ABSTRACT
Whenever a researcher is presented with data, he or she seeks to discern the existence of any patterns in the data. The nature of the data typically determines the type of analysis to use in discovering these pattern(s). For example, a survey on attitudes toward educational reform might contain open-ended questions. The resulting (qualitative) data would be subjected to a content analysis (also known as thematic analysis) in order to identify any themes (i.e., patterns) in the responses. Similarly, any questions on the survey using a Likert response format (i.e., quantitative data) could be analyzed using a mathematical model that attempts to account for or explain the variability in the data (i.e., patterns) in terms of one or more variables. These models can broadly be classified as either linear (e.g., the structural model for an analysis of variance) or nonlinear (e.g., the logistic regression model) in parameters.
