The Measurement Of Observer Agreement For Categorical Data Landis

This paper presents a general statistical methodology for the analysis of multivariate categorical data from observational reliability studies. The procedure focuses on the design of the functions of the observed proportions, which focus on the degree of consent between observers, and the establishment of test statistics for assumptions relating to these functions. The interobserver bias tests are presented in the form of first-rate marginal homogeneity, and the interobserver agreement measures are developed as general Kappa-type statistics. These methods are illustrated by an example of clinical diagnosis from the epidemiological literature. ABSTRACT: Socio-economic disability in the neighbourhood has been associated with health behaviours and outcomes. However, the socio-economic status of the neighbourhood was inconsistently measured in all studies. It is not clear whether the corresponding socio-economic indicators vary by geographic area and geographic level. The objective of this study is to compare the composite socio-economic index with six socio-economic indicators that reflect different aspects of the socio-economic environment, both geographically and by level. Based on 2000 U.S. Census data, we conducted a multivariate factor analysis to identify significant socio-economic resources, and established 12 composite indices in the county, census district, and block group levels across the country and in three states. We assessed the consistency between composite indices and individual socioeconomic variables.

The composite component varied in geographic areas. In a given geographic region, the composite index component was similar to census district and block group levels, but different from that of the district. The share of the population below the federal poverty line contributed significantly to the composite index, regardless of geographic areas and levels. Compared to non-component socio-economic indicators, the component variables of the composite index were more pleasant. Based on these results, we conclude that a composite index as a measure of socio-economic poverty in the neighbourhood is better than a single indicator and should be established on a surface and unit-specific basis, in order to accurately identify and quantify the socio-economic inequalities of small regions compared to a given study region.