Individual, social, and environmental factors associated with physical activity and walking

Background: Participation in physical activity (PA) is influenced by a multitude of factors. Traditionally, research has focused on several theoretical frameworks focusing on the individual ; however, they do not necessarily take into consideration other influencing factors such as the social environment or the physical or built environment. As such, a comprehensive socio-ecological model considering a multiplicity of factors is useful in explaining behaviour. Aims: To 1) assess the prevalence of the individual level correlates and their association with PA and walking; 2) assess the prevalence of environmental determinants and neighbourhood characteristics and the association between these and PA and walking behaviours; 3) explore within a comprehensive and socio-ecological approach, the contribution of the individual, social, and environmental factors in predicting PA and walking. Methods: The studies used in this thesis are national, random-digit dialling telephone-based surveys of a representative population sample within Canada. All research questions and procedures underwent ethics review at York University. The studies incorporated a two-stage probability selection process to select a survey respondent, and included a number of standard self-report measures across the data collection cycles. PA and all-domain walking were measured using the telephone-administered, short International Physical Activity Questionnaire, the neighbourhood environment was measured using an abbreviated version of the Neighbourhood Environment Walkability Scale (NEWS), and individual factors such as knowledge about amount of PA required for guidelines, beliefs about the benefits of PA, self-efficacy, intention, and initial behaviour changes. Walking for transport was measured through the Physical Activity Monitor and walking for recreation were measured through an adapted version of the Minnesota Leisure-Time PA questionnaire (for the 2007 collection only). Complex sampling methods were required to take into account stratification by province or territory within Canada. Complex samples cross-tabulation procedures were used to calculate the prevalence estimates of Canadians meeting the PA and walking guidelines and 95% confidence intervals. The relationship between factors predicting sufficient activity and sufficient walking were examined using complex samples logistic regression procedures that were reflect the sample design. This thesis explored associations and the relative strength of the factors as the independent measures predicting sufficient PA and sufficient walking as the dependent measures, using age, sex, and education as covariates for each of these models. Chapter Six expands this model by including walking for recreation and transportation, and examining sub-population groups. Results: Individual factors (e.g., self-efficacy, intention, and some trial behaviours) and social factors were associated with sufficient PA and certain types of walking. Relatively few environmental factors were associated with sufficient walking (all domain and domain-specific) or sufficient PA. The relationship between high density neighbourhoods and higher rates of walking (generally and specific), and the availability of supportive walking facilities with various modes of walking were evident. Proximity of many shops and the presence of sidewalks were associated with the highest quartile of walking for transport. A greater number of the individual factors predicted walking and PA compared to the environmental/neighbourhood factors, within the context of a full socio-ecological model. Findings differed when stratified by age and sex of respondents. Conclusions: The results suggest that individual factors may be more relevant for predicting activity and walking than environmental factors, or at least should be considered in their inter-relationship with environmental factors when developing environment-based interventions. Although the inter-relationship between individual factors, social factors and the built environment are important, understanding individual factors are critical for determining strategies and interventions to promote PA among certain populations with traditionally lower levels of activity. Findings suggest that within countries like Canada, with a relative abundance of supportive environments, more specific and detailed measures of the perceived and objective physical environment may be required in order to achieve sufficient variation