Graduation Year


Document Type




Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Aging Studies

Major Professor

David Chiriboga, Ph.D.

Committee Member

Kyriakos S. Markides, Ph.D.

Committee Member

Russell S, Kirby, Ph.D., MS, FACE

Committee Member

Yuri Jang, Ph.D.

Committee Member

Hongdao Meng, Ph.D.


Geographic Information System, Multilevel Mediation, Segregation Index


Research suggests that living in communities with high densities of persons from their own ethnic group improves the overall health of older Mexican Americans. One hypothesis is that residing in high ethnic density areas allows characteristics of Mexican culture such as strong social ties and social cohesion, to have a beneficial effect. The majority of investigations focused on ethnic density effects, however, have utilized relatively loose interpretations of what constitutes the appropriate social-geographic area to be studied. Moreover it is not clear how certain dimensions of residential segregation are protective or harmful toward health, particularly when measuring ethnic residential segregation from a geographic information systems (GIS) perspective. The effects of ethnic density and segregation have not been directly or quantitatively tested using the kind of multi-level methodology that can effectively capture data from both personal and environmental characteristics. The present study assessed how multiple geographic/neighborhood factors including ethnic density, neighborhood social cohesion, and social ties may serve as resources for health and health service use of older Mexican Americans.

The study had three objectives:

1. To examine whether protective/deleterious effects of ethnic density exist when we use the more commonly used approach to measuring ethnic density (proportion of ethnic group within a specific census unit), and whether the ethnic density effect is increased when an alternative measurement approach (proximity weighted density) is used that relies on more than a single, specific census unit .

2. To examine how different dimensions of ethnic segregation are related to perceived social support, number of social ties, and perception of social cohesion in their neighborhood.

3. To examine whether or not social cohesion mediate the relationship between ethnic density or segregation and health status/health service use.

The study represents a secondary analysis of data from the fifth interview wave of the Hispanic Established Populations for the Epidemiologic Study (H-EPESE; PI: Markides). Using geographic information systems (GIS), proximity weighted ethnic density and residential segregation indices were calculated, as well as more standard measures of density based on composition of the census tract in which participant lived. Since the H-EPESE dataset has a clustered structure where individuals are nested within neighborhoods, multilevel modeling techniques were employed.

Results suggest that the several approaches here employed to measure ethnic composition of the local environment are complementary. First, the proportion of Hispanics in the neighborhood as defined by the use of census tracts, is both simple and the data easily accessible to researchers. This proportion, or what is often called density, was found to associate with several outcome measures in much the same way, and with similar proportions of variance as the more complex ways of method. The latter, however, made significant contributions that often were relatively independent of the census tract based proportions and thus add significantly to our understanding of the role of the ethnic neighborhood. These more complex measures, moreover, may potentially contribute even more: analyses using these newer approaches were limited by the lack of street address or census block data. Access to such data was not possible due to confidentiality issues surrounding the use of highly specific geographic information that could potentially identify the participant. Results did strongly suggest the value of a residential segregation index as a means of demonstrating that the ethnic environment and urban-rural composition of the residential environment contributes to our understanding of the importance of social coherence and social ties. It was found for example that older Mexican Americans who lived in neighborhoods with higher exposure segregation (i.e., neighborhoods where an individual from one particular racial/ethnic group has a higher probability of encountering members of another group, rather than from their own group) have higher depressive symptoms, as measured by the Center for Epidemiologic Studies Depression (CES-D) scale. This relationship was mediated by individual level data on perception of social cohesion. While in all cases causal interpretations were limited by the lack of a true experimental design the results generally do demonstrate the value of the newer, complementary, approaches to assessment of racial/ethnic density.

Included in

Gerontology Commons