Questionnaires vary greatly in their scope and level of detail and dimensions are not recorded to the same extent by each. The face and content validity of the specific questionnaire should therefore be considered alongside the research question, for example:
Questionnaires are limited in their ability to estimate physical activity energy expenditure at individual level and have varying validity for categorising individuals into groups and ranking activity levels. The dimensions of physical activity assessed by questionnaires are described in Table P.2.1.
Table P.2.1 The dimensions which can be assessed by physical activity questionnaire.
|Dimension||Possible to assess?|
|Total physical activity energy expenditure||✔|
|Behavioural pattern: Timing of activity|
|Behavioural pattern: Bouts of activity|
|Behavioural pattern: Variability of activity|
|Contextual information: Domain||✔|
|Contextual information: Location||✔|
|Contextual information: Social||✔|
Physical activity questionnaires are completed retrospectively. In contrast, a diary or log is completed as physical activity occurs. The respondent typically answers questions about the duration and frequency of physical activity of different types, intensities and/or occurring in different domains or contexts. Questionnaires vary from very brief (1-4 items) to much more detailed (60 items).
The time frame varies from previous day to entire lifetime:
Options regarding use of questionnaires include:
Interviewer vs. self-administered
Benefits of electronic vs. paper-pencil
Proxy methods refer to the implementation of questionnaire methods in scenarios where the respondent is not the individual being assessed. The choice of proxy-reporter may be vital to the accuracy of the reports, and normally made on the basis of intimate knowledge of the individual, proximity, or their professional capacity.
Assessing physical activity by questionnaire is a complex task, which may be particularly difficult for some populations, such as: young children, adults with cognitive impairment, chronically ill, disabled. Individuals may lack the cognitive ability to recollect the intensity, frequency and particularly the duration of activities. Alternatively, they may not interpret questionnaire items as intended (especially those with complicated designs), or fully understand the meaning of abstract terms such as “moderate to vigorous physical activity”.
Based on issues of equity, inclusivity, sample size, missing data and bias, use of proxies may be preferable to excluding or not investigating individuals and populations who cannot self-report. Proxy methods also provide some degree of objectivity of estimates of physical activity – individuals do not report their own activity and should have no control over what is recorded.
Use of a proxy-reporter has limitations, such as:
Since questionnaires vary greatly in their scope and level of detail, so does their usage in different types of study. Physical activity questionnaires are the most widely used subjective instrument in large population-based cohorts and surveillance systems.
Brief 1 - 4 item instruments have been used in:
Short 5 - 15 item instruments have been used in:
Detailed 15 - 60 item instruments have been used to:
The outcome variables produced from questionnaire data vary by method, but can include:
Interpretation of data from a questionnaire is aided by additional information, such as:
It is usual for MET values to be assigned based on the intensity or type data reported by the participant. These MET intensity scores are used alongside the questionnaire reported duration and frequency to derive the volume of activity.
This approach assumes the following:
The extraction of features such as MET-hours in different types or domains is an intermediary step, these features can then combined to estimate final target variable(s), or be used in their current state.
Strengths and limitations vary by questionnaire; however, the key characteristics of questionnaire methods are outlined in Table P.2.2.
Table P.2.2 Characteristics of questionnaire methods.>
|Number of participants||Small to large|
|Researcher burden of data collection||Low|
|Researcher burden of data analysis||High if completed manually|
|Risk of reactivity bias||No|
|Risk of recall bias||Yes|
|Risk of social desirability bias||Yes|
|Risk of observer bias||When interview is used|
|Participant literacy required||When self-administered|
Table P.2.3 Physical activity assessment by questionnaire in different populations.
|Infancy and lactation||Requires proxy.|
|Toddlers and young children||Requires proxy.|
|Other adults||May require proxy depending on cognitive function.|
|Ethnic groups||Requires language/cultural specificity. If translating a current questionnaire, translation and back translation is good practice.|
A list of specific questionnaire instruments is being developed for this section. In the meantime, please refer to the overall instrument library page by clicking here to open in a new page.
If a new questionnaire is to be developed:
For proxy methods, it is not advisable to simply translate a self-report instrument to proxy-report without testing. A suitable proxy-report method must therefore: