Introduction
Several authors cite Heller and his colleagues (1982) as the source of a response to treatment strategy for identifying students in need of special education. The Heller Report was a response to equity concerns in the implementation of special education national policy. I criticized the Heller Report's recommendations for a "two-phase comprehensive assessment" as simply urging the educational system to try harder, to invest more effort in students who were difficult to teach and manage. I believed then, and no evidence has persuaded me otherwise, that such an approach disregards the fact that individual differences in responsiveness to instruction are not in any sense inside students. Rather they are aptitude-treatment interaction effects (Reynolds, 1988). Responsiveness to instruction is embedded in but not separable from a complex educational context that includes institutional as well as teacher variables.
Because teachers are conscious agents who mediate between available instructional resources and innate individual differences (see Figures 1 & 2), they learn to observe and value responsiveness to their instructional efforts. I call the range of students whom they come to view as adequately responsive - i.e., teachable - as the tolerance; those who are perceived to be outside the tolerance are those for whom teachers seek additional resources. The term "tolerance" is used to indicate that teachers form a permissible boundary on their measurement (judgments) in the same sense as a confidence interval. In this case, the teacher actively measures the distribution of responsiveness in her class by processing information from a series of teaching trials and perceives some range of students as within the tolerance.. "The validity of such a tolerance is socially and historically constructed, not psychometrically derived (Gerber, 1988, p. 311)."
A theory of tolerance focuses attention on the dynamics of classroom instruction and makes both economic and cognitive arguments about differences in responsiveness to instruction as the source of teachers' perceptions of "discrepancy" and, therefore, referral for special education (Gerber, 1984; 1988; 1995; Gerber & Kauffman, 1979; Gerber & Semmel, 1984; 1985). In these analyses, the actual intra-individual differences that we hypothesize as the source of "learning disability" remain hidden. Only their expression in transaction with instruction - as with any behavioral assessment -- can be observed and measured. We formalized these concepts in microeconomic terms that were grounded in the transaction that occurs during instruction - response from student, interpretive adjustment by teacher (Gerber & Semmel, 1985). We attempted to show why and how "referral" or "re-integration" (i.e., inclusion) were conditioned by teachers' conscious, contingent, or forced choices under conditions of resource scarcity and relatively fixed technologies of teaching (see Figures 1 & 2). Moreover, the surrounding organization of the school creates and maintains patterns of resource allocation that facilitate or constrain teachers' ability to be responsive to differences.
Figure 1. Variable student achievement as a function of instructional effort
Figure 2. Joint optimal achievement possibilities showing teacher decision point(P) for two-student instructional grouping
Recently, much of the general approach marshaled by Heller and his associates has reappeared in a series of papers that seek to promote a "resistance to intervention" 1 strategy as a more valid means to identify students with learning disabilities (Fuchs & Fuchs, 1998; Fuchs, Fuchs, & Speece, 2002; Gresham, 2002; Torgesen, 2001;Vaughn & Fuchs, 2003). I will argue in this paper that these newer incarnations of Heller et al. advance a more sophisticated measurement argument but are flawed in similar ways. Moreover, I argue that these new proposals send a mixed message about identification of learning disabilities, one that mistakes the measurement of the construct from the construct itself, a logical error reminiscent of the circular reasoning that promoted the idea that intelligence was what IQ tests measure. There is mounting evidence that learning disabilities - problem best conceived as a "spectrum" disorder like autism - have a material (i.e., neurological) basis and can be directly diagnosed (e.g., Paulesu, De'monet, Fazio, McCrory, Chanoine, Brunswick, Cappa, Cossu, Habib, Frith, & Frith, 2001).
Fuchs and his colleagues do suggest that there are "problem-solving" as well as "standard-protocol" models for using response to instruction (RTI) 2 as a means for identifying students (Fuchs, Mock, & Morgan, 2003). However, most authors dismiss these on psychometric grounds 3 and have assumed or explicitly described what amounts to standard-protocol procedures. Their apparent logic is that if classroom instruction is inadequate or somehow variable, then responsiveness, too, may vary, but not as a function of true underlying learner differences. Therefore, a formal procedure of highly controlled, increasingly intensive instructional trials is necessary to partition variance attributed to instruction and variance attributable to learner differences. That is, maintaining teacher behavior within some limits is necessary to reduce measurement error. Gresham (2002), for example, states that an RTI approach to eligibility ...identifies students as having a learning disability if their academic performances in relevant areas do not change in response to a validated intervention implemented with integrity (p. , emphasis added).
Somewhat more elegantly, L. Fuchs (2003, p. 175) proposes that "the "intervention" used to test responsiveness is a general education that has demonstrated "efficacy for the vast majority of students (emphasis added)." In contrast to this sort of "effective" general education, Fuchs suggests that: "... ineffective classrooms reveal low growth rates across many students, making detection of unrespons iveness difficult but signaling school personnel to intervene at the classroom level (p. 175)."
She goes on to suggest that it is the failure of some students in classrooms in which most students succeed that "reveals some underlying deficit." She adds "...relying on the general education program for [intervention response assessment] permits a full normative framework. That is, student responsiveness to effective general education can be estimated for all students so that a normative profile can be generated to describe the full range of performance (p. 176)."
When construed in this way, RTI constitutes a standardized measurement of individual differences based on a behavior sampling that is more applied but not different in concept from other standardized tests. That is, learning -- i.e., responsiveness -- under specific instructional conditions -- i.e., relevant, validated, and implemented with integrity -- measures the degree of difference that must exist among learners that is specific to ability to learn. RTI, therefore, when implemented according to rigorous and standardized protocol, aspires to be a reliable and valid measurement of learning disability.
There is not only a psychometric, but also an experimentalist logic to this formulation. That is, the conditions of measurement must not only be standardized (i.e., "implemented with integrity"), but also the conditions as a whole must logically exclude any other variables as plausible explanations for variations in learning outcomes. If differences in teachers' instructional behaviors during and between instructional episodes can be an important source of variance in outcomes for a single student, not to mention for different students, RTI must seek to exclude teacher differences as a competing explanation of "resistance to treatment." However, the degree to which instruction in RTI is reduced to simpler forms to maximize reliability and fidelity is the degree to which it departs from ecological validity and risks having only a trivial relationship to education in its broadest sense.
These ideas are not new and pre-date the Heller Report. In 1963, for example, John Carroll (1963) published a seminal theoretical paper on the nature of school learning in which he attempted to account for the embeddedness of individual differences. Carroll formalized the concept of learning as time actually spent in learning divided by time needed for learning. Since Carroll's paper, it has been amply demonstrated that time needed to learn (TNL) a criterion task is a better predictor of longer term learner differences than IQ and other status variables (Gettinger, 1979; 1983; 1984a, b, c) and can differentiate students with learning disabilities (Gettinger, 1991).
1 Alternative terms include "response-to-treatment," "response-to-instruction," "responsiveness-to-intervention."
2Various terms are used. They are treated here as synonymous and referred to as response-to-instruction (RTI) in this paper.
3 Many authors appear to understand that teachers' judgments are consequential (e.g., Fuchs, 2003; Speece, Case, & Malloy, 2003) but only after a baseline fidelity of treatment can be established.
Previous Page | Next Page
(Table of Contents) | (Carroll's Theory)
 |
The symposium was made possible by the support of the U.S. Department of Education Office of Special Education Programs. Renee Bradley, Project Officer. Opinions expressed herein are those of the authors and do not necessarily represent the position of the U.S. Department of Education. |
