Determinants of Residential Satisfaction: Ordered Logit vs. Regression

Models

Growth and Change, Spring 1999, vol. 30, no. 2, pp. 264-287(24)

Lu M. [1]

[1] Kansas State University, Manhattan

Abstract:

Residential satisfaction is not only an important component of individuals' quality of life but

also determines the way they respond to residential environment. An understanding of the

factors that facilitate a satisfied or dissatisfied response can play a critical part in making

successful housing policies. This study reinvestigates the effects of housing, neighborhood,

and household characteristics on individuals' satisfaction with both dwelling and neighborhood,

in order to reconcile the inconsistencies in the previous research. The empirical analysis uses

data drawn from the American Housing Survey (AHS) and ordered logit models (OLM). OLM is

more appropriate than the widely-used regression technique in such analysis due to the ordinal

nature of the dependent variables representing satisfaction. The results show that residential

satisfaction is a complex construct, affected by a variety of environmental and

socio-demographic variables. While the actual effects of the variables by and large confirm

earlier findings in the literature, significant differences between the results from the OLM and

regression models were found. This indicates that regression models should be used with

caution and their results accepted with a grain of salt.

Language: English Document Type: Research article ISSN: 0017-4815

Publisher: Blackwell Publishers Inc., Boston, USA and Oxford, UK