Features of Direct Instruction: Content Analysis

Slocum et al. (2021) looked under the hood of every major Direct Instruction program. They asked, 'What makes these lessons so powerful?'

The team mapped the tiny links between skills. They called this job 'content analysis.' No kids were tested; the paper is a recipe, not a taste test.

Strong DI programs hide a web of 'generative relations.' A child who learns A can now do B, C, and D without extra teaching.

Weak programs miss these links. Kids have to be taught every little step, so learning is slow and fragile.

Frampton et al. (2018) gives a live demo. Three autistic children first learned a problem-solving strategy. That one skill let them suddenly explain how to brush teeth, make a sandwich, and more. Their data is an example of the web Slocum describes.

Gallant et al. (2021) shows the same idea with college kids. Two equivalence-based formats built whole networks of logical-fallacy classes. Both beat self-study, proving that engineered links work at any age.

Dell’Aringa et al. (2021) seems to disagree. They found that adding ‘transfer trials’ to DTT did not speed up tact learning. But their study tested only one small link, not a full web. Slocum’s point is: map the entire maze first, then choose the fastest path.

Before you write a single program, list every skill your learner needs. Draw arrows: which skills open the door to others? Put the ‘keystone’ skills early and practice them hard. You will cut teaching time in half and watch novel responses pop out for free.