A Monte Carlo Simulation of the Intraclass Correlation in Hierarchical Linear Modeling With Missing Data

• In Event: 61.087 - Poster Session 12
  In Poster Session: 61.087-4 - Multilevel Modeling: Complex Survey Data and Missing Data

Abstract

The intraclass correlation (ICC) is a descriptive statistic used in hierarchical linear modeling (HLM) in the early stages of model building. The ICC is useful when determining whether a multilevel approach such as HLM is necessary and provides an indication of the amount of between cluster variation which can be modeled. Findings from a pilot Monte Carlo simulation conducted in the spring of 2014 indicated that the ICC is likely underestimated in the presence of data missing-not-at-random (MNAR). The proposed study aims to extend the earlier pilot simulation by isolating additional missing data conditions under which the ICC may be underestimated and to investigate the effectiveness of various missing data handling procedures for addressing bias in the estimated ICC.

Authors

• Bryden Spencer, University of Northern Colorado

• Niloofar Ramezani, University of Northern Colorado