Open Collections

The factorial urban ecology of Greater Vancouver : characteristics of the data base

University of British Columbia

The methods of factorial urban ecology are followed for Greater Vancouver for 1961. One hundred and ninety-two variables are factor analysed to produce the seven most important underlying dimensions of areal differentiation. The factors include the three constructs of social area analysis. Factor 1, Family Status, factor 3, Housing Tenure, factor 5, Families with Older Children, and factor 6, Rural-Urban Status are all required to encompass the family status construct. Factor 2, Socio-Economic Status, and factor 7, Retail and Clerical Workers, are both required for the economic status construct. The ethnic status construct is paralleled by factor 4, Ethnic Status, which includes a wide range of ethnic groups. An examination of the first four factors for correspondence with theories of urban structure leads to the conclusions that Family Status displays an "interrupted radial" spatial pattern, and that Socio-economic Status displays a sectorial pattern centred on Vancouver City, and a radial pattern centred on New Westminster. Housing Tenure shows tendencies towards a multiple nuclei structure, though the complexity of the City of Vancouver dominates the pattern. Ethnic Status is concluded to have a unique pattern, unlike those predicted by any of the urban models. The data available to factorial ecologists are rarely normally distributed. Procedures are developed with which each of the variables is transformed to a distribution which is normal or very close to normal. The factor analysis is performed using both untransformed data and normalized data. In comparing results, it is concluded that normalization aids the ability of the correlation coefficient to measure linear association among variables. This helps the rotation procedure to achieve simple structure and consequently, the factors based upon normalized data are more easily interpreted. They include variables which seemed more closely interrelated and seem to correspond more closely to theoretically derived hypotheses about urban areal differentiation. This is especially true of the Ethnic Status factor, where a considerably larger number of ethnic groups are incorporated in the factor based upon normalized data. In addition to helping in the search for parsimony and for simple structure, the normalizing of data would seem to produce results which will be more easily comparable, ceteris paribus, from city to city and hopefully among countries as well. It is tentatively claimed that the use of normalization procedures such as the ones developed here will have a beneficial impact on our ability to describe and hence understand the ecological structure of the city. The areal unit used in this analysis is the enumeration area, about one tenth the size of census tracts. This provides the opportunity to investigate the importance of the level of data aggregation. The approach used is to consider census tracts as aggregations of enumeration areas and to compare the information on areal differentiation lost through averaging with this aggregation to that lost with one achieved with an optimal analytic grouping procedure. It is concluded that the internal variability of demographic, socio-economic, and housing characteristics of urban dwellers within census tracts is high enough to mask important information about the population. The best data base is one at the lowest level of aggregation suitable for data handling and at which data error is sufficiently small. Census tracts are too large to be considered suitable as the unit of analysis for factorial urban ecologies.

Text

2010-01-22

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.

http://hdl.handle.net/2429/19016

Geography

University of British Columbia

Graduate