Properly prepared and properly presented graphics often provide highly informative visualization of statistical analysis and results with education data. The applicability of graphical representations is shown, for example, through the use of repeated traces by van den Bergh et al. (see Chap. 20) in the discussion of time series analysis in this book. This chapter demonstrates how methods of statistical graphics can be applied to the study of science, literacy, and other areas of education research. Multivariate data analysis is a course commonly offered in statistics undergraduate and graduate programs. There are many textbooks; most discuss ways to plot multivariate data in some form or another; usually a chapter is devoted to plotting methods. Unfortunately, most textbooks still focus on old methods, such as Chernoff faces, star plots, and Andrews curves, which have not been used seriously since the 1970s. A few textbooks give the topic some cursory attention and include material on static methods that are actively used, such as scatter plot matrices, trellis plots, and parallel coordinate plots. Not a single textbook discusses dynamic methods like multiple linked plots and tours. Yet these methods were invented in the early 1980s; they are commonly available in today's software and provide insight into data and theoretical concepts of multiple dimensions. Plots of multivariate data are very important-more so than for univariate data because multivariate data have more complexity and the distribution theory is less developed.
|Title of host publication||Quality Research in Literacy and Science Education: International Perspectives and Gold Standards|
|Number of pages||19|
|Publication status||Published - 2009|