You have probably had this thought. You see a number for some vitamin, compare it to what you think you eat, and conclude you are falling short.
That number is almost certainly a Recommended Dietary Allowance, and the comparison you just made is not quite the one it was built for.
An RDA really is a target for an individual. The Food and Nutrition Board says so in those words. But it is deliberately set to cover nearly everybody, which means it sits above what almost everyone needs, including you. Landing under it on a given day tells you far less than it feels like it does.
Four values do most of the work in American nutrition guidance, and a fifth was added in 2019 that the government’s own main pages have not caught up with. Here is what each one is for.
The four values, and the question each one answers
The whole set is called the Dietary Reference Intakes, or DRIs. They are developed by the Food and Nutrition Board at the National Academies of Sciences, Engineering, and Medicine, jointly for the United States and Canada, and the NIH Office of Dietary Supplements publishes them for everyone else to use.
ODS gives one-line definitions, and the differences between them are the whole point.
The estimated average requirement, or EAR, is the average daily intake estimated to meet the requirements of 50 percent of healthy people. Half the group is covered at that level. Half is not. ODS says it is usually used to assess the nutrient intakes of groups, and to plan adequate diets for them, though it adds that the EAR can also be used to assess an individual’s intake.
The recommended dietary allowance, or RDA, is defined by ODS as the “Average daily level of intake sufficient to meet the nutrient requirements of nearly all (97–98%) healthy individuals,” and the same definition goes on to say it is often used to plan nutritionally adequate diets for individuals.
The adequate intake, or AI, is a level assumed to ensure nutritional adequacy. ODS says an AI is established when evidence is insufficient to develop an RDA. It is what the Board publishes when it cannot do the arithmetic an RDA requires.
The tolerable upper intake level, or UL, is the maximum daily intake unlikely to cause adverse health effects. It is not a target and not a spending limit. Article 323 covers it properly.
Notice what is missing from that list. None of these values describes you. Every one of them is a statement about a population.
One more number gets mistaken for these constantly. The percentage printed on a Supplement Facts panel is not an RDA. It is a percentage of a Daily Value, a separate figure FDA sets for labeling, with one number standing in for a whole table of age and sex groups. The gaps between the two are real and sometimes large, and article 324 covers how Daily Values are set.
The correction most people need: an RDA is a goal set above what you need
Two things get said about the RDA, and one of them is wrong.
The true one first. The RDA is meant as a personal target, and the National Academies says so explicitly. In its 2005 report on water, potassium, sodium, chloride and sulfate, the Academies puts it in a single sentence: “The RDA is intended to be used as a goal for daily intake by individuals.” Its 2000 report on applying the DRIs in dietary assessment says the same thing in slightly different words. Anyone telling you the RDA was never meant for individuals is contradicting the source.
The misunderstanding is in the other half. An RDA is not a minimum, and it is not a measurement of your body.
Look again at how it is constructed. It is set to meet the requirements of nearly all healthy people in a given life stage and sex group, 97 to 98 percent of them. Turn that around and you get the sentence nobody prints on a label: 97 to 98 percent of the people in that group have a real requirement lower than the RDA.
So does your body, almost certainly. Not definitely, but with something like 97 to 98 percent probability.
The construction is explicit, not an interpretation. The Academies’ 2000 report on applying the DRIs to dietary assessment sets out the arithmetic: the RDA is computed as the EAR plus two standard deviations of the requirement distribution. Two standard deviations above an average is, by design, a figure that overshoots nearly everyone. That is where the 97 to 98 percent figure comes from, and the report states it as a conclusion rather than leaving readers to infer it.
The same arithmetic explains a gap readers run into later. Because the RDA comes directly from the EAR, the report notes, insufficient data for an EAR means no RDA can be set at all.
That changes what a shortfall means. A day under the RDA is not a deficiency, and it is not evidence of one. It is a day below a figure set high enough that hardly anyone needs all of it. ODS publishes its own caveat alongside the tables, noting that individual requirements may run higher or lower than the DRI figures.
