ABSTRACT
The term panel model is a generic term for a wide variety of clustered, nested, and longitudinal models. In fact, this chapter could well have been titled Modeling Clustered Data . The problem with clustered data is that they are not independent and therefore violate an essential criterion of maximum likelihood theory-that observations in a model are independent of one another. Clustered data, depending on how they are generated, tend to bias standard errors as well as parameter estimates.
