Abstract:

One of the critical assumptions of conventional statistical methods is the independence among variables. Relational data, however, violate this prerequisite. Then how to measure relationships in relational data? QAP is an applicable solution. This paper introduces the rationale of QAP and exemplifies its application in correlation and regression with UCINET. QAP is unique in that it can not only be applied to the regression of one relational variable on another relational variable but also to correlations among relational and attribute variables.

Key words: QAP, Relational data, Social Network Analysis