Physical activity behaviours can take place in different contexts. There can be several layers of information.
Behavioural context can be measured in several ways which may not capture all subdimensions of context. For example, a questionnaire may be structured by domain but may contain no questions about location or social context. Most wearable activity monitors store activity as a time-series and a GNSS sensor measures geographical location as a time-series. Using clues of time and space and allowing participants to view and annotate their own physical activity data allows the capture of contextual information:
The term “domain” is often used interchangeably with context, setting or location. Here, domain is defined as the broader, higher-level context which typically structures our daily lives. Behaviour by domain varies both within- and between-individuals. There is no definitive way of categorising physical activity behaviours into domains but commonly the following aspects are considered:
Often, occupation, voluntary work, and education are collapsed into one domain, and sometime voluntary work is considered part of the leisure-time domain. Regardless how domains are classified, each domain brings together features of the context which are similar in some way (i.e. they share key characteristics). For example, the occupational domain can be most easily differentiated by its purpose (i.e. paid work), and although the occupational physical activity of an office worker, a shop assistant, a police officer and a lumberjack would be very different and take place in very different spatial and social contexts, these jobs typically share characteristics that there are certain agreed/contractual work patterns, with specifically outlined requirements and responsibilities for individual workers, coordination of work tasks with colleagues or other parts of the organisation, and generally with less freedom and more physical constraints of behaviour change unless dictated by new policy.
On the other hand, the home domain is more related to one single physical location, and the transport domain is characterised by the purpose of getting from one location to another.
The leisure domain offers greater volition over choice of behaviour compared to the occupational domain – this may make it a more suitable target for behaviour change.
Relation to other contexts
Contextual layers of information are not mutually exclusive across layers but categories within layers should mutually exclusive to be operational. There is a degree of correlation between domain and other dimensions of context, e.g. one's job may be in a particular physical location and social context of that is likely to be colleagues, whereas behaviours in the home domain are more likely to be undertaken together with family and friends.
Since domains refer to broader typical life structures, locations and purposes of our daily endeavours, they are often used to aid recall in some forms of physical activity measurement, such as questionnaires (Craig et al., 2003; Wijndaele et al., 2014). This domain-based approach is also useful in terms of identifying intervention targets. Another way to capture information of domains is by the use of simple domain log which is a diary of the start and stop times of each domain for each day, or the use of electronic annotation tools, potentially aided by the visualisation of time-series of activity and location. An example of the use of domains in physical activity research is shown by Figure P.1.7 (Strain et al., 2020).
Figure P.1.7 Ternary plot of the relative contributions of work/household, travel and leisure moderate-to-vigorous physical activity (MVPA) to total MVPA. Mean relative contributions should be read following the direction of the arrows for each axis. For example, in the USA, relative contributions are 52% from work/household, 11% from travel, 37% from leisure. Cameroon, Central African Republic, Chad, Cote d’Ivoire, Democratic Republic of the Congo, Ethiopia, Gabon, Guinea, Madagascar, Maldives, Mali, Mauritania, Micronesia Fed. Sts. and Pakistan are subnational surveys Source: Strain et al., 2020.
The physical world can be indexed by the four fundamental dimensions of space and time. The geospatial context of physical behaviour relates to the physical setting or location of the behaviour in an instance of time. This can be strictly defined as a lattitude/longitude coordinate plus a time stamp or more loosely defined as indoors/outdoors, in green space or busy roads, urban or rural. The latter parameterisation is potentially more useful for behavioural research as these may act as facilitators or barriers to healthy physical behaviour patterns.
Measurement and estimation of features of the physical environment
A GNSS sensor measures geographical location as a time-series which can be combined with specific databases that are indexed by space and time, containing information about land-use or other features of the physical environment to derive estimates of:
Environmental information can also be captured subjectively from the study participant; this class of information is often referred to as the perceived environment. Whilst it may be less accurate than the objectively assessed environmental features, aspects of the perceived environment can be more important determinants of human behaviour than their equivalent actual environmental features, e.g. parents perception of whether it is safe for their child to walk or cycle to school or not, rather than going by motorised transport would have a stronger influence of the child's commute mode to school than official road safety or crime statistics for that route.
The psychosocial context of physical activity behaviour describes yet another aspect of contextual information, namely with whom the behaviour is undertaken and under what psychological/emotional conditions. For the social element, this can be described by the relations to other people, e.g. family, friends, colleagues, teachers, strangers. This includes relations to animals participating in the activity, e.g. horse-riding or dog-walking activities have strong elements of human-animal interaction. For the psychological/emotional element, this relates both to the degree of volition or control of the individual with respect to a particular settings, as well the emotional state, e.g. happy, sad, angry, enthusiastic etc. (Werneck et al., 2021).
Elements of social context that can be measured include:
Co-participation can be objectively measured by proximity sensors. For example, low-power radios such as Bluetooth in smartphones transmit and receive signals of nearby devices which also have low-power radios, a technology also used for contact-tracing during the Covid-19 pandemic. In population research studies, however, psycho-social contextual information is typically captured using dedicated questionnaires that target this dimension alongside recall of activity type, duration and intensity. Alternatively, such retrospective recall of the psychosocial dimension can be captured using annotation tools prompted by time-stamped activity and space information clues as described above. Finally, the information can be captured prospectively using ecological momentary assessment using apps either with random time sampling or prompts informed by detected bouts of activity.