Within-session changes in responding during several simple schedules.

Silverman et al. (1994) watched how response speed changes inside one session. They used rats and pigeons under simple FI, VR, and DRL schedules.

Some animals earned lots of reinforcers each minute. Others earned only a few. The team tracked how response patterns moved from start to finish.

When reinforcement came fast, the animals' response rates drifted in a clear direction within the same session. The pattern looked like a slope, not a flat line.

Under lean schedules the line stayed flat. Time alone did not drive the change; the local density of reinforcement did.

Lowe et al. (1995) later shifted VI rates in the middle of each session. Response rates followed those shifts, backing the idea that moment-to-moment reinforcement density is the lever.

Rutland et al. (1996) saw the same drift in autoshaping, a Pavlovian procedure. The effect is not limited to operant schedules; it shows up whenever food rate changes.

Buitelaar et al. (1999) added behavioral-economics measures. Demand curves also slid within sessions, so the drift can hide economic analyses unless you keep sessions short.

If you run functional analyses or preference assessments that last 10 minutes or more, the client's response rate may slide on its own. Check the last 2 minutes against the first 2. If you see a slope, shorten the session or split the data to avoid misreading a reinforcement effect that is really just drift.