For a project to require Institutional Review Board review, it must 1) constitute research, and 2) involve human subjects, as defined by the Code of Federal Regulations (see list of detailed definitions).
Projects conducted by students for educational purposes within a course alone do not fall within the purview of the IRB, as they do not fit the federal regulations definition of research. Such projects occur in a course designed, at least in part, to provide training in research methodology. These projects are commonly conducted as part of a research methods course or research practicum, which involve the supervised practical application of previously studied theories of research methods.
To the extent that students will be interacting with human subjects, regardless of whether their project meets the Office of Human Research Protection's definition of human subjects research, care must be taken to ensure that those subjects are treated with respect and courtesy, do not have their privacy invaded and are not subjected to unnecessary discomfort (physical or emotional).
It is important that the instructors fully understand the classroom projects that are proposed and conducted by their students, and that they provide clear and unambiguous guidelines to those students with respect to their interactions with their subjects.
If, however, a student plans to use the data collected from a course-related project for future human subjects research activities (as a pilot study, for example), the project will require IRB review before being undertaken for the class, as IRB approval is not retroactive.
Independent projects conducted by students that include systematic data collection from human subjects and that are intended to produce generalizable results are considered research as defined by the federal regulations and so do fall within the purview of the IRB. Such projects include honors theses, graduate theses and doctoral dissertations.