First, I think the way you're describing instructional/skill levels obfuscates the issue a bit. Instructional level, by definition, means the level at which a child can benefit from instruction, including with scaffolding. Frustrational, by definition, means the instruction won't work. Those levels, like the terms "reinforcement & punishment" for example, are defined by their outcomes, not intentions. If a child learned from the instruction, the instruction was on the child's "instructional" level.
Where we may be getting confused is that I think you actually are referring to teaching reading comprehension using material that is in a child's instructional level with comprehension, but on a child's frustrational level with reading fluency. This is a much different statement than what I think most teachers are getting from your messages about text complexity, to the point that I think they're making mistakes in terms of text selection.
More generally, I'd argue that there is copious research supporting using "instructional material" to teach various reading skills. Take, for example, all of the research supporting repeated readings. That intervention, by definition, uses material that is on a child's "instructional" level with reading fluency, and there is great support that it works. So, the idea that somehow "teaching a child using material on his/her instructional level is not research supported" just doesn't make sense to me.
In terms of this specific post about how much one can scaffold, I think it largely depends on the child and specific content, as Lexiles and reading levels don't fully define a material's "instructional level" when it comes to comprehension. I know many 3rd graders, for example, that could be scaffolded with material written on an 8th grade level, but the content isn't very complex, so scaffolding is much easier.
The broad point here, Dr. Shanahan, is that we're over-simplifying, therefore confusing, the issue by trying to argue that kids should be taught with reading material on their frustrational level, or on grade level despite actual skill level. People are actually hearing you say that we should NOT attempt to match a child with a text - that skill level or lexile is completely irrelevant - when I believe you know you're saying that "instructional level" is just a bit more nuanced than providing all elements of reading instruction only on a child's oral reading fluency instructional range.
First, you are using the terms “instructional level” and “frustration level” in idiosyncratic ways. These terms are not used in the field of reading education as you claim, nor have they ever been. These levels are used as predictions, not as post-instruction evaluations. If they were used in the manner you suggest, then there would be little or no reason for informal reading inventories and running records. One would simply start teaching everyone with grade level materials, and if a student was found to make no progress, then we would simply lower the text difficulty over time.
Of course, that is not what is done at all. Students are tested, instructional levels are determined, instructional groups are formed, and books assigned based on this information.
The claim has been that if you match students to text appropriately (the instructional level) that you will maximize the amount of student learning. This definition of instructional level does allow for scaffolding—in fact, that’s why students are discouraged from trying to read instructional level materials on their own since there would be no scaffold available.
Fountas and Pinnell, for example, are quite explicit that even with sound book matching it is going to be important to pre-teach vocabulary, discuss prior knowledge, and engage children in picture walks so that they will be able to read the texts with little difficulty. And, programs like Accelerated Reading limit what books students are allowed to read.
You are also claiming that students have different instructional levels for fluency and comprehension. Informal reading inventories and running records measure both fluency AND reading comprehension. They measure them separately. But there is no textbook or commercial IRI that suggests to teachers that they should be using different levels of texts to teach these different skills or contents. How accurately the students read the words and answer questions are combined to make an instructional text placement—not multiple text placements.
If we accept your claim that any text that leads to learning is at the “instructional level,” then pretty much any match will do. Students, no matter how they are taught, tend to make some learning gains in reading as annual Title I evaluations have shown again and again. These kids might have only gained .8 years in reading this year (the average is 1.0), but they were learning and by your lights, that means we must have placed them appropriately.
Repeated reading has been found to raise reading achievement, as measured by standardized reading comprehension tests, but as Steve Stahl and Melanie Kuhn have shown, such fluency instruction works best—that is, leads to greater learning gains—when students work with books identified as being at their frustration levels rather than at their so-called instructional levels. That’s why in their large-scale interventions they teach students with grade level texts rather than trying to match students to texts based on an invalid construct (the instructional level).
You write: “People are actually hearing you say that we should NOT attempt to match a child with a text -- that skill level or Lexile is completely irrelevant - when I believe you know you're saying that "instructional level" is just a bit more nuanced than providing all elements of reading instruction only on a child's oral reading fluency instructional range.”
In fact, I am saying that beyond beginning reading, teachers should NOT attempt to match students with text. I am also saying that students should be reading multiple texts and that these should range from easy (for the child) to quite difficult. I am saying that the more difficult a text is, the more scaffolding and support the teacher needs to provide—and that such scaffolding should not include reading the text to the student or telling the student what the text says.
I am NOT saying that skill levels or Lexiles are irrelevant, or that “instructional level” is simply a bit more nuanced than people think. It is useful to test students and to know how hard the texts are for that student; that will allow you to be ready to provide sufficient amounts of scaffolding (and to know when you can demand greater effort and when just more effort will not pay off).