Modelling effects of intervening variables using path analysis

Path analysis models the effects of independent variables on dependent variables via intervening or mediating variables. As such, it models pathways by which variables affect each other. As a simple example, we may expect physical proximity to services and facilities to positively impact on level of satisfaction with one’s neighbourhood via satisfaction with access to services and facilities. So, a path analysis would model the direct effects of, say, physical proximity to services and facilities on satisfaction with access to services and facilities, and of satisfaction with that access on overall neighbourhood satisfaction, as well as the indirect effects of physical proximity to services and facilities on neighbourhood satisfaction. However, other variables also affect neighbourhood satisfaction, both directly and indirectly, and so including these would lead to a more complicated path analysis. Other names associated with path analysis are analysis of covariance structures, causal analysis or modelling, simultaneous equation modelling and structural equation modelling, all of which can be used to do path analysis. Figure 22.1 shows path analysis in its simplest form. The independent or exogenous variable (x) predicts the intervening or mediating variable (y) which in turn predicts the dependent or endogenous variable (z). This approach models a path of effects, usually representing a hypothesized process where x leads to y leads to z. The independent variables (x) are called exogenous variables because they are determined outside the system.