Abstract
Correspondence analysis is a statistical method that allows researchers to explore relationships among complex categorical variables. This paper will provide researchers with the theoretical and practical foundations for understanding and applying correspondence analysis to their own research agendas. Problem: Technical communicators use a variety of research methods and collect a variety of types of data. Of particular interest to technical communicators is categorical data, or data that are not traditionally quantitative. For instance, technical communicators often collect and analyze language data from a variety of texts. Analyzing this type of data can be difficult using traditional statistical methods. Key concepts: Variable types, a priori versus exploratory research designs, contingency tables, and data visualization are central to understanding the foundations of correspondence analysis. Key lessons: To conduct correspondence analysis, a researcher must walk through a series of steps including: (1) determining whether correspondence analysis is appropriate, (2) choosing a statistical software package, (3) running the correspondence analysis, and (4) interpreting and applying the results. Implications for practice: While correspondence analysis provides many useful insights into categorical data, a researcher must consider several things when deciding to use correspondence analysis. These include the potential to misinterpret and misapply the results of a correspondence analysis.