House Floor Area as a Correlate of Marital Residence Pattern: A Logistic Regression Approach

Abstract

The goal of this article is to reflect on the relationship between average house floor area and marital residence pattern. This article brings three novel elements to this research problem: (a) it combines cases from the previous studies thus creating a larger empirical base, (b) logistic regression is used as an analytical technique which can test the significance and strength of the correlation between variables and establish explicit classification criteria with corresponding error rates, and (c) the effect of subsistence mode on the correlation between house floor area and residence is taken into account. New results confirm previous findings that there is a statistically significant correlation between house floor area and marital residence but not strong enough to warrant the use of floor area as a sole indicator of residence. Subsistence mode has a large influence on the correlation-goodness of fit of the logistic regression model and its predictive power are significantly higher in agricultural societies, whereas the correlation between house area and marital residence is not significant in the nonagricultural group of cases.