Data quality in the context of longitudinal research studies

Abstract

This paper discusses the concept of data quality in the context of longitudinal research. By deconstructing quality assurance process and data collection strategies through a case study of the "Croatian Birth Cohort Study", we try to define causes and sources of poor data quality in the context of longitudinal studies. Besides the problems discussed throughout the known literature (panel conditioning, sample attrition, recall bias, temporal and financial demands), we introduce singlesource problems, multi-source problems, security problems, design questionnaire problems and QA workflow problems as important aspects in the domain of the possible sources of errors. Additionaly we propose models for eliminating the errors through prevention and detection in order to improve data quality