Aware of the need for changes in physics instruction at my previous university (UFJF), I started researching physics education. From my personal experience, I believe that we learn by becoming a member of a community of practice where newcomers model their work after that of experts, gradually learning to use shared vocabularies, techniques, and concepts. Those beliefs led me to research active engagement and inquiry based methods that would allow students be members of small groups that reproduced the dynamics of the scientific community. Our research showed a significant improvement in students' conceptual understanding and logical reasoning, and led to a widespread change in introductory physics instruction at the Federal University at Juiz de Fora [6, 7].
I have also been interested in students' epistemological views. Informal interviews with physics pre-service and in-service physics teachers suggested a fragmented epistemological view, with three distinct knowledge systems about physics: (i) the "college-level physics" needed at exams, (ii) the "high-school physics" taught at schools, and (iii) the "real physics" applied to the everyday phenomena. The difference between those physics was not one of difficulty or sophistication, but of different concepts and law. For example, even though a student correctly used Newton's Third Law in his exams, this same student taught his pupils, incorrectly, that weight and normal forces were Third Law pairs. Therefore, this epistemological fragmentation poses a problem to departments involved in teaching certification programs.
To better understand how students form their epistemological views, we started a program of mapping students reasoning skills and their relationship to epistemological views. We gave physics introductory students a series of tests evaluating their epistemological views and reasoning skills. For epistemological views, we used the Colorado Learning Aptitudes about Science Survey (CLASS) and the Maryland Physics Expectations (MPEX). For reasoning skills, we used Tobin's Test of Logical Thinking (TOLT) and a test we created to examine student's abilities to reason with formal logic. We are in the middle of the data collection stage, but preliminary analysis shows almost no correlation between epistemological views and reasoning [8]. We also found out that, for senior students, their reasoning skills didn't change significantly, with substantial problems with common logical fallacies, but their epistemological views approached that of an expert. However, their "expert" epistemological views measured the above tests did not prevent those students from having an actual fragmented epistemology of physics. It is our interest to investigate these matters further and to understand how we can improve students actual epistemological beliefs.
More recently, I am concerned about how to teach science to pre-service teachers in a way that demystifies it. In particular, more and more people see science as an indoctrination (a la Feyerabend), and, as any indoctrination, not necessarily carrying an element of truth. I believe that science teaching needs to address this misconception that science is just for a few initiated scholars, and not for everybody, by focusing on the process by which science constructs knowledge.