Data Aggregation in Multilevel Research: Best Practice Recommendations and Tools for Moving Forward

The multilevel paradigm is omnipresent in the organizational sciences, with scholars recognizing data are almost always nested – either hierarchically (e.g., individuals within teams) or temporally (e.g., repeated observations within individuals). The multilevel paradigm is moored in the assumption that relationships between constructs often reside across different levels, often requiring data from a lower-level (e.g., employee-level justice perceptions) to be aggregated to a higher-level (e.g., team-level justice climate). Given the increased scrutiny in the social sciences around issues of clarity, transparency, and reproducibility, this paper first introduces a set of data aggregation principles that are then used to guide a brief literature review. We found that reporting practices related to data aggregation are quite variable with little standardization as to what information and statistics are included by authors. We conclude our paper with a Data Aggregation Checklist and a new R package, WGA (Within-Group Agreement & Aggregation), intended to improve the clarity and transparency of future multilevel studies.