|
|
|
GIS
|
Spatial Data Annalysis and GIS Applied to Study the Urban-Rural Linkage in Change Mai-Lamphun Area, Thailand
5. Spatial Statistical Model
Because the flow of labor is a spatially dependent process, explanation is not completed without some characterization of spatial intersection. The highly-significantly clustered spatial distribution of working-out population - as shown in Figure 2 and indicated by positive strong Moran's index of 0.4096 and Geary's index of 0.5869 respectively-suggested similar destination and source tambols for labor flows may occur in spatial groups. Inclusion of the spatial lag of response variable -a variable representing the neighborhood effect of the working-out population-may help to explain some of the residual variation. Moreover, the high levels of urban-biased (factor 1) and industrial based economic (factor2) in a tambol not only pull back the local labor, but also attract the additional labor from neighboring tambols. The spatial lags of the factor 1 and factor 2 - representing the 'pull' effects of neighboring tambols -thus, were calculated for each tambol (using adjacency weight matrix) and included as additional explanatory variables into the spatial statistical model. This model, termed the regressive-spatial-autoregressive model (Anselin, 1988) was calculated though maximum likelihood estimation in SpaceStat 1.80 (Anselin, 1995) with results summarized in Table 4. The explanatory variables insignificantly influencing on the realization of the response variable were excluded from the model based on the probability of the z-statistics (using criteria p=0.1). The spatial lag term for response variable was highly significant and, more importantly, its addition reduced the spatial autocorrelation in the model residual to an insignificant level (Table 4). As expected the overall fit of the models were improved with the addition of spatial lag variables: the log-likelihood increased and the AIC significantly decreased (Table 3 & 4). The estimates of the model parameters in the spatial lag model were more precise, i.e., the standard errors of the estimates were lower.
Figure 2 Spatial pattern of working-out population (in %) by tambol in Chiang Mai-Lamphun.
Table 4 Results of spatial lag regression analysis solved through maximum likelihood, with diagnostics of residuals.
| Response Variable: In(working-out pop. +1) |
| Pseudo R2=0.4921 Log-likelihood=-88.428 AIC=122.855 |
| Variable |
Coefficients |
Std. Err. |
z-value |
Prob |
| Lagged |
| Working-out pop. |
1.39591 |
0.232631 |
6.000540 |
0.000000 |
| Constant |
0.0497432 |
0.0126568 |
3.930154 |
0.000085 |
| Rural-Urban Indicator |
0.464627 |
0.187835 |
2.473583 |
0.013377 |
| Factor3 |
-0.322413 |
0.0597546 |
-5.395623 |
0.000000 |
| Pop. Density |
-9.88531E-05 |
5.02207E-05 |
-1.968374 |
0.049025 |
| Lagged Factor2 |
0.186305 |
0.105104 |
1.772585 |
0.076297 |
Regression Diagnostics
Breusch-Pagan (heteroskedasticity) = 36.347 (p = 0.000)
Langrange Multiplier test (spatial error dependence)=(p=0.125) |
6. Interpretations and Conclusions
Given the satisfactory diagnostics tests, the relationships between the response and explanatory variables can be interpreted with some degree of confidence. Approximately half of the variation in the working-out population in 1994 is accounted for by the demographic and economic 'push-pull' factors (a reasonably good fit among social science studies). According to the sign of estimated parameters for explanatory determinants and the meaning of the model (Table 4), the significant 'pull' factors were factor 3, factor 2, population density, and significant 'push' factors are rural-urban dichotomy, spatial lag of response variable itself. A closer look at the original economic variables behind factor 3 (Table 1) showed that the accessibility was indeed a crucial factor in providing rural people opportunity to go out for employment, i.e., intensifying the urban-rural interaction. Moreover, the model was also in support of the conventional wisdom that the land pressure was the real force affecting the outflows of free agricultural labor (Table 1). It was found that the population pressure was not a factor affecting the outflow of rural labor. The rural population tend to rushing out to seek for an employment than the urban population as the urban centers were providing good employment opportunity. Concerning the economic factors, the urban-biased economy (factor 1) appeared not significantly affecting the outflows of labor from rural areas, while the industrial-based economy (factor 2) was significantly attracting the rural labor. This findings have more spatial contexts since the urbanization in fact is concentrated mostly around the Chiang Mai city an the rapid industrialization process in the study area since 1986 appeared to have favorable impacts on employment generation for rural population.
In summary, as an important linkage type in the urban-rural interaction, the spatial statistical model of working-out population had provided insights into the mechanism of influences of demographic, economic and social factors upon the outflow of rural labor. In this study, accounting for the spatial association in the data resulted in a spatial model that better extracts information from the variables and has more precise estimates of model coefficients than does the OLS model. By this case study, GIS had proved to be efficient in managing data set (combined both physical and socio-economic spatial variables) for spatial Stastical modeling and for visualization of results for developing and verifying geographical hypotheses.
References
- Anselin L., 1995. SpaceStat, A Software Program for the Analysis of Spatial Data, Version 1.80
Regional Research Institute, West Virginia University, Morgantown, West Virginia.
- Anselin L., 1988. Spatial Econometrics: Methods and Models. Studies in Operational Regional Science. Kluwer Academic Publishers, London
- Sriboonruang S., 1992. Chiang Mai province and its emerging development problems. Faculty of Economics, Chiang Mai University.
- Suwan M., et al., 1992. Impacts of industrialization upon's the village's life in Northern Thailand. Faculty of Social Science, Chiang Mai University.
- Tran Hung, 1997. Integrating GIS with spatial data analysis to study the development impacts of urbanization and industrialization : case study of Chiang Mai-Lamphun area, Thailand. Ph. D. Dissertation, AIT, Bangkok, Thailand (Forthcoming)
|
|
|
|
|
|
|
