Accommodating Individual Differences in Software Designed for Location-Based Survey Tasks

点击翻译按钮获取中文摘要

The use of digital maps, imagery and global positioning system (GPS) receivers to support a range of field survey activities is perceived as an area that has potentially high payoff in terms of reduced costs and improved data quality for national surveys. At present, however, appropriate survey methods for effectively using these resources are lacking. The long-term goal of this research is to develop a theoretically grounded yet practical approach for designing computer-assisted survey instruments that rely on geospatial resources to support location-based field tasks. The project will contribute to this goal by developing naturalistic decision-making models for distinct location-based survey tasks that reflect the diversity of strategies that result from individual differences (e.g., in spatial ability), and by developing model software interfaces that accommodate variation in strategies for using computers and integrating environmental and digital geospatial information. The work will begin with building naturalistic decision-making models to describe a survey task referred to as address verification, which involves evaluating whether the location of a housing unit on the ground is accurately depicted on a map. The model will describe how task implementation varies in relation to individual abilities, such as mental rotation of spatial objects, perspective taking, and spatial memory. The approach involves outlining a high-level naturalistic decision-making model to describe cues (information) available to and selected by a field staff member, hypotheses (if any) that are formulated from this information, and the action selected that leads to completion of the task. Using this template, a series of more specific models will be created that describe how variation in spatial ability affects the decision-making process. Expert evaluation and field-based cognitive walk-through methods will be used to further build and refine models. With the understanding gained from these decision-modeling investigations, models of two user-interface views will be developed for the address verification task. The first user-interface view will be a basic interface that provides the required map functionality without taking any user differences into consideration. The second user-interface view will be designed to be an inclusive interface using what was learned in our modeling research. Information on variation in field behaviors (e.g., preferred cues, map use, task order) in relation to spatial ability will be used to create the set of screens for the inclusive user-interface views. Survey field experts and cognitive psychologists will evaluate these draft screens in relation to their knowledge and experience with field staff behaviors. This research is an important first step of an iterative process to develop interface design principles for location-based survey field tasks that rely on an inclusive software design approach to accommodate individual differences. This research will make several substantive contributions to both basic and applied science. First, the project will begin developing a theoretical approach for effectively incorporating geospatial information resources in computer-assisted instruments for field surveys. Through this process, naturalistic decision-making theory will be extended to include heterogeneity in individual abilities, specifically spatial ability. In addition, user interface design approaches will be broadened to handle the context of geospatial data for a task involving scientific protocols. The practical impact of the research will be felt in fielding surveys with improved field operations efficiency and enhanced quality of data and survey estimates for the national statistical infrastructure and other large-scale surveys. This project will provide students in human-computer interaction, computer science, and statistics with a substantive experience with interdisciplinary and inter-organizational research collaboration. The research is supported by the Methodology, Measurement, and Statistics Program and a consortium of federal statistical agencies as part of a joint activity to support research on survey and statistical methodology.