In multilevel analyses, the aggregation of lower-level data to a higher level is commonly justified via rWG and ICC(1) indices. In this article, the authors analyze the relationship between these two measures theoretically and by means of Monte Carlo simulations. Simulation results show that only ICC(1) affects decision quality in terms of statistical power, whereas the height of the mean rWG does not impact power estimates. The authors discuss ambiguous settings in which common interpretation of rWG and ICC(1) values are inconsistent and suggest that researchers interested in analyzing the adequacy of aggregating data to a higher level should primarily employ the ICC(1) and only use the rWG as an additional source of information.
|Veröffentlicht - 2010
|70th Annual Meeting of the Academy of Management - Dare to Care: Passion and Compassion in Management Practice and Research, AOM 2010 - Montreal, QC, Kanada
Dauer: 06 Aug. 2010 → 10 Aug. 2010
|70th Annual Meeting of the Academy of Management - Dare to Care: Passion and Compassion in Management Practice and Research, AOM 2010
|06.08.2010 → 10.08.2010
ASJC Scopus Sachgebiete
- Technologie- und Innovationsmanagement