Conditioned place preference (CPP) or aversion (CPA) methods are potentially useful tools in animal welfare assessment because they permit measurement of the reinforcing properties of a stimulus in the absence of the stimulus itself. We used CPP/CPA techniques in a series of experiments to assess the preference of food restricted broiler breeders for increased food quantities or avoidance of aversive stimuli. In all experiments, 6-10 week old Ross308 pairs of broiler breeders were housed in pens divided in half. Pen sides were visually differentiated and birds were trained with different stimuli on each pen side, correcting for possible side biases, either different food amounts for CPP or aversive stimuli on one pen side and none, or 'neutral', on the other, for CPA. To test if a preference for a pen side had been formed, the pen divider was removed when no stimulus was present and the amount of time birds spent on each pen side was recorded. Each experiment had a factorial treatment structure (n=10 replicate pens per treatment combination) and the proportions of time spent by birds on the 'positive' pen side (i.e. increased food amount for CPP or 'neutral' for CPA) were analysed. In experiment 1 on CPP (180 birds forming 90 pairs), three different training regimes in combination with three different testing methods were trialled: whilst during training of all birds, on one pen side the birds received the commercially recommended, restricted amount of food (R) and on the other pen side they received twice that amount (2R) (no food was present during testing). In experiment 2 (110 birds forming 40 pairs and 10 individually housed) and 3 (80 birds forming 40 pairs), further refinements were made to the experimental methods and birds were allocated to CPP treatments with food amounts 2R or 3R (vs. R) or to CPA treatments, 'social isolation' or 'unpredictable wind'. Overall, there was no evidence of aversion at testing to the pen sides with aversive stimuli during training and little evidence of preference at testing for the pen sides with increased food amounts during training. Furthermore, where statistical significance was achieved for CPP the preferences shown were very small. The most consistent result was a strong preference for the pen side birds were not previously housed on immediately before each test (P<0.001 in all experiments). It appeared that birds were motivated to explore a location where they had not just been housed in an attempt to find food and this motivation seemed to overshadow other effects. This series of experiments demonstrates some limitations of CPP/CPA techniques for welfare assessment and the learning problems experienced by chronically food restricted animals