Geary Rummler, who died last week, has influenced my thinking since a 1977 workshop on programmed instruction that used his research (and that of others). This afternoon I came across a yellowing handout from that workshop: Selected NSPI Talks, 1965.
NSPI, the National Society for Programmed Instruction, was the predecessor to ISPI, the International Society for Performance Improvement. “Programmed instruction” today sounds as archaic as coal-oil lanterns, but that workshop and its concepts have paid off throughout my career.
In “The Economics of Lean Programming,” Rummler demonstrates a technique for creating programmed instruction, a kind of print ancestor of computer-based learning.
He begins, “This is a sequence from a program.” Then he reproduces four frames (chunks of information). Whoever wrote the frames never read Cathy Moore’s advice about chopping. So Rummler “leans them up” — he chops. He offers a lean approach to the same learning sequence (ringing a sale on a cash register when given merchandise with a price tag). He has two frames, not four, and maybe a third of the words.
As the learner, you might think a bit harder, but you think sooner as well. And you don’t have to navigate through swamp gas like “A price tag has three (3) numbers on it… Look at this price tag. How many sets of numbers are on it?”
One pages three and four, he does the same with a sequence for explaining logic circuits. In the original version, there’s this valuable instruction:
One of the most common types of logic is the AND circuit. A typical _ _ _ circuit is identified by the following symbol…
As Bill Deterline said about a similar frame, this is the instructional equivalent of “copy the word ‘AND.'” By the way, that’s one of five frames in the sequence. Rummler’s revision has one frame:
An AND circuit will produce an output only if all inputs are present. Which of the following AND circuits will produce an output?
(followed by three circuit diagrams)
Only on the sixth page of the article does Rummler begin explaining the lean programming process. So he let you work with it first, and then talks about it, so you can check your understanding. Programmed instruction was criticized for proceeding in tiny increments. Rummler says that lean programming has “the idea in mind theat the ‘small steps’ … should be the largest step the learner can take without failure.”
Rummler strongly advocated developmental testing — trying out very lean drafts on a learner to test assumptions about what’s needed. In the workshop, the facilitators explained it this way: if you write all the instructional stuff first, you don’t know if people pass because you taught them, or because they knew this stuff already — in which case you were wasting your time and theirs.
He also urged you to write criterion frames (test items) before writing any teaching frames. Whether you’re talking about job-related training or learning (I don’t want to get too philosophical), there’s great value in this, which I see as working backward from the result you want to achieve. It ought to be good instructional design, but all over North America today, people are sitting or clicking through “learning” that’s full of the history of widgets, great moments in widgets, appreciation of widget-making, and other flapdoodle.
When you do enough of this spoon-feeding, you drive away many people with working brains, and you train the others to be spoon-fed. Because, as Rummler once said, “If you pit a good performer against a bad system, the system will win almost every time.”