Barbara Herr Harthorn
Institute for Social, Behavioral, and Economic Research
& Department of Anthropology
UC Santa Barbara

POSITION STATEMENT

What are your research interests in the areas of inequality and equity, and what spatial dimensions do you currently or potentially see in them?

My research in the past 5 years has centered on several issues surrounding the production of health inequality. I am involved in 3 separate but related research projects, 2 of which have significant spatial dimensions. The one most directly relevant is a collaborative project with my UCSB anthropology colleague, Susan Stonich, in which we are examining the possibilities for using Public Participation GIS (PPGIS) as a community capacity building tool in mediating community conflict over perceived health consequences of pesticide drift on the agricultural/urban interface. This ongoing study of the northern Santa Barbara Co. town of Lompoc is building a comprehensive GIS to share with the community; we plan to include time series analyses by census block group of ICD-9 diagnoses, sociodemographic data, pesticide application data, meteorological data, air quality data (as a measure of exposures), and other relevant information as it becomes available. This involves close and sometimes contentious work with the interagency working group of regulators and politicians that has formed to assess the conflict and to oversee the government's assessment process. The Lompoc Interagency Work Group includes: the federal EPA, Cal EPA and its Department of Pesticide Regulation and Office of Environmental Health Hazard Assessment, the state Department of Health and Human Services, county and local government health, air quality, agricultural, and other regulators, local legislators, and community organizations. As one component of the PPGIS project, we are developing a program to work w/ high school and junior high teachers in the community to enable them to access and manipulate the data files from our study so that students in the community can begin to develop their own GIS maps on the community. We are also conducting cultural-model analysis of this environmental problem to examine variation among residents' and experts' views about the problem. In addition, we are planning to seek more funding to pursue new primary data collection on perception of risk in the community and plan to use GIS statistics to examine the spatial relationships between exposure to risk and perception of risk.

My own ethnographic research for the past 5 years has focused on aspects of farmworker health in central coastal California using an environmental justice approach. This work has included a study of the public health system's and private nonprofit organizations' roles in the diagnosis and treatment of tuberculosis among farmworkers, and a prospective study of farmworker maternal and newborn health. Both studies examine links between farmworker living and working conditions, health care access, health care delivery practices, and particular health consequences and treatment outcomes. We have used GIS in one of these studies in a rudimentary way to look at the distribution of TB cases in the county. As an offshoot of these studies, I have also embarked on an historical study to track the emergence of racialized discourse in biomedicine and public health about the etiology and treatment of infectious disease among Latino immigrants in California in the first 3 decades of the 20^th century. The purpose of this work is to demonstrate that recent efforts to deny care to Mexican-origin immigrants (e.g., the passage of Prop. 187) reproduce historical xenophobic processes in California that have resulted in profound ethnic disparities in health and health care. There is not yet a GIS component to this work.

What kinds of spatial data, models, techniques, software, etc. do you use or have considered using in your research? Which of these work well for you? Where do you see problems and/or shortcomings?

The technical difficulties we have encountered in conducting our (limited) spatial analyses have been substantial. With CSISS and the NCGIA on campus, we have extraordinarily good access to tools and advice, but both of us, though solidly trained quantitative behavioral scientists, need to learn a lot to become capable users of spatial statistics. The learning curve for competent use of ArcInfo and even ArcView is particularly an impediment to dissemination at the community level, compounded by scarcity of resources (like accessible computers adequate to the task, inability to print GIS maps w/out access to expensive color printers, etc.). The third world contexts where we and many of our colleagues in anthropology work provide even more difficult access problems.

In the Lompoc pesticide drift case study, we have had significant problems with data management and incommensurate levels of analysis. We had difficulty both strategically and technically in accessing individual patient records with the address identifiers needed to geocode the diagnostic data. Our research assistant (UCSB geography graduate student Rebecca Powell) was compelled to aggregate the data to block group level on site (at the county) so that we would not have any data sets at the university with individual patient identifiers in it. This means we have no point data with which to conduct the sorts of statistical analyses more familiar to us. The GIS to date on the project has been compiled using ArcInfo. We're now (following colleague Stonich's participation in the August CSISS workshop) beginning spatial statistical analyses of particular diagnoses in order to better understand apparent clusters. We will probably use SpaceStat and tools provided at the workshop (which are not generally available) for these analyses. We are still not certain if census block group is the ideal level to aggregate to in terms of balancing invasion of privacy issues with community desires to know the spatial organization of risk in a fine-grained way. For low incidence, high severity disorders (e.g., childhood cancers), this level will probably not protect privacy very well in a community of 40,000 people, but going to census tract level loses much of the neighborhood level specificity that community members desire. We're also exploring possible use of ArcExplorer as a free tool that K-12 schools and teachers could use with the PPGIS.

In the second study, we used ArcInfo to develop a GIS on TB cases in SB Co. The data from the county were extremely difficult to manage for the purposes of this study--half of the county's data on TB at that time was not even digitized, much less in a database format we could work with--and we ended up w/ zip code level data (i.e., not very fine grained) that wasn't particularly useful�it simply confirmed what we already knew by then from our ethnographic work about where clusters of TB cases (among different ethnic groups) were located and their distance from care.