Abstract:Hierarchical Linear Models have been widely used to analyze multi-level datawith “nested” structure. In Hierarchical Linear Models, we take the “low level”covariance into account in estimating the “high level” parameters. That is thereason why Hierarchical Linear Models can be used to deal with the “nested”data set, in which the “identity of variance” assumption has been destroyed. Inthe thesis, to explain why the Hierarchical Linear Model does better than OLSestimation in estimating the coefficient of interaction variables, a consumerdata was used.