Iron shows how much the age and sex band matters. ODS gives an RDA of 18 milligrams a day for women aged 19 to 50, 8 milligrams for adult men, 8 milligrams again for women aged 51 and over, and 27 milligrams in pregnancy. Four different answers to the same question, printed on one table. Any single figure lifted out of it and handed to a general audience is wrong for most of that audience.
The same page adds a wrinkle no single number can carry. ODS reports that the iron requirement is 1.8 times higher for people following vegetarian diets, because the nonheme iron in plants is absorbed less readily than the heme iron in meat. The table figure assumes a mixed diet. Change the diet and the number underneath it moves.
The value built to judge whether intake is inadequate is the EAR, and it works on groups.
The EAR is where the alarming statistics come from
Nearly every headline announcing that some large share of Americans do not get enough of something is an EAR calculation.
The method is sound. Researchers estimate usual intake from survey recalls, then count what share of a population falls below the EAR for that nutrient. Because the EAR sits at the midpoint of requirements, that share approximates the share with inadequate intake. It is the right tool for the job, and it is the one the framework hands you.
Two published examples show what the output looks like in practice.
A 2025 analysis in the Journal of the Academy of Nutrition and Dietetics applied the method to 15,029 adults from NHANES 2007-2018. Among adults eating a carbohydrate-restricted pattern, the prevalence of intake below the EAR was 46.7 percent for vitamin A and 42.5 percent for vitamin C, against 33.6 percent and 23.3 percent in adults who met the macronutrient ranges. The same paper reported that 92.1 percent of the carbohydrate-restricted group exceeded the sodium Chronic Disease Risk Reduction Intake, which we will come to in a moment.
A 2020 paper in Nutrients ran similar numbers on 26,282 adults from NHANES 2005-2016 and reported inadequacy prevalence of 45 percent for vitamin A, 46 percent for vitamin C, 95 percent for vitamin D, 84 percent for vitamin E and 15 percent for zinc. It is worth knowing who produced it. All five authors were employed by a dietary supplement manufacturer, and the paper argues that supplements lower the share of people falling below the EAR. The arithmetic is not in dispute. Who commissioned the arithmetic is still useful context.
Neither paper is a deficiency survey. An intake below the EAR is an intake figure, not a blood test and not a diagnosis. Article 321 deals with what happens when people treat the two as the same thing.
When there is no RDA at all
Some nutrients never get an RDA, and the reason is specific.
An RDA is derived from an EAR. If the data are not good enough to estimate the average requirement, there is nothing to derive the RDA from, so the Board publishes an Adequate Intake instead, based on observed intakes in apparently healthy populations.
Vitamin K is the textbook case. ODS reports that insufficient data were available to establish an EAR, so the Board set AIs for every age group: 120 micrograms a day for men aged 19 and over, 90 micrograms for women.
Chromium is structurally the same and substantively stranger. The Board found the data insufficient for an EAR and set AIs of 35 micrograms a day for men aged 19 to 50 and 25 for women of the same ages, dropping to 30 and 20 micrograms at age 51. ODS also reports that chromium deficiency has not been reported in healthy populations and that no clinically defined chromium deficiency state exists. The number on the table is an observation about what people eat, not a requirement anyone has demonstrated.
Potassium is the AI with the biggest practical consequences, because the figures changed and the old ones are still in circulation. A National Academies committee revised the sodium and potassium DRIs in 2019. ODS reports current potassium AIs of 3,400 milligrams a day for men aged 19 and over and 2,600 for women, with 2,900 milligrams in pregnancy and 2,800 in lactation. The committee found the data insufficient to derive an EAR for potassium, and did not set an upper limit either.
The practical difference is this. An RDA carries a statistical claim about how much of the population it covers. An AI carries no such claim. It is the Board’s best observation-based estimate, labeled as one.
There are five DRI values now, and the main federal pages still list four
This is the newest piece of the framework and the least reported.
