Predictors of physical activity in residents of Ontario: An analysis of the 1990 Ontario Health Survey.

Abstract

Objectives. The primary objective of this study is to identify predictors of leisure time physical activity (defined as the reported level of energy expenditure in non-work activities) in the adult population of Ontario. The secondary objective of this study is to compare the predictors in different subgroups of the Ontario population. Subgroups divided by age group, gender and whether the subject lived in a rural or an urban area will be examined specifically. Methods. Two different software packages were used for this thesis. Statistical Analysis Software (SAS) was used for preliminary analyses. Final models were tested using Statistical Analysis for Correlated Data (SUDAAN) software which is designed to adjust for the multi-stage sampling design. Subjects were included in this study if they were over the age of 18. Energy expenditure (NEW), the outcome variable, was calculated by summing the 20 different activity specific scores. Twenty-nine independent variables were considered as potential predictors. Three variables were excluded because they were found to be collinear with other potential predictors. The main statistical method used was multiple linear regression. (Abstract shortened by UMI.)