Reading a claim about your food

Eight glasses of water a day, and tea and coffee don't count. You have probably heard it, and you may have lived by it. In 2002 H. Valtin, in a review for a physiology journal, went looking for the evidence behind the advice to "drink at least eight 8-oz glasses of water a day". He searched the databases, took what he calls a cursory look at older literature, and consulted "several nutritionists who specialize in the field of thirst and drinking fluids". He reported: "No scientific studies were found in support of 8 x 8."¹ That's one thing a claim about food can turn out to be. Not a hoax: a sentence that travelled a long way from wherever it started, with nothing behind it when somebody finally looked. Others turn out to have a great deal behind them, and the method is how you tell which. You'll run it on that sentence first, then on caffeine and a headline, and finish with a decision only you can make.

The method, with this course's question in front

The institute's sort was introduced in Digital Literacy lesson 8 and sharpened in Using AI Effectively lesson 11, and you've used it at the end of most courses since. Three bins. Measurable: some result could count against the claim. Unmeasurable as stated: nothing in it could come out the other way. False: it was checked and did not hold.

Each course since has added one question for measurable claims. Here they are, with the place each one bit in this course.

Each example is sourced, with its read level, in the lesson named.¹⁶

Mental Fitness put its question last, as the final thing to ask of a measurable claim. Focus and Deep Work asked its instrument question before the sort. This course puts "instead of what?" first too, and the reason is lesson 1's: you can't change one thing. "Cut saturated fat" was never one claim. With polyunsaturated fat in its place it was one finding, with refined starch in its place another. Until you know what came in, you do not know which claim you're sorting. That ordering is this course's way of joining its question to the sort.¹⁶ It often splits one sentence into several, each with its own verdict. And the answer is often "not stated", which is itself a finding.

Then comes the grade lessons 6 to 8 used, for a measurable claim about a known swap: the institute's three kinds of claim. Established: settled across the experts who work on it, as industrial trans fat's harm is. Contested: an empirical question serious people argue over, as the low end of sodium is. A value question: one evidence can inform and not decide, as the meat dispute showed, where a value question sat on top of an empirical one. Many real claims are two of these at once. "Measurable" only means somebody could check. The grade is where you find out whether anybody has.

Eight glasses, run through the method

The claim as you'd meet it: "Drink at least eight glasses of water a day. Tea and coffee don't count."

Instead of what? Two swaps hide in it: water in place of other drinks, and water on top of whatever you already get, food included, since the claim never mentions food. Two claims already.

Sort. Measurable, and here's what would settle it: look at what healthy people drink and whether those drinking less run short. Valtin's review is that look, and this course read its abstract in full.¹ On the first half: no study supported it, and "surveys of food and fluid intake on thousands of adults of both genders, analyses of which have been published in peer-reviewed journals, strongly suggest that such large amounts are not needed because the surveyed persons were presumably healthy and certainly not overtly ill."

He fences the finding himself: "It is to be emphasized that the conclusion is limited to healthy adults in a temperate climate leading a largely sedentary existence, precisely the population and conditions that the "at least" in 8 x 8 refers to". And the other way, "large intakes of fluid, equal to and greater than 8 x 8, are advisable for the treatment or prevention of some diseases and certainly are called for under special circumstances, such as vigorous work and exercise, especially in hot climates."¹

On the second half he cites published studies showing that "caffeinated drinks (and, to a lesser extent, mild alcoholic beverages like beer in moderation) may indeed be counted toward the daily total".¹ This course read his abstract, not those studies.

That is lesson 5's Adequate Intake, the number used when nobody can measure a requirement. What does the day-to-day work, in the Academies' words, is "fluid intake, driven by the combination of thirst and the consumption of beverages at meals".² Europe's food safety authority, EFSA, printed 2.0 litres for women and 2.5 for men in 2010; this course read only the press release, not the opinion, so it can't say what those figures count.³ And the newest US guidelines, as of September 2026, give no volume at all: "Choose water (still or sparkling) and unsweetened beverages."⁴

The verdict, half by half, in this course's reading.¹⁶ "At least eight glasses" is measurable and unsupported: the review this course read found nothing behind it, and the surveys it cites point the other way, for healthy, mostly sedentary adults in a mild climate. It is not shown false for everyone, the difference lesson 10 drew between "no" and "can't tell", and Valtin says why: "it is difficult or impossible to prove a negative".¹ Heat, hard work and some illnesses change the answer. If a doctor has given you a fluid amount, more or less than usual, that number is about you, and this section isn't.

