Much of the literature on immigrant integration suggests that such processes involve multiple dimensions of integration outcomes (i.e. economic, social, political, cultural). In Canadian context, while much of the literature has traditionally focused in economic dimension, a growing body of literature addresses non-economic dimensions of integration. Yet, less research has explored to model the multi-dimensionality of immigrants’ integration processes by various social attributes. Of the scholarship on immigrants’ integration there is still little consensus on the measure to be used for each dimensions, as well as how to reconcile the limitation of multiple outcomes of integration.

The previous research uses either item-specific regression techniques to describe social group differences in one dimensional integration outcome or factor analyses in order to combine multi dimensional outcomes of integration in order to creating composite integration score. Both of these approaches cast limitations. In the case of the former, complexity of interrelationships among multiple outcome measures are lost by estimating one outcome various at a time. Also, reporting of results can be cumbersome when various measures of outcome measures are analysed in a set of regression equations. With respect to factor analysis, more complex understandings can be gleaned, however detail and specificity of measures is compromised.

In this paper, the Structural Equation Modelling is examined to explore its potential for understanding immigrant integration. In particular, the confirmatory factor analyses allows us to model multi-dimensionality of immigrants’ settlement outcomes. Moreover, by comparing the models by social groups, it highlights the difference/ similarity of integration processes among various social groups (i.e. race, ethnicity, generation status). Using the Statistics Canada’s Ethnic Diversity Survey (2003), the paper explores the potential use of the SEM, testing a conceptual model, which contains 2 latent variables for socio-cultural integration with 10 observed items. The models are compared by gender, ethnicity, and generation status.