Dietary intake assessment - Glossary of terms

Outcomes of interest

Energy intake

This is measured in kilojoules (kJ) or Calories (cal/kcal) and forms the basis of many dietary studies as one side of the energy balance equation. If more energy is taken in than is required, excess weight gain ensues. If more energy is expended than is taken in, sub-optimal weight gain, weight loss or poor growth can occur. This may be desirable (eg during obesity treatment) or may indicate risk of malnutrition. During childhood, energy needs vary widely depending on the stage of growth, gender and physical activity level. A sub-variable can be created in dietary studies known as kJ/kg, which is total energy intake adjusted for a child's body weight which will allow for better comparisons to be made between children and/or over time.

Total energy intake could also be reported relative to lean body mass or fat free mass to adjust for body composition. Another approach is to report total energy intake relative to multiples of resting energy expenditure, if it has been measured, or to compare the reported intake to a standardised value. This approach may allow for comparisons to be made between studies over time.

Energy density

Energy density is a measure of the amount of caloric energy from a given volume of a food or beverage.

Back to Outcomes of Interest Matrix

Food & beverage intake

Food / foods groups

The basic food groups include: breads and cereals, fruit, vegetables, meat and alternatives, dairy and extras (energy-dense nutrient-poor) or sometimes referred to as non-core foods. Intakes of specific foods can be assessed directly depending on the nature of the research question and may include items such as sweetened drinks (soft drinks, cordials and fruit juice) or dairy foods (milk, yoghurt, cheese).

Assessment and reporting of actual food intake is essential for continual development and refinement of food-based guidelines used for effective prevention of childhood obesity. People eat food and not nutrients.

Dietary patterns

This includes the assessment of whole diet, or quality and variety not just specific components. The study of eating patterns is similar conceptually, to the more traditional dietary approach used in intervention trials which assesses foods and nutrient intakes. However rather than creating or prescribing a specific diet, this approach uses statistical methods such as factor or cluster analysis, usually within a large data set from observational studies, to derive eating patterns or demographic characteristics that are located together. For dietary intake, this ideally leads to the reporting of a meaningful set of eating patterns to describe the way groups of people actually eat.

Eating patterns

Understanding how people actually eat in natural settings through the use of eating pattern methods provides an advantage over single nutrient approaches in nutritional epidemiology. This includes the assessment of eating patterns such as breakfast consumption, the inclusion of snacks, skipping meals, timing of eating, size of meals.

Back to Outcomes of Interest Matrix

Nutrient intake


This category includes the energy yielding nutrients of protein, fat and carbohydrates. These macronutrients are the basis of most foods and beverages in the diet and contribute to an individual’s energy intake. These are often reported as mean daily intakes (g/ day) and/or their individual percentage contribution to total energy intake. This provides direct comparison with national food selection advice, including The Australian Guide to Healthy Eating, National Dietary Guidelines for Australian Children and Adolescents and the Nutrient Reference Values (NRV).


Due to the smaller molecular size and contribution to dietary intake when compared to macronutrients, they are usually measured in smaller units such as milli (mg) or micrograms (μg) per day. If trying to determine usual or habitual consumption of micronutrients, longer reporting periods of dietary intake, often greater than 8 days are required due to the large daily variation in intake.
To estimate intakes of micronutrients a food database that reports the nutrient profile for a nationally representative range of foods consumed is required. Therefore it is important to select the most accurate and current database available for the nutrient or nutrients of interest.

Back to Outcomes of Interest Matrix

Individual influences

Food / eating behaviours

It is becoming more common to try and assess food behaviours as part of dietary intake and obesity related studies. Food behaviours influence food intake and preferences and may include behaviours such as eating in front of television, pace of eating, picky or fussy eating, eating as a family or at a dinner table.

Back to Outcomes of Interest Matrix

Environmental influences

There is a need to consider other influences on food intake and eating behaviours, including neighbourhood, school, social and home settings. This is particularly for children and adolescents who will have less control over these environments compared to adults. Examples include the number and variety of fast food restaurants in the local area, food availability in the home and school, price of food, peer/social pressures, purchasing practices, food marketing and advertising.

