An assessment of token value and effectiveness: A replication

Hangen et al. (2023) asked a simple question: do paired tokens really work as reinforcers? They ran a progressive-ratio schedule with three kids. The kids had to press a switch more and more times to earn either candy, paired tokens, or unpaired tokens.

Each child worked in all three conditions. The team counted how many responses the child made before quitting. More responses meant the item had higher value.

Candy won every time. Kids pressed far longer for candy than for any token type. Paired tokens sometimes kept the kids going, but the effect jumped around. Unpaired tokens barely worked at all.

The takeaway: tokens can help, but they are weaker than primary reinforcers. Check value first before you bank on them.

Cihon et al. (2019) also used tokens, yet they saw great results. The difference? They kept the exchange number secret and changed it on the fly. That mystery kept kids talking during snack. Hangen’s fixed, known schedule may explain why their tokens looked weaker.

Richling et al. (2019) warn that standard rules like 80 % over three sessions can fail. Hangen’s lab-style value check is one way to probe before you lock in a program.

ASutton et al. (2022) reviewed dozens of studies and found most packages worked, but few teams checked each part. Hangen’s piece-by-piece test fits that call for tighter assessment.

Before you set up a token board, run a quick progressive-ratio probe with your learner. Let them work for candy, then for tokens. If tokens fall flat, pair harder or add more backup reinforcers. A five-minute value check can save weeks of weak programming.