Experiment Design: Motivating metadata contributions for data re-use and reproducibility.
Presenter
Doctoral Candidate at the School of Information
Time and location
1:00-2:00 pm on Zoom: https://zoom.us/j/320085151
Abstract
The design is at its early stage and we welcome all comments and suggestions.
Background: what is metadata about? Metadata of a dataset is “data about data”. Study-level metadata usually includes various descriptive information of the study that generates the dataset, and variable-level metadata contains documentation for variable labels as well as value labels. High quality metadata is indispensable for data re-use and reproducibility of the studies.
How do we motivate researchers to provide better metadata after they have published their papers? We plan to leverage social preference and social information (role models) to increase researchers’ contribution to metadata, which are stored as data deposits for their published paper. We will introduce a novel randomization scheme to rule out potential spillover effects governed by the observable networks. We also would like to discuss alternative designs that can motivate metadata contributions to all existing data deposits.