This presentation focuses on measurement error in survey data to ensure validity and reliability by firstly clarifying core concepts of validity, reliability, and measurement quality, exploring estimation approaches of specification and measurement errors (validity and reliability estimation, multi-trait multi-method and hidden markov models) and covering tools for assessing measurement quality before data collection begins (survey quality predictor). Apart from that similarities and differences of error sources in digital trace data compared to survey data are identified, and traditional tools are adapted to address both the advantages and challenges of applying traditional tools in the context of digital trace data. The concept of the “multiverse of measurements” will be introduced. Challenges posed by the non-linearity, non-normality, and heterogeneity of error processes in digital trace data are also addressed. This presentation is the third part of a set of presentations on the topic of "Indicators & Metrics of Data Quality". Presentation available after free registration: https://elearning.gesis.org/course/section.php?id=1211#module-3192
Peter Lugtig (Wed,) studied this question.