"Tea and coffee don't count" goes against both sources: the studies Valtin cites found caffeinated drinks "may indeed be counted", and the Academies' figure counts every beverage. That is the nearest thing to a false claim in this lesson, on the strength of an abstract and a summary chapter.

One caution the other way. The Academies set no upper limit for water, "as healthy persons can adapt to higher intakes from foods and beverages".² No limit is not a statement that any amount is safe, and this course read nothing on drinking too much, so it says nothing about it. The same goes for "detox" diets and drinks, the other claim people ask about most: this course read no source on them, so it says nothing about them either way.

Caffeine: one number, three framings

Caffeine shows something water doesn't: bodies can agree on a number and still mean different things by it.

EFSA's caffeine topic page summarises its 2015 opinion; this course read the page, not the opinion. "Intakes up to 400mg per day (about 5.7mg/kg bw per day) consumed throughout the day do not raise safety concerns for healthy adults in the general population, except pregnant women." The "mg/kg bw" is milligrams per kilogram of body weight. The page puts a 200 ml cup of filter coffee at 90 mg, so 400 mg is about four and a half cups (this course's arithmetic).⁵ The page gives a single-dose figure too: "Single doses of caffeine up to 200mg … from all sources do not raise safety concerns for the general healthy adult population". That is about one sitting, not the day.⁵ The US Food and Drug Administration also cites 400 mg, and adds the caution that travels less: "However, there is wide variation in both how sensitive people are to the effects of caffeine and how fast they eliminate it from the body."⁶

For pregnancy the same number, 200 mg a day, comes from three places in three framings.

• The NHS: "you should not have more than 200mg of caffeine a day", because "Having too much caffeine has been linked to a higher chance of miscarriage and stillbirth."⁷
• ACOG, the American College of Obstetricians and Gynecologists, in a 2010 opinion its page says was reaffirmed in 2026: "Moderate caffeine consumption (less than 200 mg per day) does not appear to be a major contributing factor in miscarriage or preterm birth."⁸
• EFSA: "Caffeine intakes from all sources up to 200mg per day consumed throughout the day do not raise safety concerns for the foetus."⁵

The NHS frames 200 mg as a ceiling with a risk above it, ACOG as a level below which harm is not apparent, EFSA as a level without safety concerns. None says harm begins at 200. ACOG adds: "A final conclusion cannot be made at this time as to whether there is a correlation between high caffeine intake and miscarriage."⁸ Its evidence is from 2010, and this course didn't read the newer studies that have argued for a lower figure. So, on the lesson's grades, "three bodies use 200 mg" is established and "harm begins above 200 mg" is contested: one number, two claims.

A headline, with every question at once

Most claims arrive as headlines about a new study. This one is invented for the lesson, of the kind lesson 8 took apart, and built on a real review:

"Ultra-processed food linked to anxiety in huge study of nearly 10 million people. Time to clear out the cupboards."

The review is Lane and colleagues' 2024 umbrella review in the BMJ, which lesson 8 read for its heart findings; this course read the abstract.⁹ It reports "higher risks of prevalent anxiety outcomes (odds ratio 1.48, 1.37 to 1.59; low)", among the findings its authors class as "convincing evidence (class I)". An odds ratio compares the odds of an outcome in two groups; 1 means no difference. The 1.37 to 1.59 is its 95 percent confidence interval, and "low" is the GRADE rating.

Now the other questions. Not every one bites on every claim, and noticing which do is part of the skill.

Who? The "nearly 10 million" is the whole umbrella, 9,888,373 people across all 45 pooled analyses.⁹ The abstract does not say how many were in the anxiety analyses. The headline has put the review's biggest number beside one of its findings.

Which instrument? Exposure was "defined by the Nova food classification system", and the abstract doesn't say how people's food was recorded.⁹ Lesson 2's self-report problem and lesson 8's specialists, who agreed on NOVA groups at a kappa of about a third, are this course's cautions, not findings about this review.

Which came first? This is not one of the six. It is Sleep lesson 3's direction problem, and it bites here. The abstract calls some outcomes "incident" and others "prevalent". In this course's gloss, incident means new cases arising while people were followed, and prevalent means cases people already had when studied. A prevalent finding can't say which came first: anxiety might change what people eat as readily as the reverse. The abstract doesn't discuss direction; that reading is the course's.

How big? Logic and Argument's question, not one of the six. An odds ratio, and the abstract gives only relative measures, so there is no absolute figure of the kind Logic and Argument lesson 9 taught you to want.

Where, and against what? Two of the six barely bite. Note-Taking's setting question has no single answer for a review pooling so many studies, and the abstract doesn't say where they ran. Mental Fitness's comparison-group question has nothing to grip: these are observational studies, so nobody was assigned to a comparison group. The people with lower exposure are just people who ate less of it, and who they were is the "instead of what?" you've already asked.

