| Summary: | Thesis (M.Sc.)--Memorial University of Newfoundland, 2011. Geography Bibliography: leaves 108-114. Traditionally, the majority of quantitative studies that analyze labour market phenomena have utilized global data and top-down methods. Often, however, labour market dynamics observed on the national or provincial scale are the result of processes operating on a lower-level, local scale. This discrepancy between the scale at which labour market processes operate and the scale at which they are studied reduces the resolution of analysis, and may result in faulty policy development. In addition, because labour markets normally operate in discrete space, traditional local methods which operate in continuous space, such as geographically weighted regression (GWR), are not appropriate. In this thesis, a method of adapting the GWR model to discrete space is described and tested. To evaluate its effectiveness, the discrete-space GWR (DGWR) technique is applied to the aftermath of the 1992 cod moratorium in Newfoundland and Labrador. To do this, a theoretical economic model of moratorium susceptibility and impact is constructed and tested using a DGWR, ordinary least squares, and continuous GWR model. Upon conducting the analysis, it is found that the DGWR is the superior technique with respect to both model fit and the mitigation of spatial effects. In addition, the DGWR model also produces more realistic and easily applicable empirical results than the existing alternatives.
|
|---|