Medium Node: Reducing Error in Computerized Survey Data Collection

调查数据收集成本的上升和研究人员对调查数据质量的日益关注，要求开发创新的数据收集方法。这项研究将着重于改进从计算机辅助方法收集的调查数据。该项目将包括与互联网数据收集和计算机辅助电话采访（CATI）数据收集系统有关的研究，以及这些计算机辅助数据收集系统的变体。该研究的总体目标是提高调查所得数据的质量，重点是实现三个目标。首先，该研究将评估四种诊断工具的使用情况，以确定计算机辅助的、采访者管理的数据收集工具的测量误差，并将利用这些发现为计算机辅助的、采访者管理的数据收集工具的共同特征的视觉重新设计提供信息。其次，该研究将评估自适应/响应式设计的使用，其中收集数据的动态建模用于在收集数据时修改问卷。第三，本研究将评估基于日历和时间日记的数据收集方法的应用，通过根据个人受访者的需求定制问题来帮助行为自我报告的准确性。这些目标下的任务将是综合的，允许研究人员改进调查设计/工具，并通过减少访谈者和受访者的负担来提高数据质量。跨学科的项目团队包括统计学、心理学、社会学、调查研究和方法论以及计算机科学方面的专家。该团队将利用与行业领导者（盖洛普和Abt SRBI）以及普查局的长期合作和合作伙伴关系来实现其目标。这项研究将促进调查方法和相关领域的科学知识。减少调查数据中的测量误差对于从调查数据中准确推断至关重要，并且确定最小化或减少调查数据中的测量误差的新工具将改进跨领域调查得出的结论。除了促进调查数据收集的发现和理解外，该项目还将通过实施一项教育计划，为培训未来的调查研究专业人员做出重大贡献，该计划旨在：(1)从相关社会科学和统计学科招募和留住多样化的以方法为导向的学生群体；(2)将研究整合到UNL现有和新课程中；(3)为所有参与的学生提供实践研究的机会；(4)培养学生科研人员的专业技能。这些结果将在多学科和跨学科的会议和讲习班、联邦统计机构和调查行业的研究人员中广泛相关和广泛传播。这项活动是由nsf人口普查研究网络资助的机会。

Rising costs of survey data collection and researchers' growing concerns about the quality of survey data necessitate the development of innovative approaches to data collection. This research will focus on improving survey data collected from computer-assisted methods. The project will include research related to internet data collection and computer-assisted telephone interviews (CATI) data collection systems, along with the variants of these computer-assisted data collection systems. With an overall goal of improving the quality of data derived from surveys, the research will focus on accomplishing three objectives. First, the study will evaluate the use of four diagnostic tools for identifying measurement errors in computer-assisted, interviewer-administered data collection instruments, and it will use these findings to inform visual redesign of common features of computer-assisted, interviewer-administered data collection instruments. Second, the study will evaluate the use of adaptive/responsive designs in which a dynamic modeling of collected data is used to modify the questionnaire as the data are being collected. Third, the study will evaluate the application of calendar- and time diary-based data collection methods to aid in the accuracy of behavioral self-reports by tailoring questions to the needs of individual respondents. The tasks under these objectives will be integrative, allowing researchers to improve survey designs/instruments and enhance data quality by reducing interviewer and respondent burden. The interdisciplinary project team includes experts in statistics, psychology, sociology, survey research and methodology, and computer science. The team will leverage long-term collaborations and partner with industry leaders -- Gallup and Abt SRBI -- as well as Census to accomplish its goal and objectives.This research will advance scientific knowledge in survey methodology and related fields. Reducing measurement errors in survey data is critical to accurate inference from survey data, and identifying new tools that minimize or reduce measurement errors in survey data will improve conclusions made from surveys across fields. In addition to advancing discovery and understanding in survey data collection, the project will contribute significantly to the training of future survey research professionals through implementation of an education plan designed to: (1) recruit and retain a diverse methodologically-oriented student pool from related social science and statistical disciplines; (2) integrate research into existing and new curricula at UNL; (3) provide all participating students with hands-on research opportunities; and (4) develop student researchers' professional skills. The results will be broadly relevant and broadly disseminated at multidisciplinary and interdisciplinary conferences and workshops, to the federal statistical agencies, and to researchers in the survey industry. This activity is supported by the NSF-Census Research Network funding opportunity.