Analysis of precurrent skills in solving mathematics story problems.

Two third-grade students with developmental delay could not solve arithmetic story problems.

The team broke each problem into four tiny skills: find the starting number, find the change, pick the operation, and name the ending number.

They taught the four skills one at a time with short discrete-trial drills until each child hit 100 % correct for three sessions in a row.

Both kids mastered every piece.

After the last piece, they could solve brand-new story problems the teacher had never shown them.

Accuracy jumped from 0–20 % before training to 80–100 % after training.

Winett et al. (1972) used the same step-by-step idea in writing class. They reinforced “total words,” then “different words,” then “new words.” Writing output soared, showing the sequence trick works across subjects.

Kully-Martens et al. (2018) later packaged component math lessons into the MILE program for kids with FASD. Their larger quasi-experiment echoed the 2003 finding: teach the pieces, the whole problem follows.

Knutson et al. (2019) looks opposite at first glance. They found kids learn fastest when you give ONLY new tasks, no mixed review. The 2003 study, however, chained four NEW skills in order, so both papers agree—keep the focus on acquisition targets, not old mastered ones.

If a learner is stuck on word problems, stop guessing and slice the task. Run quick trials on “find the start number,” then “find the change,” and so on. Once the chain is clean, untaught problems solve themselves. You can finish the whole sequence in a week of 15-minute drills.