Dietary assessment and physical activity measurements toolkit

Choosing a dietary assessment method

The aim of the interactive decision matrix is to help researchers ascertain the aspect(s) of diet that are fundamental to the research question. To use the matrix:

  1. First decide if the analysis will be at an individual or group level and go to the relevant section of the matrix; some dietary assessment methods may vary in suitability for outcomes at an individual versus a group level .
  2. Enter the primary, and up to two secondary, desired dietary outcomes.
  3. Think about the time frame of the desired outcomes carefully.
  4. Each method is then considered in the light of practical considerations which will ultimately determine the choice of method.

Methods of dietary assessment included in the decision matrix

The methods in the decision matrix are confined to those most commonly used in population health research; therefore duplicate diets and observation are not included. Due to the limited number of biomarkers currently available to assess diet these are also not included. Emerging technologies are likely to change how dietary data are collected with the web and mobile phones and personal device assistants being increasingly utilised. To date these emerging technologies are at an early stage and until their respective efficacies have been established they have not been included in the matrix. Many dietary questionnaires exist which have been designed to answer specific questions, these are diverse tools and cannot be covered in the decision matrix. It has been assumed that a chosen food frequency questionnaire is appropriate for the foods eaten by the population or individuals being investigated, and a diet history follows the classical approach of a 24 hour recall and a detailed history about diet in the past.


Cost will ultimately dictate which method can be used. If the ‘best method(s)’ is not possible for financial or logistical reasons, it may mean compromising and choosing a ‘less good method(s)’. The time and skill and hence money required to undertake dietary data processing from diaries or recalls should not be underestimated.

Seemingly simpler and cheaper methods such as FFQs are expensive at the outset as ideally they should be validated in a population similar to the study population. Analysis of the data is often more complex than anticipated and may require the development of a computer program; scanning systems are expensive. FFQs are only cost effective if researchers have access to an already developed tool for large study groups. There may be instances that due to the study budget or study population it is necessary to re-visit the research question and primary outcome measure, to ensure that they are compatible with methods which are feasible.


For each of the diet and physical activity assessment methods there is a detailed section on the toolkit which should be read in conjunction with the matrices.
Users are advised to refer to the specific pages for each assessment method as well as reading the information provided in the matrices. 

Web design by Studio 24