Chinese researchers have long preferred substantive research (i.e., examination of relationships among variables) to construct validation research, while failing to realize that construct validation research is the foundation of substantive research. This paper endeavors to answer a question: Do Rumelt’s (1974) categorical and Standard Industrial Classification (SIC) and skill-based continuous measures capture the same construct of diversity in the context of China? We take two steps to investigate the question. First, are continuous measures (SIC & skill-based entropy and Montgomery product count) appropriately differentiated among Rumelt’s categories? Second, are the categories converted from SIC & Skill-based measures convergent to Rumelt’s categorical measures? Results show that the two measures are consistent with each other.