Didaktik21 – wissenschaftliches Programm
DD 21.2: Vortrag
Dienstag, 23. März 2021, 10:40–11:00, R2
Fails & Fixes for Data Encoding - Estimating Encoding Reliability — •Florian Genz, Kathleen Falconer, Jacob Beautemps, Lars Möhring, and André Bresges — Universität zu Köln
"Data encoding is often neglected or taken for granted, even though this step forms an important bridge between the data collection and analysis." (Springuel, Wittmann & Thompson, 2019) How can I estimate the encoding quality of my research data? Even complete double-coded or digitally collected data can profit from further layers of error redundancy. We present our biggest fails in data encoding and transformation as well as techniques for fixing them. We will do this exemplarily by giving insights into the development of the new Flight Physics Concept Inventory (FliP-CoIn). We describe how to estimate encoding errors in your research data, probability of remaining scoring errors, and influence on the analysis. We close with best-practice advise for secure and redundant data encoding. We will elaborate our methods to an extent they become practically useful for all kinds of publications dealing with huge quantitative data.