile (p-value for trend=0.008).”
“A plant-based diet worked more powerfully to lower sleep apnea risk in men”	“Overall PDI is primarily associated with reduced odds of OSA risk in males, not females.”

 

 

 

 

Suggested Answers:

a) The popular article title is “Cutting Out Meat Might Help Prevent Snoring: Study”. The verb is “prevent”, which is a causal claim verb. The empirical article title is “Plant-based and vegetarian diets are associated with reduced obstructive sleep apnoea risk”. The verb is “associated with” which is an association claim verb.

b)

Name of the variable (Conceptual level)	What are the possible levels of this variable?	Is this variable measured or manipulated?
Level of Adherence to a General/Healthy Plant-Based Diet (PDI)	Low to High	Measured
Level of Risk for Sleep Apnoea (OSA)	Low to High	Measured

c) This was a correlational study because all the variables were measured. In an experiment, at least one variable is manipulated.

d)

e)

Name of causal criterion	Application to this example
Covariance	The results show that level of adherence to a general/healthy plant-based diet was associated with level of sleep apnoea. Yes, covariance is achieved!
Temporal Precedence	This one is tricky! The researchers asked about both variables in the same questionnaire, at the same time. However, PDI was assessed based on foods consumed in the past 24 hours, which therefore took place before the STOP-BANG tool was administered. Yes, temporal precedence is achieved!
Internal Validity	This is a correlational study, so we can’t rule out any third variables that might be responsible for the underlying relationship. For example, people who adhere more to a plant-based diet may also choose to engage more in exercise and healthy behaviors that reduce risk for sleep apnoea. Alternatively, people who have a higher socioeconomic status (SES) might have better access to and means to afford fresh fruits and vegetables and also have better access to healthcare services leading to better overall sleep hygiene. No, internal validity Is not achieved!

f) The key takeaway uses the verb “reduce” which is a causal claim. The causal claim made by the popular article cannot be supported because the study did not manipulate a variable to establish internal validity. In fact, the researchers even write, “The cross-sectional nature of the study prevents us from inferring causality between dietary indices and OSA risk.”

g) 

Type of Source	“Degree” of Source	Authors of Source	Intended Audience of Source	Examples of this Type of Source
Scholarly	Primary	Researchers who conducted a scientific study	Other scientists, researchers, readers of this blog and students who have taken research methods!	European Respiratory Society, Nature, Science, Cell, Journal of the American Medical Association
Journalistic	Secondary	Journalists or reporters, typically employed by a news organization	General population	Health Day: News for Healthier Living, The New York Times, The Washington Post, BBC News, CNN

g) You should only use empirical/scholarly/primary sources for a literature review.

h) Journalists write for a general audience that might not have specialized knowledge in a particular field. They might oversimplify the research by avoiding technical jargon and statistical terms to make it less precise but more accessible. Journalists typically want their stories to be impactful for people’s daily lives. They might make causal interpretations of correlational studies to be more actionable.