RESEARCH ON AUTHENTIC ASSESSMENT USING A VIRTUAL WORLD FOR LEARNING GEOLOGY

The Geology Explorer (http://oit.ndsu.edu) is an immersive virtual environment where learners take on the role of geologists on an expedition to a mythical, newly discovered Earth-like planet (Planet Oit). Students are given authentic geologic goals, e.g., to locate a particular mineral found in a geologically reasonable context, or to map and interpret the geologic history of a metamorphic terrane. Accomplishing such goals entails becoming adept in several geologic concepts and procedures, and demonstrates student mastery of the material.

We are researching the use of scenario-based assessment to gauge student learning in this type of environment. These assessments are a type of "authentic assessment," which means that student performance on intellectual tasks is directly examined. The general protocol is to create matched groups (Planet vs. Control) and pose an authentic open-ended, commonplace problem-solving situation (a 'scenario'). Scenarios which differ in detail, but which are designed to test the same concept are posed to students both before and after their experience in the virtual environment. The student is asked to provide a narrative solution, as well as a list of questions relevant to the problem. Student responses are evaluated, according to a pre-determined protocol, by content experts looking for evidence of problem solving technique and intuition. The scoring protocol stresses the importance of reasoning over the recitation of facts. Scores are given in reference to an ideal answer a professional geologist might provide.

Over the past seven years, 12 scenarios in a variety of pre- and post- experience combinations have been applied to over 2000 students taking introductory Physical Geology at NDSU. Statistical analysis of performance of the matched groups shows that the virtual environment positively affects student learning. Gender was significant only in one year (with females obtaining higher scores) and not significant for the other years. An area of ongoing research is the highly significant statistical effect of scenario combination – the choice and order of scenarios presented pre- and post- scenario.