EAGER: Development of a Novel Online Visual Survey Data Analysis Tool and Assessment of its Capabilities to Enhance Learning of Quantitative Research Methods

点击翻译按钮获取中文摘要

As public opinion and other surveys become a ubiquitous tool in many contexts and disciplinary domains, the need for online survey analysis systems that can be used by non-professionals to quickly and accurately analyze patterns of responses in different subsets of respondents is also increasing. The system developed by the project will vastly expand opportunities for survey data access and online data sharing and exploration in several domains that rely on questionnaire data collection. The ability to annotate patterns of responses and share the annotations with others leads to a valuable new quality in survey analysis, making it a collaborative and interactive online experience. Important benefits are expected in education and learning. Online learning of research methods is often hampered by a deep-seated fear of quantitative approaches reported by many social science majors. The ability to relate inferential results with easy to use visual interactive exploration of social datasets will help students test distributional assumptions, compute models for graphically selected groups of respondents, and find and explain outliers. This combination of features has the potential to transform how survey data are shared and analyzed online, and increase the value of research investment in sociological and related studies by promoting reuse of available data. This project's goal is to build an online tool for exploratory survey data analysis, which can be used by researchers and students to examine, analyze, share, and publish public opinion and other questionnaire surveys. The online tool offers a new visual metaphor for analyzing surveys and intuitively uncovering patterns of relationships, making survey data easy to share, appealing to explore, and simple to navigate from any subsets of respondents to individual cases. The project leverages technical approaches from image analytics, faceted search, and online map navigation, combining them into a novel survey authoring and online publication system which allows users to juxtapose interactive exploration with common inferential statistics, and share and annotate analytical results. The system is evaluated as a learning tool in several research method classes for both graduate and undergraduate students. In addition, it is field tested in a number of surveys conducted by the project and the advisory team members. A key novel feature of the survey analysis system is its efficient and intuitive exploratory data analysis interface accessible by a wide group of expert and non-expert users. Another key novel feature is its support of standards-based data sharing, collaboration, and communication among social survey analysts and students. The project includes development of architecture and information models for a cyberlearning-focused survey analysis system; integration of inferential statistics and exploratory approaches in a single easy to use visual analytical interface; and exploration of survey data sharing, collaborative online analysis and annotations.