Chapter 12: Reading Nutrition News and Studies Without Getting Fooled
A diet claim is only as good as the study behind it, and four questions (who was studied, what was compared, for how long, and what was measured) catch most bad claims.
On this page
- The short answer
- The four questions to ask about any claim
- Study types, from weakest to strongest for diet questions
- What a headline leaves out
- Who is selling, and what "do your own research" means
- How to find the study yourself
- Worked examples from this guide
- The drill-downs in this chapter
- Related
The short answer
Most diet news comes from real studies. The problem is what gets dropped on the way to the headline: the number of people, whether they were people at all, and whether anything was actually tested.
Before you change what you eat because of a news story, find the study and read its abstract. An abstract is the short summary at the top of a paper, and it is free to read on PubMed for almost every diet study.
KetoDick on this: his channel is built on this habit. He tells viewers to look studies up on Google Scholar and PubMed and to check that a study states its protocol (video 1JbK1_ClDmY), and he says he shares his sources so viewers and their doctors can judge them (video UygKGoP_BtA).
The four questions to ask about any claim
Ask these in order. If a story cannot answer them, it has not told you enough to act on.
- Who was studied? People or mice? How many? Healthy adults, people with diabetes, older men only?
- What was compared with what? A diet against another diet, against no change, or nothing at all?
- For how long? Days, months or years? Weight results often change between 6 and 12 months.
- What was measured? A real outcome (weight, heart attacks, death) or a lab marker that may or may not lead to one?
These four questions are the core of How to Read a Nutrition Study Abstract, which walks through two real abstracts step by step.
Study types, from weakest to strongest for diet questions
Not all evidence is equal. This table ranks the common types for answering "will this diet change what happens to me?"
| Study type | What it does | What it can show | Main weakness |
|---|---|---|---|
| Cell or animal study | Tests a substance on cells or mice | A possible mechanism | People are not mice |
| Observational (cohort) study | Follows people and records what they eat | That two things move together | Cannot prove one causes the other |
| Randomized controlled trial (RCT) | Assigns people to diets by chance | That the diet caused the difference | Often short, small or hard to stick to |
| Metabolic ward study | Feeds people every meal under supervision | Exact effects of food on the body | Very small, very short |
| Meta-analysis or review | Pools many studies | The overall pattern | Only as good as the studies pooled |
The full comparison, with examples where the two main types disagreed, is in Observational Studies vs Randomized Trials.
What a headline leaves out
Headlines use words like "linked to," "boosts" and "slashes risk." "Linked to" almost always means an observational study found an association, not that one thing caused the other.
Three things headlines often skip:
- The species. Some of the most-cited ketone papers were done in mice. Youm et al., 2015 found that the ketone BHB blocked an inflammation pathway in mouse models and in human immune cells (monocytes), not in people on a diet.
- The size of the effect. "Lower risk" can mean a small change. A hazard ratio of 0.92 means 8% lower relative risk, which says nothing about how many people that is out of 100. See relative vs absolute risk.
- The outcome. A change in a blood marker is not the same as a change in heart attacks or lifespan.
Each of these is covered in Diet Headlines: What "Linked To" Really Means.
Who is selling, and what "do your own research" means
The question "who benefits if you believe this?" is part of reading a study. A paper's funding and conflict-of-interest sections tell you who paid for it. A product page or social post usually tells you nothing.
Common red flags:
- A claim backed only by testimonials or before-and-after photos
- A product sold through people who also recruit sellers (multi-level marketing)
- "Do your own research" offered in place of a source
- A forum or group where every question gets the same answer
KetoDick on this: he calls "do your own research" a red flag and says whoever makes a claim should share their sources (video UygKGoP_BtA). He says he does not recommend exogenous ketones, pointing to their price and multi-level-marketing selling (video cTGYIFOTPl8), and warns that a keto-only forum will answer every question with keto, which he calls confirmation bias (video 1m1hPgaRIg4).
