Abstract: Hierarchical Linear Model (HLM) is a new tool for analyzing hierarchical structure data. It is superior to traditional statistical methods in that HLM based hypothesis is more compatible with facts and that the results generated from the model are reasonable. Therefore, HLM has been applied more and more wildly in social science. This paper is an introduce to HLM which introduces the fundamental hypothesis of the model, parameter estimator and hypothesis testing methods, as well as its advantages over the traditional regression method in dealing with hierarchical structure data.

Key words: hierarchical linear model, regression analysis, least square estimation, iterative least square estimation.