Back to Outcomes of Interest Matrix

Practical considerations


This captures the time and effort required to complete the dietary assessment method. Two aspects need to be considered:


How much time do researchers need to dedicate to conducting the dietary assessment method within the study. A researcher needs to consider standardisation and training of personnel (if required), implementation of the method, data entry and analysis, time, analysis method and cost.


  • How and what is the participant being asked to do?
  • How long is the survey?
  • How literate is the participant?
  • How much time do they have to commit to complete the method?
  • Do they have to travel to a venue?
  • Do they need to have any special equipment such as food scales or standard household measures to complete the method? Do they require access to a computer?

Back to Practical Considerations Matrix

Sample size

How many people are going to be part of the study and how many are going to have dietary intake assessed? It is highly recommended before the study commences that a power calculation is conducted. The estimation of sample size from the outset will help in planning the study timetable and articulate the number of participants required to detect a change or difference in the dietary variable that is considered to be important within the study. Freely available software can assist with this or a statistician can be consulted.
It is not recommended to simply select a number of participants that can be recruited. If a change in fruit and vegetable intake by 0.5 serve/day over a three week period is to be detected, there is a need to determine how many participants are required before study commencement to detect a change so that a study is not conducted with inadequate numbers to conduct the statistical test with sufficient power.

Back to Practical Considerations Matrix

Potential for bias

Do participants report accurately? There is always a need to consider key potential biases influencing estimates of dietary intake. This may include:-

Recall bias: Can people accurately remember what they ate?

Responder bias: Are the people participating in the study different from the general population? Are they more motivated or different in some way?

Reporting bias: Are people reporting what they think is important in relation to the research study they have consented to participate in?

Social desirability: Are people reporting what they think other people do, i.e. the ‘norm’, or what they think the researchers want to hear?

Back to Practical Considerations Matrix

Represents usual intakes

How reflective is the dietary data collected? Does it capture one day, many days or the average over three or 12 months?

Back to Practical Considerations Matrix

Type of data

Does the research require data such as how the consumer feels about foods available in their environment or the way foods are cooked. This type of information can be obtained by individual interviews or from focus groups with analysis identifying key themes, or some data could be collected as part of a diet history. Does the research require quantitative data such as the daily intake of specific foods e.g. breads, macronutrients (grams of protein) or micronutrients (mg of calcium). These quantitative methods will require access to a National specific food and nutrient database for the analysis.

Back to Practical Considerations Matrix


How much finance is available to undertake assessment of diet? How much time needs to be allocated by researchers for the data collection, entry and analysis? Do researchers need to have pre-requisite skills to conduct the dietary assessment such as being an Accredited Practising Dietitian (APD)? Do they need to be trained in order to standardise the dietary data collection technique? Is there a cost to use the dietary tool? Is there a cost associated with analysis? Are participants going to need to be provided with equipment such as food scales or standard household measures in order to complete a weighed food record? Will they be out of pocket for any items related to assessment of their dietary intake and if so will they be reimbursed?

Back to Practical Considerations Matrix

Populations of Interest


The age of children needs to be considered as their attention span, cognitive abilities, literacy skills will influence the selection of the dietary methods and the quality of the dietary information that is to be collected. It is generally accepted that children under the age of 9 yrs cannot accurately report their own diet intake and a parent should be used as a proxy. With increasing age, children become more accurate when reporting over shorter periods of time such as the previous days or week. However, accuracy declines as the time period over which the child is asked to recall intake increases. Parents and/or carers can be used to supplement the child's information.
The Populations of Interest Matrix provides details of who can be used as the reporter for each diet method.

Back to Populations of Interest Matrix


There is a need to consider where the research is going to be conducted, how many people are going to be involved and how much space, how the location may need to be set out for an interview versus completion of a questionnaire and the time needed to complete the tasks.


