Developing a GIS-Based Online Survey Instrument to Elicit Perceived Neighborhood Geographies to Address the Uncertain Geographic Context Problem

Abstract

Although neighborhood factors have been consistently associated with health, technological difficulties in eliciting self-defined neighborhoods from large cohorts have compromised the interpretability of this research. Here, we offer a mixed-methods approach to elicit and validate self-defined neighborhoods. Participants used a customized Google.Maps interface to “draw” their neighborhood and answered questions about perceived map accuracy, neighborhood definition, and neighborhood activities. We compared geographic concordance of drawn and narrative neighborhood definitions, quantified differential accuracy by demographic characteristics, and examined factors influencing neighborhood definitions. We found similar geographic concordance between narrative and mapped boundaries in two cities, with no differences by neighborhood size. Self-reported neighborhoods had greater concordance with larger administrative areas (e.g., police precincts) than for smaller units (e.g., census tracts). To delineate their neighborhood boundaries, participants reported using administrative definitions, walking distance, their familiarity with people and structures, where they spend time, and physical landmarks. In New York City, participants also reported considering sociodemographic characteristics and transportation. Our method demonstrates the feasibility of collecting perceived (egocentric) neighborhoods through online mapping surveys, adaptable to many study settings.

尽管邻里因素持续与健康相关, 但从大型群体中诱发自我定义的邻里之技术困难, 却使得此般研究的可诠释性有所妥协。我们于此提供一个诱发并验证自我定义的邻里之混合方法取径。参与者使用客製化的谷歌地图界面来 “绘製” 其邻里, 并回答有关感知的地图正确性、邻里定义和邻里活动的相关问题。我们比较绘製和叙述的邻里定义的地理相符性, 以人口特徵来量化差异的正确度, 并检视影响邻里定义的因素。我们在两个城市中发现叙述和绘製边界的地理相符性, 且邻里规模不具差异。自我报导的邻里, 相较于较小的单位而言 (例如人口统计单位), 与较大的行政面积较为相符 (例如警察管辖区)。参与者在描绘其邻里边界时, 回报使用了行政定义、步行距离、他们与人和建物的熟悉度、他们在哪消磨时间, 以及物理空间的地标。在纽约市, 参与者同时回报将社会人口特徵和交通纳入考量。我们的方法証实通过网上地图绘製调查蒐集感知的(以自我为中心的)邻里的可行性, 并可适用于诸多研究设定。

Aunque los factores de vecindario han sido asociados consistentemente con la salud, dificultades tecnológicas para sonsacar vecindarios autodefinidos desde cohortes grandes han comprometido el grado de interpretación de esta investigación. Ofrecemos aquí un enfoque de métodos mixtos para sacar y validar vecindarios autodefinidos. Los participantes utilizaron una interfaz adaptada de Google.Maps para “dibujar” su vecindario y respondieron preguntas sobre la exactitud percibida del mapa, la definición del vecindario y las actividades vecinales. Comparamos la concordancia geográfica del dibujo y narrativa de las definiciones vecinales, cuantificamos la exactitud diferencial por las características demográficas y examinamos los factores que influyen las definiciones vecinales. Hallamos una concordancia geográfica similar entre la narrativa y los límites mapeados en dos ciudades, sin diferencias por tamaño del vecindario. Los vecindarios auto-reportados exhibieron concordancia más grande con las áreas administrativas más grandes (e.g., distritos policiales) que con unidades más pequeñas (e.g., secciones censales). Para delinear sus límites de vecindario, los participantes informaron el uso de definiciones administrativas, distancia del recorrido, su familiaridad con la gente y estructuras, dónde pasaban el tiempo, y puntos de referencia físicos. En la Ciudad de Nueva York, los participantes informaron también la consideración de características sociodemográficas y el transporte. Nuestro método demuestra la viabilidad de recopilar vecindarios percibidos (egocéntricos) a través de levantamientos cartográficos online, adaptables a muchos marcos de estudio.

Key Words:

• geographic information systems (GIS)
• mixed methods
• neighborhood scale
• uncertain geographic context problem (UGCoP)

关键词:

• 地理信息系统 (GIS)
• 混合方法
• 邻里尺度
• 不确定地理脉络问题 (UGCoP)。

Palabras clave:

• sistema de información geográfica (SIG)
• métodos mixtos
• escala vecinal
• problema del contexto geográfico incierto (UGCoP)

Additional information

Funding

Funding for this research was provided by a University Center for Social and Urban Research Steven Manners Faculty Development Award, a Central Research Development Fund pilot award from the University of Pittsburgh Office of Research, and the U.S. Environmental Protection Agency (R834576).

Notes on contributors

Jessie L. C. Shmool

JESSIE L. C. SHMOOL was a Research Analyst at the University of Pittsburgh Graduate School of Public Heath, Department of Environmental and Occupational Health, Pittsburgh, PA 15602, during the time when this work was conducted. E-mail: [email protected].

Isaac L. Johnson

ISAAC L. JOHNSON is a PhD student in the Department of Computer Science, Northwestern University, Evanston, IL 60660. E-mail: [email protected]. His research interests include representation learning, GIScience, and algorithmic auditing.

Zan M. Dodson

ZAN M. DODSON is a postdoctoral associate in the Public Health Dynamics Laboratory at the University of Pittsburgh and adjunct faculty at the H. John Heinz III College, Carnegie Mellon University, Pittsburgh, PA 15261. E-mail: [email protected]. His research interests include social ecological processes affecting health, spatiotemporal modeling, and spatial optimization for health care accessibility, and remote sensing applications for urban environments.

Robert Keene

ROBERT M. KEENE is a Database Administrator/Research Programmer for the University Center for Social & Urban Research at the University of Pittsburgh, Pittsburgh, PA 15260. E-mail: [email protected]. His research interests include research software development, survey methodology, user interface design, data visualization, data analytics, research data security, and cloud computing technologies.

Robert Gradeck

ROBERT GRADECK manages the Western Pennsylvania Regional Data Center at the University Center for Urban and Social Research at the University of Pittsburgh, Pittsburgh, PA 15602. E-mail: [email protected]. The Regional Data Center is a collaboration of the University of Pittsburgh, Allegheny County, and the City of Pittsburgh. Through the project, he works to make public information more available, accessible, and useful.

Scott R. Beach

SCOTT R. BEACH is Interim Director and Director of Survey Research at the University Center for Social & Urban Research at the University of Pittsburgh, Pittsburgh, PA 15602. E-mail: [email protected]. His research interests include survey methodology, psychosocial aspects of family caregiving, elder abuse, and technology and aging.

Jane E. Clougherty

JANE E. CLOUGHERTY is an Associate Professor in the Department of Environmental and Occupational Health at the Drexel University Dornsife School of Public Health, Philadelphia, PA 19104. E-mail: [email protected]. Her research interests include urban health, spatial analysis, air pollution exposure assessment, and population susceptibility by social factors and chronic stress.