Abstract
Quantifying the level of poor housing within the United States has traditionally been based on the use of combinatorial analysis of census-derived indicators of housing conditions. Units falling below a threshold value on each indicator are defined as being of poor quality. Housing experts from the U.S. Bureau of the Census have produced data showing that census variables, when used as housing quality indicators, must be viewed as random variables, not definitions. The paper that follows explains the use of combinatorial analysis where the indicators are recognized to be random variables. Based upon the results of the Five City Housing Survey of the U.S. Census, the level of error associated with a set of combinatorial rules is estimated and the principles needed to locate the optimal rule established.
Get full access to this article
View all access options for this article.
