[영문]Multi-level structures are common in social work administration research, where individuals are hierarchically nested within organizations. Hierarchical linear modeling (HLM) is a statistical method developed to address issues associated with the unique nature of multilevel data. We introduce the basic notion of HLM, highlighting the need for using this technique in multilevel data. We also present analytical procedures, illustrating the major issues involved in the application of HLM to social work administration research. Currently, there has been an increasing interest in promoting a multilevel approach in social work administration research. To comprehend and benefit from the results of research dealing with the multilevel approach, administrators as well as researchers are expected to understand the basic logic of HLM. This paper delivers sufficient practical knowledge for research consumers to appreciate the results of research that has employed HLM, especially for those consumers who are not familiar with advanced statistics.