In 2019 the National Academies published revised DRIs for sodium and potassium and added a new category: the Chronic Disease Risk Reduction Intake, or CDRR. Every value before it answered a question about adequacy or about safety at the top end. The CDRR answers a third question. At what intake does the risk of chronic disease start to come down?
For sodium, the committee set one. In the Academies’ announcement of the report, the recommendation for people aged 14 and older is to reduce sodium intake if it is above 2,300 milligrams a day.
For potassium, it did not. ODS’s potassium fact sheet describes the expanded model, then reports that the committee found the evidence insufficient to derive a CDRR for potassium. Same report, same committee, two different answers, because the evidence underneath them was different. That asymmetry is worth holding onto. A missing CDRR is a statement about the evidence, not about the nutrient.
How the category came to exist is documented in the peer-reviewed literature. A 2019 paper in the American Journal of Clinical Nutrition from the Joint Canada-US Dietary Reference Intakes Working Group describes the two governments developing guiding principles for setting DRIs based on chronic disease risk reduction, to be applied for the first time in the sodium and potassium review.
Now the inconsistency, and it sits inside the source of record.
ODS’s main page on nutrient recommendations and its Dietary Reference Intakes landing page each define four values: EAR, RDA, AI and UL. Neither mentions the CDRR. ODS’s own potassium fact sheet explains it at length.
So the federal government’s most-read summary of this framework is one value out of date, while another page on the same site is current. This is not a scandal and not a disagreement about science. It is a reminder that reference pages age at different rates, and that the date on a page matters nearly as much as the agency name at the top of it. Researchers have already moved on: the 2025 carbohydrate-restriction analysis cited above reports its sodium results against the CDRR as a routine measure.
The upper limit is measuring something else entirely
One paragraph of framework, then a pointer, because article 323 owns this.
A UL is the maximum daily intake unlikely to cause adverse health effects, in the Office of Dietary Supplements’ phrasing. It is not a goal to aim at.
The clearest proof that an RDA and a UL measure different things is magnesium, and it looks like a mistake until you read the footnotes. ODS gives adult magnesium RDAs of 400 to 420 milligrams a day for men and 310 to 320 milligrams for women. It gives an adult UL of 350 milligrams. The ceiling sits below the target.
The explanation is in what each value counts. ODS says the magnesium ULs only include magnesium from dietary supplements and medications, and do not include magnesium found naturally in food and beverages. The RDA counts everything. ODS states the reason on the page: the RDAs include magnesium from all sources, meaning food, beverages, dietary supplements and medications.
Two numbers, two different accounting rules. Comparing them head to head produces nonsense, which is exactly what happens when people try.
Four questions to ask of any nutrition number
When you meet a figure in an article, on a panel, or in a conversation, these four sort out what it is.
- Which value is this? EAR, RDA, AI, UL, CDRR or Daily Value. They are not interchangeable, and the article you are reading may not say which one it means.
- Who is it for? Every DRI is published by age band and sex, with separate figures for pregnancy and lactation. The adult male and adult female figures differ for most nutrients. A number quoted without its band is incomplete.
- What does it count? Food only, supplements only, or everything. Magnesium’s UL counts supplements and medications. Magnesium’s RDA counts all sources. Same mineral, same page, different accounting.
- What year, and from whom? The potassium figures changed in 2019. The CDRR did not exist before then. Pages go stale at different speeds, and federal pages are not exempt.
There is a fifth question, and it is the one that decides anything. What applies to you, given your age, your health conditions, your medications and your pregnancy status if that is relevant? No table answers that. It is a clinician’s judgment, which is the main reason these values are published for clinicians and dietitians rather than as consumer advice.
That matters most in the places where the standard figures stop applying. Pregnancy and lactation have their own DRI columns for nearly every nutrient. So do infants, children and older adults, and the adult figures do not transfer to them. Reduced kidney function changes what is safe with several minerals, which is why the 2019 committee flagged potassium supplements for people with compromised kidney function even while declining to set a UL. Those are conversations with a clinician, and they are the specific conversations these tables were built to support.
