The primary results of PER have been met with skepticism by many in the physics community. To remove this skepticism the PER research community must do more thorough, more controlled experimentation.

Education is a profoundly complicated undertaking. Educational systems deal with how human beings think and how those thinking processes evolve. A science class attempts to impart a amazing amount of information that somehow results in the students being able to perform a vast number of new tasks. If one counts tasks by the number of new types of steps that can be found an expert's problem solutions, we estimate that a science class allows a student to perform about 1500 new tasks. The class does this by exposing the new tasks in a variety of environments; lab, lecture, reading, and homework. In total, the student sees perhaps 2000-5000 experiences, assignments, and information chunks that attempt to impart skill in these tasks. A concept inventory requires only 40-50 of these new skills for its solution. This means that an extremely minor shift in the coverage of the class, the selection of 2000-5000 experiences, could greatly modify the concept inventory outcome. Therefore, for educational research to be believable the class being used for the research must be characterized at a level where changes on the order of 40 tasks can be detected.

As the first step in meeting this requirement our research class, University Physics II(UPII), is characterized so that all planned student interactions are captured and available on CD-ROM to anyone wishing to carefully evaluate our research. The idea of a task, called a process in our research, is being refined and the class currently being evaluated to determine how many tasks are present, how they are distributed in class materials, and how student performance is affected by their distribution. The classification should be complete by the end of the Spring 2005 semester. Once classification is complete, UPII will serve as a functioning educational laboratory producing research of a quality commensurate with the best physical science.

One educational laboratory is not enough. We seek partners to help build and fund a network of educational laboratories at a diverse set of institutions. Our characterization scheme is not unique and other better schemes may exist. All student populations are different and all classes different, our results may only apply to our students in our class. Research at many institutions is vital, but that research must be brutal in its attention to detail. The physics education research community has a great opportunity to increase our understanding of education dramatically by applying the precise standards of physics research to the field.