eric.ed.gov har udgivet:
The stability of classroom behavior is examined from several perspectives: (1) the relative consistency of teacher behavior in two different sections of the same course taught concurrently; (2) the relative consistency of student behavior in math and English classes attended concurrently; and (3) differences in student and teacher behavior in math vs. English classes (to determine the effects of subject matter on teacher and student behavior). In general, stability coefficients obtained here were much higher than those expected on the basis of earlier research on stability in courses taught successively rather than concurrently. Even so, high inference ratings were more stable than low inference counts of discrete behaviors, and many behaviors did not occur often enough to allow stable measurement, despite intensive observation. The data are discussed with reference to implementing different treatments in experimental studies in order to document the differential effects, and in reference to the possibility of linking teacher stability on clusters of variables with information about student outcomes. (Author/MV)