Rejoinder to Comments on Task Features in Simulations and Games

First, we would like to thank all of the people who took the time to write commentaries on our article. There is a rather famous map of the United States drawn by Daniel K. Wallingford from a New Yorker’s perspective and a similar one drawn by Mark Storm from a Texan’s perspective; each has the home state much larger than the rest of the country. The humor in these maps comes from the natural tendency of people to know things close to them in great detail but to only have sketchy knowledge of things that are far away; that is, our natural view of any topic is like a fisheye lens with things in the center magnified and things at a distance compressed. The fish-eye lens also applies to assessment design: Psychometricians, test developers, domain experts, computer programmers, game designers, test users all have their unique view of the test design process. Evidence-centered assessment design (ECD) looks different from these different perspectives, and it is hard to put the whole picture together without looking at ECD from many directions. Therefore, we are truly grateful to have such a wide array of perspectives on our article. One thing that was obviously not clear in the original article was the purpose of including both the mathematics word problem and the game examples. Some of the earliest ECD presentations by Mislevy, Steinberg, and Almond1 included a triangle. The 3 vertexes were purpose (formative versus summative), score reporting model (unidimensional versus multidimensional) and task type (multiple choice versus simulation based). The key idea was that changes in any one vertex would necessitate changes in the others as well. In particular, at the time (late 1990s) both simulation-based assessment and cognitively diagnostic assessment were becoming more popular and would induce changes in one (and as a consequence all) of the vertexes. ECD was