Factor Analysis
As mentioned previously, neighborhood stress is directly linked to social and economic conditions but on the first run of the factor analysis only one variable on income (income status) was included. This gave results of a principle component factor that correlated highly with Minors Population. The resultant index map is shown below.
The variables used in this analysis were extracted from Canada Census Data 2001 at the dissemination area level. Extraction of the variables and sorting of the numerators and denominators and organising of the tables was done. The Z-Scores were then calculated so as to standardize the variables. Next I performed factor analysis on the variables with the aim of finding the principle component.
In the analysis, five factors accounted for 70% of the variability in the variables. In otherwords, the extracted factors represent the variables with 30% loss of information.
The Scree Plot helps determine the optimal number of components and eigenvalue of each component is plotted in decreasing order. In this case the components extracted were those with eigenvalues greater than one although sometimes the components on the steep slope are extracted.
The Rotated Component Score Matrix helps to determine what each component represents.
- The first component is highly correlated with Low Income Incidence and Employment
- The second component is highly correlated with Education and Language
- The third with Elderly Population and Dependency
- and so on
