Online Program

Saturday, February 22
PS3 Poster Session III & Continental Breakfast Sat, Feb 22, 7:30 AM - 9:00 AM
Bayshore II-IV

Conceptualizing Statistics: A Heuristic Approach (302825)

*Heidi Reichert, University of Michigan 
Kathy Welch, University of Michigan 

Keywords: Communication, communicating, multilevel models, random effects, heuristic, teaching statistics, learning statistical concepts

Communicating statistical concepts can be a challenging task for applied statisticians and often requires more than one strategy. A heuristic approach, which encourages self-learning through discovery, can be a useful tool for getting complex ideas across to researchers and others. A heuristic approach conveys a reasonable approximation of the method, without providing an exhaustive technical explanation. In this presentation, we illustrate a heuristic approach to explaining random effects in a multilevel model. By using a thoughtful combination of real-world data, illustrative graphs, and pointed questions, the fundamental concepts behind random effects are revealed. This technique has been applied in a graduate class in longitudinal/clustered data analysis for non-biostatistics public health students.