The grade. You've met this pair in lesson 8: "convincing" on the authors' scheme, "low" on GRADE. Across the whole review only four of the 45 analyses reached moderate on GRADE; the other 41 were low or very low.⁹

The advice. "Time to clear out the cupboards" is the welded second half, the advice joined to the finding by "time to", in Note-Taking's word. The authors conclude that their findings "provide a rationale to develop and evaluate the effectiveness of using population based and public health measures to target and reduce dietary exposure to ultra-processed foods for improved human health", and "support for urgent mechanistic research".⁹ So they do want exposure reduced, through public measures whose effectiveness is evaluated, and they ask for the causes to be studied. That's a different thing from telling one household to act on this finding.

Others do go further. Monteiro, arguing the "yes" side of the debate lesson 8 quoted, holds that "existing evidence is sufficient to recommend the avoidance of ultra-processed foods to optimize health".¹⁸ And the 2025–2030 US guidelines tell readers to "Avoid highly processed packaged, prepared, ready-to-eat, or other foods that are salty or sweet".⁴ Both rest that advice on more than this one finding, and neither tells you what goes in the cupboard instead.

Nothing in that headline is false, and the review is serious. Every problem sits in the distance between the review's sentences and the headline's.

What no study here can tell you

Lesson 4's Ornish dieters averaged 3.3 kg lost, and one standard deviation either side ran from a 4.0 kg gain to a 10.6 kg loss. Which of them would you be? This course can't say, and here's why, in three findings it has shown you.

The evidence is about groups. Lesson 4's diets had averages about a kilogram apart, 1.2 kg at most, while people inside one diet spread over many kilograms. No average says where you would fall.

The swap that matters is yours. Every finding here was about a particular replacement in particular people: polyunsaturated oil for animal fat, a potassium salt substitute for table salt, low-fat advice against a usual diet. What you would actually eat instead depends on your kitchen, budget and week, and no study tested that.

And much of the rest is value. Lesson 8's two meat bodies read overlapping evidence differently and also needed different premises about what matters, and the premises did work the data couldn't. When the evidence is low certainty and the effect small, whether to change is partly about what you care about.

So the course won't bless the way you eat, and it won't condemn it.

Who can, and what their titles mean

Some people can look at one person's diet, health and medicines together. In the UK and the US, the title that tells you most is "dietitian", and the one that tells you least is "nutritionist".

In the UK, the regulator, the Health and Care Professions Council, protects "Dietitian" and "Dietician": "To use one of these titles, an individual must be registered with the HCPC."¹⁰ The British Dietetic Association adds: "Please note the titles ‘nutritionist’, ‘nutritional therapist’ and ‘diet expert’ are NOT protected and can be used by anyone regardless of training and/or qualification."¹¹ The BDA is the dietitians' own body, so it has an interest in that distinction; the HCPC page is the neutral statement of the law. Its advice to the public: "check your health professional is registered with the HCPC or UK voluntary register for nutritionists."¹¹ A nutritionist may well be trained and on that register. The title alone won't tell you.

In the US, "registered dietitian" (RD) and "registered dietitian nutritionist" (RDN) are national credentials from the Commission on Dietetic Registration, and since 2024 new candidates need a graduate degree: "Effective January 1, 2024, the minimum degree requirement for eligibility to take the registration examination for dietitians changed from a bachelor's degree to a graduate degree." In the commission's words, "all registered dietitians are nutritionists but not all nutritionists are registered dietitians."¹² Whether "nutritionist" is regulated where you live depends on your state. This course saw that only in a search result and read no state's rules, so check with your state's licensing board. Outside the UK and the US it read nothing on titles.

So, in this course's summary, two checks: the HCPC register in the UK; the RD or RDN credential and the state licence in the US. Neither title guarantees good advice. The protected UK title does guarantee training, a code of conduct and a regulator you can complain to; the US credential guarantees the training and the commission's standards, and whether there is also a state regulator depends on the state.¹⁶ And for anything on the callout's list, a doctor is the other person to ask, often first.

When it stops being a self-help question

The callout is the list. Here are reasons behind some of its items, and a few it does not name, from official pages this course read.

Losing weight without trying. The NHS: "Unintentional weight loss is when you lose weight without changing your diet or exercise routine. It can be a sign of stress or a serious illness. See a GP just in case." And: "It's important to get it checked if you're losing weight without trying."¹⁴ Here weight loss is itself the warning.

