BEHAVIOR STABILITY AND RESPONSE RATE AS FUNCTIONS OF REINFORCEMENT PROBABILITY ON "RANDOM RATIO" SCHEDULES.

Six pigeons pecked a key for grain under a random-ratio (RR) schedule. The chance of food after each peck varied from a large share to a large share. Birds lived in the chamber for 40-50 days to reach steady performance.

The team wanted a clean curve: higher ratio probability should mean faster pecking. They recorded daily response rates to see if the curve appeared.

No smooth curve showed up. Some birds sped up a little, others slowed, and most rates drifted sideways across the whole range of probabilities.

Even after seven weeks the data stayed messy. The authors concluded that RR probability alone does not set a dependable pace of responding in pigeons.

Iwata et al. (1990) later saw tidy rate differences when toddlers worked under RR versus random-interval (RI) schedules. Kids always responded faster in the RR component. The pigeon data look messy because the 1964 study tested only one schedule type; the toddler study compared two.

Winterling et al. (1992) used a changing-criterion VR schedule to raise exercise rates in obese boys. Each new target was a large share above the prior mean and pedaling climbed steadily. Their orderly gains show that when you shape ratio requirements in steps, rate control is achievable even with children.

Barnes et al. (1990) let pigeons choose between concurrent VR and VI keys. Birds switched toward the VR key when sessions were short or food portions small. This orderly preference again shows ratio schedules can control rate and choice when contrasted with interval schedules.

If you expect a straight line between ratio size and client speed you will be disappointed. Pair ratio schedules with interval schedules or shape ratio requirements in small steps to see cleaner rate changes. When designing interventions, compare schedule types rather than testing one ratio value alone.