Behaviour is a significant component of well-being and should be properly considered in a pig welfare assessment. However, interpretation of variations in behaviour is usually not straightforward, especially when dealing with animals housed and managed in a variety of systems. The fourth principle of the Welfare Quality® project, labelled “Appropriate behaviour”, was assessed on 21 farms (11 extensive and 10 intensive) with a total of 25,856 Iberian pigs (Sus scrofa). The methodology included scan samplings of social and exploratory behaviour, human–animal relationship tests and qualitative behaviour assessments (QBA). Changes in occurrence of behaviour and qualitative measures were evaluated and discussed when comparing Iberian pigs either in intensive or extensive rearing conditions. Negative as well as positive social behaviours were significantly (P<0.001) more frequent among intensive Iberian pigs compared to extensive animals. In intensive Iberian pigs, both types of social behaviours were highly correlated (rs=+0.62). No significant differences were observed in the occurrence of exploratory behaviour and in the proportion of animals exhibiting a panic response to the presence of humans between intensive and extensive Iberian pigs. It is suggested that the high occurrences of social negative behaviours recorded in intensive conditions are clearly an indicator of poor welfare; but that interpretations of variations in “positive” social and explorative behaviour are not straightforward and may lead to misinterpretations. Despite its subjectivity, the qualitative behaviour assessment appears useful to discriminate farms.