More on each of these in Red Flags in Diet Advice.
How to find the study yourself
You do not need a science degree or a paid account to check a claim.
- Pull the key words from the story: the food or diet, the outcome and the year.
- Search them on PubMed or Google Scholar.
- Match the result to the story: same authors, same journal, same year.
- Read the abstract and answer the four questions above.
- If the story gives no study at all, treat the claim as unverified.
Step-by-step search tips are in how to search PubMed for diet studies.
Worked examples from this guide
Every evidence claim in this guide links to its source, so each chapter doubles as practice. These examples show the four questions at work.
| Study | Who and how long | What it shows | What it does not show |
|---|---|---|---|
| Shai et al., 2008 (DIRECT) | 322 adults, 2 years, randomized | Mean loss 2.9 kg low-fat, 4.4 kg Mediterranean, 4.7 kg low-carb | That every person loses that much |
| Gardner et al., 2018 (DIETFITS) | 609 adults, 12 months, randomized | Healthy low-fat −5.3 kg vs healthy low-carb −6.0 kg, not significantly different | Anything about a strict ketogenic diet; its low-carb arm was moderate |
| Hall et al., 2016 | 17 men, metabolic ward | Energy expenditure rose 57 kcal/day after a switch to ketogenic; body-fat loss slowed | Long-term results in women |
| Mursu et al., 2011 | 38,772 older women, observational | Multivitamin use associated with slightly higher total mortality (HR 1.06) | That multivitamins caused it |
| O'Donnell et al., 2014 (PURE) | 101,945 people in 17 countries, observational | Sodium of 3–6 g/day associated with fewer deaths and cardiovascular events than higher or lower intakes | Exact sodium intake; it was estimated from one morning urine sample |
KetoDick on this: on DIETFITS, he questions how the article he was rebutting used it, pointing to self-reported intake and wide differences between individuals (video 1JbK1_ClDmY). The published result is the one in the table: the two healthy diets did not differ significantly at 12 months. Both readings are in the DIETFITS explainer.
Where his stated position and the literature differ, this guide says both. One example: he says cholesterol markers may spike early on keto and improve over 6 to 12 months (video UygKGoP_BtA). A National Lipid Association statement reports that low- and very-low-carb diets lower triglycerides, raise HDL-C and have a variable effect on LDL-C. See cholesterol on keto.
When to bring a study to your doctor
A study is a reason to ask a question, not a reason to change a prescription. If you take diabetes or blood pressure medication, have kidney disease, are pregnant, or have a history of an eating disorder, bring the study to your doctor before acting on it.
KetoDick on this: he tells viewers to check blood work, talk to a doctor and follow up on symptoms rather than asking a forum (video 1m1hPgaRIg4). See Chapter 11: risks.
The drill-downs in this chapter
| Page | What you will learn |
|---|---|
| How to Read a Nutrition Study Abstract | Pull population, diet, comparison, outcome, size and length out of any abstract |
| Observational Studies vs Randomized Trials | Why cohort studies and trials can disagree, using multivitamins and fiber |
| Diet Headlines: What "Linked To" Really Means | Association vs cause, mice vs people, and lab markers vs real outcomes |
| Red Flags in Diet Advice | Sponsors, ketone MLMs, testimonials and forum echo chambers |
Related
Guide: The KetoDick Guide · Risks · Cholesterol on keto
Blog: How to search PubMed for diet studies · "Do your own research" is a red flag · Relative vs absolute risk · DIETFITS explained · Kevin Hall's metabolic ward studies · The PURE sodium study explained · Keto forums and confirmation bias · The Men's Health keto article · Is keto dangerous?
Videos: Men's Health hack job on keto · Keto basics for beginners · Keto is dangerous and warnings from friends · Beginner keto questions and tips · Common beginner questions
Previous: Chapter 11: Risks · Next: Back to the guide contents
Last verified: 2026-10-31