A medicine that meets food. The NHS on warfarin: "Foods high in vitamin K, such as broccoli, spinach and other green leafy vegetables, can affect how warfarin works. You can still eat these foods, but talk to your warfarin clinic for advice." And: "Talk to your doctor before making any big changes to your diet when taking warfarin."¹⁵ So "eat more leafy greens", ordinary advice for most people, is a change to talk over first for someone on warfarin, though the claim is the same.

Pregnancy. Beyond caffeine, the NHS says "Get help from 111 if you're pregnant and" you "feel unwell after having food or drink that should be avoided in pregnancy". (NHS 111 is the UK's non-emergency medical line; elsewhere, your midwife, obstetrician or doctor.) It is reassuring about the ordinary slip: "If you've already had a small amount of food or drink that should be avoided during pregnancy, it's unlikely to affect your baby if you did not get ill."⁷

Children. The FDA reports that "Medical experts advise against energy drinks for children and teens because of the levels of sugar and caffeine, according to the American Academy of Pediatrics."⁶ Beyond that this course read no guidance on children's diets, which is why the callout sends those decisions to a doctor or dietitian.

Diabetes and kidney disease. For diabetes, the American Diabetes Association says that on a very low carbohydrate plan "consultation with a knowledgeable practitioner at the onset is necessary to prevent dehydration and reduce insulin and hypoglycemic medications to prevent hypoglycemia" (lesson 9).¹⁷ For kidney disease, the kidney guideline advises against high protein (lesson 5), and potassium salt substitutes carry a caution (lesson 7). This course found no official page on seeing a dietitian for either, and says so.

Food, eating or weight that feels out of control. The NHS: "An eating disorder is a mental health condition where you use the control of food to cope with feelings and other situations." Among its signs are "spending a lot of time worrying about your weight and body shape" and "having very strict habits or routines around food". Its advice: "If you think you may have an eating disorder, see a GP as soon as you can." And: "With treatment, most people can recover from an eating disorder." For someone else: "Let them know you're worried about them and encourage them to see a GP. You could offer to go along with them."¹³ The helplines are in the callout. Beat's number there is its England line, and its helplines page lists the numbers for Scotland, Wales and Northern Ireland.¹⁹ ANAD's is peer support, not treatment.

One point belongs to this course's own method. Reading claims closely can turn into checking every label and every meal. If rules about food have started to take up more of your day than you'd like, that is close to the NHS's "very strict habits or routines around food", and it's the callout's question, not this course's.

Deciding

The course project, One claim you acted on, ends in one of three decisions about one claim: change something, keep something deliberately, or stop acting on the claim. If the callout's list applies to you and touches the claim, the decision is "ask first", and the project scores that in full.

How does the method bear on the choice? In this course's reading:¹⁶

• An established finding about a swap you would actually make can carry a change, weighed against what it costs you.
• A contested finding can't settle it, so the decision rests on a premise of yours. Lesson 8 reconstructed one for each meat body: for NutriRECS, a very small, uncertain benefit doesn't justify asking people to give up what they value; for WCRF, when a food is a cause of serious disease and no safe level has been shown, eat little or none of it, because small risks add up across many people. Write yours down.
• A value question is yours; evidence only informs it.
• A swap that's "not stated" means the claim can't bear a change by itself. That does not show your present way of eating is fine either, so the decision rests on what else you found (the guides the project asks you to read) or on a premise you write down.

A worked case. Sam, invented for this lesson, has made himself drink eight large glasses of water a day for years, not counting his three mugs of tea. He's healthy, works at a desk, lives somewhere mild, and finds the water a chore. The section above has already sorted his claim.

Sam's case was easy: one-sided evidence and low stakes. Most won't be, and the project is where you take one through all seven steps.

What people get wrong

"A nutritionist is a qualified professional." Some are. But in the UK the title is open to "anyone regardless of training and/or qualification", in the BDA's words, and in the US it depends on the state.¹¹ "Dietitian" in the UK, and RD or RDN in the US, are the titles that tell you something.

"If the evidence is uncertain, nothing is known." Almost every dispute in this course sat on agreed ground. Salt raises blood pressure; the fight is over the low end. Saturated fat raises LDL; the fight is over hard outcomes at ordinary intakes. Contested usually means narrower than the headline, not blank.

"Measurable means it's been measured." The water rule was measurable all along, and when somebody looked, no study supported it. That is why the grade is a separate step.

"The course should have told me what to eat." It would have had to know your swaps, your health, your medicines and what you value, and pretend that group averages describe you, the error it spent eleven lessons teaching you to catch.

