This excess correlation might manifest itself in clustering – groups of observations that are more similar to one another than to the remaining data. This excessive correlation can indicate an incomplete model – predictors or interactions omitted, the need for transformations, or the presence of outliers.
week #9 – Overdispersion
- January 30, 2015
- , 11:02 pm
In discrete response models, overdispersion occurs when there is more correlation in the data than is allowed by the assumptions that the model makes.