Dietary assessment: 24-hour recall, food diary and food frequency questionnaire
Before judging a diet you have to measure it, and every way of asking people what they eat makes its own kind of error.
Key facts
- 24-hour recall: a structured interview about all food and drink from the previous day, usually midnight to midnight; it takes about 20 to 60 minutes.
- Food record (diary): the person writes down food as it is eaten, with portions estimated from pictures or models, or weighed; it does not rely on memory.
- Food frequency questionnaire: a fixed list of foods, usually 80 to 120 items, asking how often each was eaten over the past month or year.
- One day is not usual intake: day-to-day variation is large, so at least two non-consecutive recalls or record days are needed to estimate the spread of usual intake.
- Energy is commonly under-reported, most consistently by people with a higher body mass index; sweets and desserts are under-reported more than fruit and vegetables.
The three main tools side by side
| Feature | 24-hour recall | Food record (diary) | Food frequency questionnaire |
|---|---|---|---|
| Time covered | Yesterday | The days recorded, as they happen | The past month or year |
| Depends on | Specific memory and an interviewer or automated tool | Writing everything down at the time | Generic memory and averaging |
| Main error | Random (day-to-day variation) | Reactivity; quality falls over days | Systematic, from a fixed food list |
| Best use | Mean intake of a group; several non-consecutive days for usual intake | Detailed short-term intake in motivated people | Ranking people by long-term intake; case-control studies |
| Weak point | One recall cannot describe a person's usual diet | Changes the diet it measures | Poor for absolute intake; list may not fit the local diet |
In detail
- 1. Random error (day-to-day variation) makes one measurement imprecise but not biased; averaging repeated days or statistical modelling reduces it.
- 2. Systematic error pushes every measurement the same way and cannot be removed by repeating the tool; it needs an unbiased reference measure.
- 3. The flattened slope: people who truly eat more tend to under-report and those who eat less tend to over-report, so differences between people look smaller than they are.
- 4. Reactivity: keeping a diary changes behaviour (people eat less or simpler food), and quality drops as the days add up; an unannounced recall avoids reactivity.
- 5. A multiple-pass recall walks through the day several times and probes for forgotten items, preparation and portion size, so fewer foods are missed.
- 6. Doubly labelled water measures energy expenditure in weight-stable people and is the most accurate way to detect energy misreporting, but it is costly and covers only about two weeks.
- 7. Judging the result: usual intake near or above the RDA makes inadequacy unlikely; below the EAR it is more likely than not; above the UL there is risk of harm (see balanced diet).
- 8. Food-based guidelines turn this into advice: in Colombia, ICBF's guide for people older than 2 years uses the «Plato saludable de la Familia Colombiana» and messages such as fresh varied foods, dairy and egg daily and legumes at least twice a week.
Clinical notes
- A diet history is one part of nutritional assessment: combine it with weight, height, BMI and waist (anthropometry), laboratory markers and examination for signs of deficiency.
- In children, recalls are given by the caregiver; growth charts and z-scores show whether intake is enough over time.
- Under-reporting is expected in people with obesity, so a «low» reported energy intake does not explain away weight gain.
- Ask about supplements as well as food: total nutrient intake needs both.
Common traps
- A single 24-hour recall shows a person's usual diet: it shows one day; a group mean can come from one recall, but an individual's usual intake needs several non-consecutive days.
- A food diary is the gold standard because it does not rely on memory: writing food down changes what people eat (reactivity), and records become less complete over the days.
- A food frequency questionnaire gives accurate absolute intakes: it carries systematic error and is best for ranking people from low to high intake, not for exact amounts.
- Reported energy intake below expected requirements proves the person eats little: energy under-reporting is common, especially at higher body mass index; only biomarkers such as doubly labelled water measure expenditure.
- Intake below the RDA means deficiency: the RDA covers 97–98 % of healthy people, so many people need less; intake below the EAR makes inadequacy more likely.
Test yourself
Think of the answer, then tap the question to see it.
Why are at least two non-consecutive 24-hour recalls needed to estimate the distribution of usual intake in a group?
Intake varies a lot from day to day; repeated days let that within-person variation be separated from true differences between people.
Which tool is most suitable for asking cases and controls about their diet years before diagnosis?
A food frequency questionnaire (or a screener): it asks about long-term habitual intake and is the only practical option for retrospective case-control studies.
A woman keeps a 7-day food diary and reports eating much less than usual. What error explains this?
Reactivity: the act of recording changes eating behaviour, and record quality also tends to fall as the days go on.
Why can repeating a food frequency questionnaire not remove its main error?
Its main error is systematic (the same direction every time, from a fixed food list and averaging); only random error shrinks with repetition.
Reported energy intake in a man with obesity is far below his estimated needs, yet his weight is stable. What is the likely explanation?
Energy under-reporting, which is common and linked to higher body mass index; a weight-stable person's intake must match expenditure.
Connections
Educational information, not medical advice. For any symptom, see a health professional.
Sources: 24-hour Dietary Recall (24HR) At a Glance (National Cancer Institute (NCI), Dietary Assessment Primer, 2026); Food Record (National Cancer Institute (NCI), Dietary Assessment Primer, 2026); Food Frequency Questionnaire (National Cancer Institute (NCI), Dietary Assessment Primer, 2026); Types of Measurement Error (National Cancer Institute (NCI), Dietary Assessment Primer, 2026); Misreporting (National Cancer Institute (NCI), Dietary Assessment Primer, 2026); Dietary reference intakes tables: Overview (Health Canada, 2025); Guía de Alimentación para la población Colombiana Basada en Biodiversidad y Alimentación Real (Instituto Colombiano de Bienestar Familiar (ICBF), 2024).
Spotted a mistake?We are always improving for the community. Tell us what is wrong and our team will review it.Report a mistake
Representative model, for educational purposes only.