Often research studies are conducted in clinical environments such as hospitals or clinics. Hospitals can be a convenient location for participants, particularly if they already have scheduled medical appointments. They are often centrally located and there may be facilities for a range of measurements to be collected such as height, weight or blood testing facilities and able to cater for a number of participants (individual or group) and have appropriate public transport services.


Issues to consider for data collection at an individual's home include the usual Ethics Committee requirements as well as Occupational Health & Safety issues. An allowance for time for researchers to travel to and from participant homes as well as provision of all equipment needed for assessment will need to be addressed. Access to a vehicle to allow researchers to travel will be necessary. Organisation of suitable times for both researchers and participants can timely, along with having a high researcher burden if visits are needed on more than one occasion.


This type of study involves recruitment of a moderate number of free living community participants (20-200?). The aim of undertaking a study in this setting is to reduce the recruitment bias in order to establish ‘true’ or real diet relationships.


Large scale population surveys with participants greater than 1000 require detailed planning.

Issues which should be considered:-

What is the procedure and method of recruitment? How will data be collected and analysed for such large numbers? What data quality checks will need to be made? Is there appropriate funding, time and expertise available for this to occur?

There are a number of benefits to this type of study including large numbers, potential for the detection of diet disease relationships over long periods of time and population consumption patterns.

A key consideration if participants are going to be followed up at a later time will relate to repeated use of the same dietary intake assessment tool and ensuring the participant can be tracked in the future. If so, secondary contact details of other family members may be crucial to establish.

Back to Populations of Interest Matrix

Administration method

Dietary information can be collected and obtained from participants in different ways, some of which may help to reduce participant and researcher burden.


This is a common dietary collection method and involves the participant reporting their own dietary intake. This is commonly done via a face-to-face interview, but can also be performed by a mailed survey, telephone interview or via newer electronic methods. For children of younger ages, it is recommended that a parent or main carer act as a proxy for the child. Older children and adolescents are able to self-report.

If using a mail survey then it is recommended that the surveys are pre-coded with a participant ID number to avoid confusion when surveys are returned and to avoid lost data if the survey comes back with no name. To increase the likelihood of participants returning surveys, it is also useful to send a reply paid pre-addressed envelope with the possibility of an incentive or reward, to improve the response rate or ease the burden on the participants or compensate them for their time.


Dietary information is collected at an in-person interview or over a telephone which may or not be assisted by CATI*. To ensure that data is collected effectively and efficiently, research personnel conducting the interview should be trained and quality checking of the process throughout the collection period. This will help to eliminate errors and reduce biases (see bias above). Due to the high costs involved in face-to-face interviews and the time needed to collect information from participants, there has been a trend away from this method. Travel costs to participant’s homes and personnel costs involved in face-to-face methods are typically higher than they are for mail, telephone, or web surveys.

*CATI - Computer Assisted Telephone Interviewing allows a dietary interview to be conducted with a computer assisted prompting system. This system increases the efficiency and cost effectiveness of conducting large scale surveys. It reduces coding and data entry errors and also improves data management and processing capabilities since these tasks are automated, avoiding the necessity for separate coding and data entry steps.

Direct observation

This method is resource intensive and should involve trained researchers directly observing a participant’s food intake and behaviours for defined periods to ascertain foods, brands and portions consumed. This may or may not be within a research environment and it is often preferable if a participant is not able to identify who the observer is so as to not alter their dietary intake.


This is a newer means of collecting dietary information but becoming more popular due to the benefits that it can provide and the increasing computer literacy of the population.

Study websites can be set up that rely on passwords and/or PIN numbers to ensure survey responses are confidential, that only respondents selected for the study have access to the survey and that there is no duplication of survey respondents.

It is useful for studies where there is no central location for data collection or where it is difficult for participants to meet for a face-to-face interview. Through liaison with a web development team, basic online surveys for food intake or behaviors can be set up. Food diaries or weighed food records can be entered with information supplemented from food scales or standard household measures but little is known on the validity of these methods currently.

Back to Populations of Interest Matrix