One boundary deserves stating plainly, because it is where these values stop applying. The DRIs describe apparently healthy people. The 2000 National Academies report says directly that RDAs and AIs are not expected to replete individuals who are already malnourished, and are not intended to be adequate for those with increased requirements because of certain disease states. Someone recovering from a documented deficiency, or living with a condition that affects absorption, sits outside the range the tables were built to cover. That is the stated scope, not a loophole.
What the numbers are for
The DRIs were built to do two jobs: plan diets for groups, and assess whether groups are getting enough. The RDA is offered to individuals as a goal on top of that, with a margin deliberately built in.
Treat them as what they are and they are genuinely useful. A published figure, from a named body, for a named group, in a known year, with a stated rule about what it counts.
Treat them as a personal report card and they mislead in both directions. Under the RDA does not mean deficient, because your own requirement is very likely below it. At or above the RDA does not mean optimized either, because ODS says plainly that individual requirements run higher and lower than the published figures.
If one thing survives from all of this, make it the shape of the RDA. It was set high on purpose so that nearly everyone would be covered. That is why falling under it on a given day is weak evidence of anything, and why the EAR, not the RDA, is the number researchers reach for when they want to know whether intake is genuinely short.
Sources
- National Institutes of Health, Office of Dietary Supplements. “Nutrient Recommendations and Databases.” ods.od.nih.gov
- National Institutes of Health, Office of Dietary Supplements. “Dietary Reference Intakes.” ods.od.nih.gov
- National Academies of Sciences, Engineering, and Medicine. “Dietary Reference Intakes for Water, Potassium, Sodium, Chloride, and Sulfate, Chapter 1: Introduction to Dietary Reference Intakes.” nationalacademies.org
- National Academies of Sciences, Engineering, and Medicine. “Dietary Reference Intakes: Applications in Dietary Assessment, Chapter 1.” nationalacademies.org
- National Academies of Sciences, Engineering, and Medicine. “Sodium and Potassium Dietary Reference Intake Values Updated in New Report, Introduces New Category for Sodium Based on Chronic Disease Risk Reduction.” nationalacademies.org
- National Institutes of Health, Office of Dietary Supplements. “Potassium: Fact Sheet for Health Professionals.” ods.od.nih.gov
- National Institutes of Health, Office of Dietary Supplements. “Vitamin K: Fact Sheet for Health Professionals.” ods.od.nih.gov
- National Institutes of Health, Office of Dietary Supplements. “Chromium: Fact Sheet for Health Professionals.” ods.od.nih.gov
- National Institutes of Health, Office of Dietary Supplements. “Magnesium: Fact Sheet for Health Professionals.” ods.od.nih.gov
- National Institutes of Health, Office of Dietary Supplements. “Iron: Fact Sheet for Health Professionals.” ods.od.nih.gov
- U.S. Food and Drug Administration. “Daily Value on the Nutrition and Supplement Facts Labels.” fda.gov
- PubMed. “MacFarlane AJ et al. "A report of activities related to the Dietary Reference Intakes from the Joint Canada-US Dietary Reference Intakes Working Group." American Journal of Clinical Nutrition, 2019;109(2):251-259. PMID 30721931. DOI: 10.1093/ajcn/nqy293.” doi.org
- PubMed. “Ilayan A et al. "Micronutrient Inadequacy Differs by Intake of Fat Amount and Class Among Adults That Consume a Restricted Carbohydrate Diet: National Health and Nutrition Examination Survey, 2007-2018." Journal of the Academy of Nutrition and Dietetics, 2025;125(4):501-514. PMID 39521380. DOI: 10.1016/j.jand.2024.11.001.” doi.org
- PubMed. “Reider CA et al. "Inadequacy of Immune Health Nutrients: Intakes in US Adults, the 2005-2016 NHANES." Nutrients, 2020;12(6):1735. PMID 32531972. DOI: 10.3390/nu12061735.” doi.org