Document Type

Conference Proceeding

Publication Date



This study continues prior work of the investigation and modeling of Incubation Effect, a phenomenon in which a momentary break helps the generation of a solution to a problem, among students using in a computer-based learning environment called Physics Playground. This paper attempts to improve the detection of IE-False by identifying notable features among instances of unsuccessful incubation by using a combination of t-SNE dimensionality reduction and x-means clustering techniques. We found that there are overlaps on some characteristics of IE-True and IE-False incidences but discovered features that do not make a break beneficial which are low success rate prior to post-incubation, too many levels played during the incubation phase even if some of these are similar to the unsolved problem, a lengthy incubation duration, and too much attempts on the level which has been previously related to frustration.