by Richard L. Allington
The latest educational fad seems to be the Three Tier intervention model. I say fad because I have been unable to find a single research study that supports this intervention design. No one, it seems, has actually tested the Three Tier model against competing intervention models.
With all the ruckus about using scientific research to inform our efforts to close the achievement gap, one would think someone would have designed at least one experiment documenting the effectiveness of the Three Tier model before state and federal education agencies began recommending—or mandating—the model’s use. Without such evidence to support its effectiveness, it seems premature to recommend this intervention model.
I had optimistically hoped that the Congressional emphasis on the importance of research evidence would slow down the faddish adoption of unproven program models. But it hasn’t.
In addition, I have real concerns about how the Three Tier model is being interpreted (or misinterpreted) by state and local education agency officials.
My first concern is with the unwarranted emphasis often placed on purchasing different commercial instructional packages for each tier. I say unwarranted because using three different instructional packages is inconsistent with the earlier intervention research on this issue.
In a chapter in the book Title I: Compensatory Education at the Crossroads (Erlbaum, 2005), Geoff Borman and his colleagues analyzed the impact of coordinating Title I–funded remedial reading instruction with the classroom reading instruction. They found that increases in coordination were “associated with an increase of 4.7 to 7.1 normal curve equivalents in classroom mean reading achievement...and with a reduction of the Title I student achievement gap by 0.2 to 0.3 standard deviation units.”
In short, they found that when remedial reading teachers and classroom teachers used a reading curriculum (I said curriculum, not program) that was the same or similar, they increased the achievement levels of all students and reduced the achievement gap between struggling readers and their normally progressing peers. Unfortunately, fewer than one third of all Title I students experienced such instructional designs.
This finding demonstrates what Peter Johnston, Peter Afflerbach, and I had hypothesized about curricular coherence in an article titled “The Congruence of Classroom and Remedial Reading Instruction” (Elementary School Journal, 1985). We suggested that placing at-risk students in two or more different reading curricula would have a less positive effect on student achievement than simply personalizing, extending, and intensifying the reading instruction offered in the classroom reading lessons.
However, we cautioned that any intervention had to build primarily around students’ instructional needs. For instance, if students struggled because they did not use effective comprehension strategies, then expanding comprehension strategy instruction in the intervention lessons would seem a good idea. If students struggled because they did not self-monitor, then that would become one focus of the intervention lessons. Designing effective intervention requires that a student’s specific instructional needs be identified.
Borman and his colleagues found that our hypothesized model did, in fact, enhance the reading achievement of struggling readers, and it resulted in improved reading achievement for students whether they participated in the Title I (Tier 2) intervention or not. That result seemed to stem from the necessary collaboration between the Title I reading teachers and classroom teachers. In other words, achieving instructional coordination required that the classroom and Title I teachers meet on a regular basis to plan together. This joint planning enhanced achievement, though the study did not allow the researchers to identify just how that occurred.
Nonetheless, after examining the results of their two-level hierarchical analysis of the available data, Borman and his colleagues concluded that struggling readers “benefit most from having extra time to master the developmental material of the regular classroom” reading curriculum.
What concerns me, then, about the way the Three Tier model is being implemented is that the most commonly recommended designs replicate a long-standing flaw of traditional Title I and special education programs—the fragmentation of the reading instruction struggling readers receive. In the full-blown version adopted by my state department of education in Tennessee, those struggling readers receiving all three tiers of instruction would encounter three different commercial reading programs every day!
That seems like a good plan if you wanted to confuse a struggling reader. If, however, your goal was to accelerate the reading development of struggling readers, then the Borman analysis suggests that such planned fragmentation is, in fact, a bad plan. Indeed, that study suggests such a plan will produce a far more modest impact on achievement than if the Three Tier model were redesigned to provide additional, expert, coordinated reading lessons every day.
I will suggest that a Three Tier model based on the research available might look more like this:
Tier 1: High-quality comprehensive classroom reading instruction. I believe this is the most critical tier. Classroom reading instruction should be of high quality and informed by the research available. This would mean, for instance, that classroom reading instruction would provide lessons that were personalized to students’ needs as opposed to simply moving all students through a single, standard core reading program. There is sufficient research (as summarized in the IRA position statement Making a Difference Means Making It Different) demonstrating the critical nature of matching learners to texts of appropriate levels of difficulty. In an article that will appear in a forthcoming issue of Elementary School Journal, Anne McGill-Franzen and her colleagues show that mandating a one-size-fits-all classroom core curriculum left many children behind in a statewide analysis of reading achievement in high-poverty schools.
Tier 2: Small-group (five students or fewer) supplemental instructional support. Tier 2 expands the reading instruction provided for those students failing to make adequate progress with the Tier 1 classroom lessons. This instruction must be personalized as well and must complement, improve upon, and/or extend the classroom reading lessons. This additional instruction might be provided by the classroom teacher during or after the school day. It might be provided by a reading specialist in the classroom or in a pull-out program. In a typical scenario, an additional 30 minutes of reading instruction is provided daily. There is little evidence indicating that providing such instruction in groups larger than five produces accelerated reading growth.
Tier 3: Intensive very small-group (two or three students) or one-on-one tutorial instruction. Tier 3 provides even more daily reading instruction for students who are failing to make adequate progress with Tier 1 and Tier 2 efforts in place. Following the supplemental educational services model from the Reading First program of the No Child Left Behind Act, this third tier of instruction is most profitably provided in an extended school day model. Again, the intervention should support and accelerate progress through the classroom reading program. Very small intervention groups and/or expert tutoring have consistently produced accelerated reading growth.
This Three Tier model is better supported by the research than any model that has struggling readers experiencing three different commercial reading programs, especially if three different teachers are teaching the three tiers. As a researcher, I would prefer to see state education agencies funding multiple intervention designs and rigorously gathering evidence on the outcomes of each variation. We might see competing Three Tier models (fragmented vs. coherent instructional plan) implemented and closely monitored for effects. Or perhaps we could test having certified reading specialists provide the second and third tiers of intervention against having nonqualified personnel or teachers with other certifications provide it.
Proponents of the Three Tier intervention design have seemingly convinced federal and state education officials that their model is based in research, even though no experimental studies have tested that model against other intervention models. A commercialized version of the Three Tiers framework is now being mandated in state after state. This is unfortunate because those struggling readers really do need help. Unfortunate because many well-intentioned professionals will work hard implementing the commercialized Three Tier model only to have the flawed design undermine their good efforts. Unfortunate because schools will spend far too much money purchasing all those commercial intervention packages that are unneeded and may be more the source of the problem of continuing underachievement than a solution.
Struggling readers need larger amounts of more expert, more personalized, and more intensive reading instruction. In the end, the quality of that instruction is critical, and high-quality instruction for struggling readers cannot simply be boxed up and shipped to a site. High-quality reading instruction, especially for struggling readers, requires the expertise to identify just where the reader has gotten off-track and then to design instruction that moves the reader back onto an accelerated track of development. This could be provided in a Three Tier model, but not in a Three Tier model that simply provides larger amounts of fragmented, inexpert, one-size-fits-all instruction that leaves far too many children behind.
IRA President Richard L. Allington is a professor of education at the University of Tennessee.
Research and the Three Tier model. (April 2006). Reading Today, 23(5), 20.
Journals
Reading Today