Reference no: EM133336300
1.
Motivation for Analyzing Change Many of the most interesting research questions in education and the social sciences involve the measurement of change. Often educational researchers want to understand how people develop or change over time. As Willett (1988) rightly points out, "the very notion of learning implies growth and change" (p. 346). Thus, questions about learning rates, trends across time, and growth or decline in some areas implicitly pose questions that involve the measurement of change. For example, to understand how students' skills develop across time, we need to consider several key issues: What is the shape of this growth trajectory? Is this growth steady or does the rate of growth change over time? Do people tend to change in the same way over time? Or is there a great deal of variability between people in terms of their rate of change over time? In addition, identifying factors that help to predict the rate at which change occurs, or which variables help us to understand interindividual differences in the rate of change is often critically important. In higher education, there are many research questions that involve change over time. Do student grade-point averages increase over time? Do the grade-point averages of males and females increase at the same rate? To what degree has the diversity on college campuses increased over the last two decades? Why are some universities more successful than others in increasing the diversity on their campuses? Such questions are best answered using analytic methods that explicitly model change over time.Over the past two decades, individual growth models have become one of the most common methods to analyze change. This chapter introduces three of the most common of these models and highlights common issues in the analysis of change. Given the complexity of the topic, this brief chapter serves as an amuse-bouche, rather than a full treatment of the topic. Readers who are interested in pursuing such models in more depth are encouraged to consult Bollen and Curran (2006), Duncan et al. (2006), Little (2013), McArdle (2015), or Singer and Willett (2003).Why Do We Need Growth-Curve Modeling?Before embarking on a journey into individual growth-curve modeling, it is important to understand the inadequacies inherent in using two wave studies to measure change. The simplest type of change measure, a difference score, is literally the difference in two scores between pretest and posttest. Many researchers model the difference between posttest and pretest scores as a function of the presence of a treatment or some other relevant variable. Although difference scores are simple to calculate, they are inadequate measures of growth (Cronbach and Furby, 1970).
As you reflect on this reading, which topic or issue resonates the most with you? Why?
What big question remains unanswered? The question can be about something that you want more information about, or it can be something that takes the conversation in an entirely different direction.
What did you gain from this reading?
Include your own experience as well.