Do not think about, write about or deal with human behavior without determining the effects of incentives.
More from the interview with Siegfried Engelmann. The whole interview is like a field manual on programming, using feedback and developing effective programs.
”In fact, in a field tryout of one of our programs, we had one kid tell us this: The kid made a mistake, and the teacher said, "No. Sound it out." And the kid looked at her and he said, "Tell me the word, and I’ll sound it out." And I thought, what a smart-ass kid, but then thought, No! He just said it all! That’s it! That’s what he believes! He has to know the word to sound it out!’
Since “the student is always right” in terms that his response is always lawful, sometimes it makes sense to listen to them.
“Another example is when kids read words in lists they make fewer mistakes than when they read the same words in context. Why? Because in grade one, teachers did eminently stupid stuff, like have them look at the picture, discuss the picture, and then read the words. I’m sorry, Virginia, pictures do not generate specific words! But specific words generate certain features of pictures! So the proper order is: Read the words. What are you going to see in the picture? Here’s the picture. Not the other way around!”
Whole Language never dies. It doesn’t cling to life because of utility, but because the theoreticians, who can’t teach students, love it.
“Another thing teachers do is to always discuss, discuss, and discuss. Frame, think, then read. Wrong! That’s what these poor readers are already trying to do. They’re trying to figure out if there’s some kind of crazy set of rules for them. And obviously, Shaywitz and a lot of others have demonstrated that this approach is inappropriate because it’s asking a kid to read the context just to find out ‘what could that word mean?’ So they’re forever making guessing mistakes and missing words. But if you have them read words in a list, they do just fine.
So, what implication does that have for a reading program? All kinds. It means that you need unpredictable sentence structures. Why? Because they’ll guess on the basis of sentence structure. "Tim and John said, ‘Let’s go to the lake.’ So, Tim and John"…everybody could complete that sentence. No. So, we would design that sentence so that if the kid said, "went to the lake," it would be wrong. Okay? Why? So we can provide the kid with information at a high rate to change their guessing behaviors.
Related to that is how you would reconfigure in the program, what they read in text. You know that kids reading words in lists do better than reading the same words in the text. So, what we did was we had a series of stories about this dog named Chee. It’s named Chee because they had trouble discriminating "Chee" and "she". Everything is for a reason. And it’s largely influenced by the feedback we use from a kid’s performance.
So when Chee gets upset or angry, Chee says random words: "Oh, of, come, for, to, go" — all the words that these kids have had since the first grade, and they’re forever missing when they read them in connected sentences. Talk about a hell of a time! They would try to read those words and they would stumble on them simply because they were nonsense statements in a deceptively sensible context for what they thought. But after awhile, they would get proficient at it.”
If you want to do what works, you get feedback from what you’re doing. You don’t decide, a priori, on the basis of your theory. You look at the evidence. Public Education is run on the basis of INTEND–what we want to do, translated into IS–what we are doing and bothers not at all with DOES–what we’re actually doing. INTEND and IS are surrounded by spin and the usual guardians of nonsense, hyperbole and catchwords. This is all, naturally, the result of perverted incentives in politics and anything funded by the political system is political. How can it be otherwise?
Cheerio and ttfn,
Grant Coulson
Cui Bono–Cherchez les Contingencies
