As far as the heteroskedasticity is concerned, it will cause the estimated variance of the coefficient to influence both direction, upward and downward. In other words, the estimated of variance of the coefficient can be overestimated or estimated depends on the nature of causing factors in the model. If we persist in using the usual testing procedures despite heteroskedasticity, whatever conclusions we draw or inferences we make be very misleading. As a result, we can no longer rely on the conventionally computed confidence intervals and the conventionally employed t and F tests. The effect of heteroskedasticity on the test for homogeneity across samples has concentrated on the case of heteroskedasticity between samples but not within samples in the previous research. The most important conclusions of this paper in terms of methodological point of view can be summarized as follows: (i) Hedonic models of housing prices must be corrected for heteroskedasticity to ensure greater efficiency in the estimation of hedonic prices wherever the heteroskedasticity does exist; and (ii) After careful detection of heteroskedasticity, a combination OLS(Ordinary Least Squares) with WLS(Weighted Least Squares) estimates provides the greatest predictive power with respect to estimate of housing prices in the study area. In other words, if there has no evidence of heteroskedasticity, the OLS should be utilized; otherwise, the WLS must be utilized.
I. Introduction
II. The Problem of Heteroskedasticity
III. Research Method
IV. Data Description and Result of Analysis
VI. The Comparison of the Three methods
VI. Conclusions
References
(0)
(0)