抄録
This paper is a part of study on the residential distribution of employees, which attempt to make clear the employment effect on population distributionchange. In this paper, we attempted to make clear the preference structure on housing location choice of employees, as the case study in Samarinda Municipality, the Republic of Indonesia.
To get the samples, we selected 29 establishments in Samarinda Municipality, and we inquired to the employees about the experience of residential move after employed, the location of new housing lot and socio-economical attribute by questionnaire. The total number of samples became 272.
As the analytical method, we applied multinomial logit modeleq. (1) which adopted “the location of new housing lot” as the dependent variable and also “the geographical attributes which might influence the preference of residential location choice” as the explanatory variables.
(1)
(2)
P : the probability that zone i will be selected by individual n as the new housing lot X : the k-th attribute of zome i for individual n β : parameter fork-th explanatory valiables
The study area was divided to 32 zones (i=1 to 32) according to the administrative boundary. We adopted 8 variables as explanatory variable of eq. (1), which were commuting distance, CBD distance, former address distance and so on.
The results are concluded as follows.
1. From the calibration of multinomial logit model, the preference structure of housing location was formulated. It showed that commuting distance, CBD distance, former address distance were main factors for preference ofresidential location. And it indicated that the location where closed to work place and former address and also apart from CBD was preferred.
2. From the result of principal component analysis on the parameters calibrated by multinomial logit model, the characteristics of the residential preferences which related to work place location, former address and ownership of new house were made clear.
3. From this study, the significant relationship between the preference of residentioal location and individual characteristics of employees wes not founded.
4. From the simulations, the relationship between location of establishment and population distribution change was visually understood.
