Environmental enrichment is important for the welfare of farm animals, which are often kept in barren environments (e.g. Young, 2003). In 2001 the European Commission adopted a directive (2001/93/EC) which states that: ‘‘Pigs must have permanent access to a sufficient quantity of material to enable proper investigation and manipulation activities, such as straw, hay, wood, sawdust, mushroom compost, peat or a mixture of such, which does not compromise the health of the animals.’’ The directive leaves considerable room for interpretation, as it is not clear what is proper investigation and manipulation. Furthermore, the value of enrichment material is most likely determined not only by the type of material as listed in the directive, but also by other material properties such as the amount and frequency of material provision, hygiene, destructibility and responsiveness. A semantic model, called RICHPIG, was constructed based on a systematic and formalized analysis of scientific information collected in a database (Bracke et al., 2007a,b). The relative importance of three assessment criteria (destructibility, hygiene and sound) were studied as treatments applied to groups of growing pigs in a 2 x 2 x 2 factorially designed experiment measuring a range of behavioural parameters. The objectives of the study were to examine how the treatments affect pig behaviour, how intensity-related measures relate to AMI, what this means for the relative importance of the three assessment criteria, and by implication to help further validate RICHPIG. These material properties were studied using a specially constructed object consisting of a piece of sisal rope, metal wire and three fixed chain links hanging in the pens. The object was considered to be not destructible (ND), hygienic (HY) and not making sound (NS). After a habituation period of 18 h treatments were applied in that the object was (or was not) made destructible with a partial cut in the rope (DE) and/or was soiled with excreta (not hygienic, NH) and/or was allowed to make a tinkling sound by releasing the chain links (SO). The three treatments were applied in a 2 x 2 x 2 factorial design on a commercial farm in seven replicates using seven different units containing eight pens per unit. At five moments in time, ranging from 18 h before until 1 h after treatment, a range of behaviours was recorded including the frequency-related parameter AMI (animal–material interactions) and four intensity related parameters. Repeated measures ANOVA’s showed significant effects of time and hygiene as well as interactions between time and hygiene, between time and destructibility and between destructibility and sound. Soiling (NH) significantly decreased, and destructibility (DE) significantly increased attractiveness, while sound (SO) was not significant. Only moderate correlations between AMI and the four, intensity related parameters were found, indicating that frequency-related parameters alone may not suffice to determine behavioural importance for animal welfare. This study showed that it is in principle possible to study material properties independent of material type and that it is in principle possible to measure behavioural intensities on a commercial farm. Furthermore, the finding that hygiene and destructibility were more important for pigs than tinkling sounds provided preliminary support for the RICHPIG model.