Practice

Connections

Back. This lesson drew on most of the ten before it: lesson 1's swap and designs, lesson 2's instrument, lesson 4's diets and spread, lesson 5's Adequate Intake, which the water figures turned on, lessons 6 to 8's grades and value premise, lesson 9's guides, which the project asks you to read, and lesson 10's difference between "no" and "can't tell".

Forward. Strength and Fitness, next on the Core, is where training belongs. Cooking Fundamentals, later this term, is where the swap you write down turns into something on a plate. Statistics for Citizens and Medical Literacy, planned later on the Core and not yet written, will want the comparison and swap questions again.

Go deeper

• Valtin, 2002, American Journal of Physiology. Abstract read in full here. A model of hunting a piece of advice to its origin and stating the limits of what you found.
• Dietary Reference Intakes for Water, free online from the National Academies. The summary chapter and the opening of the water chapter were read here. The plainest official text on why an Adequate Intake isn't a requirement.
• HCPC, professions and protected titles, read here. The legal list, and where any UK check on a title starts.
• NHS, eating disorders overview, read here. Plain signs, where to go, and what to say to someone you are worried about.

Sources

1. H. Valtin, "'Drink at least eight glasses of water a day.' Really? Is there scientific evidence for '8 x 8'?", American Journal of Physiology: Regulatory, Integrative and Comparative Physiology 283(5), 2002, pp. R993 to R1004, doi 10.1152/ajpregu.00365.2002. Read: the abstract in full.
2. Institute of Medicine, Dietary Reference Intakes for Water, Potassium, Sodium, Chloride, and Sulfate, National Academies Press, 2005 (released 2004), record 10925. Read: the summary chapter and the opening sections of Chapter 4, "Water". The share from food is this course's arithmetic.
3. European Food Safety Authority, press release, 26 March 2010. Read: the release only; the 2010 opinion on water was not read.
4. US Departments of Health and Human Services and of Agriculture, Dietary Guidelines for Americans, 2025–2030, January 2026, PDF. Read: the whole document. The hydration line is from the January 2026 edition, current as of September 2026.
5. European Food Safety Authority, caffeine topic page, summarising the 2015 opinion (EFSA Journal 13(5), 4102). Read: the topic page only; the opinion was not read. The four and a half cups is this course's arithmetic.
6. US Food and Drug Administration, "Spilling the Beans: How Much Caffeine is Too Much?", current as of 28 August 2024. Read: the page.
7. NHS, "Foods to avoid in pregnancy", last reviewed 15 June 2026. Read: the page.
8. American College of Obstetricians and Gynecologists, Committee Opinion No. 462, "Moderate Caffeine Consumption During Pregnancy", 2010, reaffirmed 2026 per the page. Read: the page and its abstract. Newer studies on caffeine in pregnancy were not read.
9. M. M. Lane and colleagues, "Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses", BMJ 384, 2024, e077310, doi 10.1136/bmj-2023-077310. Read: the abstract. The glosses of odds ratio, incident and prevalent, and the reading of direction, are this course's. The headline is invented.
10. Health and Care Professions Council, "Professions and protected titles". Read: the page.
11. British Dietetic Association, "What is a dietitian?" and "Journalist FAQs". Read: both pages.
12. Commission on Dietetic Registration, "2024 Graduate Degree Requirement" and "RDN credential FAQ". Read: both pages. US state rules on "nutritionist" were seen only in a search result and not read.
13. NHS, "Eating disorders: overview", last reviewed 23 January 2024. Read: the page.
14. NHS, "Unintentional weight loss", last reviewed 28 July 2025. Read: the page.
15. NHS, "Warfarin", last reviewed 24 February 2026. Read: the page.
16. This course's own constructions, labelled where they appear: putting "instead of what?" before the sort; the table's examples, each sourced with its read level in lessons 1, 2, 4 and 8; the verdicts on the corn oil, water and headline examples; the summary of what the two titles guarantee; the mapping of grades onto the project's decisions; and Sam, who is invented.
17. A. B. Evert and colleagues, "Nutrition Therapy for Adults With Diabetes or Prediabetes: A Consensus Report", Diabetes Care 42(5), 2019, pp. 731 to 754, doi 10.2337/dci19-0014. Read: the full-text sections on eating patterns, macronutrients and carbohydrate.
18. C. A. Monteiro and A. Astrup, "Does the concept of "ultra-processed foods" help inform dietary guidelines, beyond conventional classification systems? YES", American Journal of Clinical Nutrition 116(6), 2022, doi 10.1093/ajcn/nqac122. Read: the abstract.
19. Beat, "Helplines". Read: the page, checked 24 September 2026. Hours change, so this lesson doesn't print them.