The rank of each Center according to grade level gain or loss in Table 3 follows exactly the rank order of Centers in Table 2. But Table 3 shows exactly how much of the average growth per Center in Table 2 is accounted for by teacher-school effect. Note, teacher-school effect cannot be separated from the High Intensity Learning System, so that the most accurate interpretation of Tables 2 and 3 together must be stated in such form as:

Using a High Intensity Learning System pedagogy in Center A, .99 of the 1.73 average growth--about 57% of the average gain--is accounted for by the interaction of High Intensity with teacher-school effect." If, for example, Mr. Smith's personality is warmer" than Miss Jones', and Smith ran Center A while Jones ran Center V, and those personality traits were in fact the main components of teacher-school effect, then .99 of the 1.73 gain in Mr. Smith's Center was the effect High Intensity has when Smith used that pedagogy.

On the other hand, .49 loss was the effect of Jones' personality when she

ran a High Intensity Learning Center, leaving a .50 gain after her personality took its toll. The data in Table 3, however, do not tell us exactly what the factors were in teacher-school effect. For that information, we must analyze the constraint measurements presented below. Those constraint measurements already tell us that the major components of teacher-school effect in Centers W and X were tests that could not detect gains.

The "System" and Teacher-School Effect

One of the ultimate tests of a "systems approach" to any endeavor is the attempt to skew the results so that a minimum level of negative results is attained regardless of personality variables. Specifically, in education, a "true" system insures a certain level of results regardless of teacher-school effect. In a sense, a systems approach to instruction reduces teacher-school variability in one direction (negative) only, leaving variability open in the positive direction. Tables 2 and 3 show this "systems effect" dramatically.

The worst" Centers (T, U, V) had .67, .58 and .50 average years gain This is not only above national norm expectancy of .45, but significantly above the .25 gain usually reached by this school population.

in .45 years.

At the

In other words, given a teacher-school effect with HILS that cut .31, .33 and .49 off the mean gain that the 2102 students as a total group made, the students in the worst Centers" still achieved above expectancy. other extreme, given a system that reduced variability at the lower end (below the average gain for the total population of 2102), the same system opened up the upper end where teacher-school effect with HILS increased average gain from 29 to .99 years above what the total population average showed. In a sense, the systems approach accomodates individual teacher-school differences, but reduces the "negative" effects of these differences.