Graded differential reinforcement: Response-dependent reinforcer amount.

The team worked with pigeons in a small lab chamber.

Each peck on the key added 0.25 seconds of grain.

More pecks meant longer grain time.

The birds could control how big the snack was.

A second group got the same grain times no matter how much they pecked.

Birds worked harder when grain time grew with every peck.

They pecked more than the yoked birds who got identical grain no matter what.

Linking reinforcer size to response count pushed the behavior up.

King et al. (1990) later tested adult humans with points instead of grain.

They saw the same boost when people had to press a button to collect the points.

The idea crosses species: let the learner do a quick "collect" action and the amount matters more.

Pliskoff et al. (1972) also used pigeons and differential reinforcement.

They showed that reinforcing high rates in one setting can speed up responding in other settings.

Both papers tell us to watch for spill-over when we tie bigger rewards to faster or more responses.

Fortes et al. (2015) seems to disagree at first.

They found that making birds peck a lot during a delay made the big reward less attractive.

The key difference is timing: D et al. tied amount to the main response, while Inês et al. piled extra work into the wait.

Effort during the task helps; effort during the delay hurts.

You can use graduated reinforcement when you want more effort from a client.

Let each correct response earn a slightly bigger slice of preferred item.

Keep the collection step visible: have the learner drop a token in a jar or press a button to hear the tally.

Watch for induction: if you reinforce fast math facts, check that other tasks don’t speed up too much.

Finally, don’t load extra responses into wait time; save the work for the target behavior itself.