Patterns of behaviour under different concurrent reinforcement schedules

Abstract

Patterns of behaviour of seven pigeons across various arrangements of concurrent RI RI (random interval, random interval) and RI RR (random interval, random ratio) reinforcement were examined. Three separate analyses were carried out. Global analyses of behaviour revealed that subjects generally matched their behaviour to reinforcer ratios across both condition arrangements, as predicted. Local analyses of preference showed that across all conditions, initial responses following reinforcement did not change as a function of where that reinforcement was obtained. This disagreed with previous findings, where a shift in preference towards the most recent source of reinforcement had been observed. It was concluded that methodological factors that were not present here may have influenced subjects’ behaviour in previous experimentation. Lastly, analyses of switching behaviour revealed the ‘fix and sample’ pattern across the majority of subjects. This consisted of responding solely on one alternative, interrupted by brief visits of around one response to the other alternative. This pattern was consistent across both condition types, and agreed with past research. Duration of fixing behaviour varied between subjects, and was generally longer during RI RR conditions than RI RI conditions. Results overall suggested that choice behaviour is controlled by global, long term accumulators of reinforcement. Behaviour will follow a general rule when analyzed at a response-by-response level, but will show variation between individual subjects.