Decision based learning for a sophomore level thermodynamics courseAn electronic tutor was designed and used to study the effectiveness of a new pedagogy calledDecision Based Learning. Decision based learning (DBL) has similarities to existing activelearning methods, but differs in the following important ways: 1) learning consists of developinga general to specific decision set that students can use to solve novel problems 2) students practicemaking the decisions in the instructor decision set, and are given help with understanding. Thegoal of this method is to improve student understanding through the process of decision makingso that they have a better chance of solving novel or complex problems.Students were asked to draw a T-v phase diagram for a refrigeration cycle they had never seenbefore. At any point in the drawing process, students could submit their work to receive feedback.Upon submission, the tutor evaluated the student drawing to determine the most general/importantunderstanding for which the student needed help. Rather than showing the correct answer, the tutorasked additional thought questions designed with the attempt to improve the student’sunderstanding to the point where he or she could make a decision. The decision making would, inturn, advance the student toward a correct solution. Students were asked to work at least 40minutes on the activity.Students completed a pre-tutor-test and a post-tutor-test for the purpose of determining the impactof the tutor in furthering students’ understanding regarding (P,T,v) property relationships forthermodynamic components. A significant amount of learning was demonstrated using DBL assuggested by a Cohen’s d=1.77 for 88 students, where d>0.8 shows a large effect. The pre-testresults indicate that on average only 25% of students were able to identify all three relations forcomponents, before the activity. Liquid gas separator, evaporator and condenser were thecomponents that were most misunderstood by students. The post activity test showed significantlearning for component relations. As an example, 58% of the students had a misconception aboutpressure for heat exchangers. This was reduced to 18% using a single activity. More detailedanalysis investigated how students learned using this activity.