The first objective was to test if pigs prefer a rooting material with food feedback to one without, and if they prefer to search for food while rooting compared to obtaining food without rooting. For measuring preferences we used cross-points between double demand curves. Demand curves describe the change in consumption with changing workload. Double demand curves are obtained by allowing animals to work concurrently for two resources. Their cross-point depends on the relative attractiveness of resources and therefore provides a measure of preference. The second objective was to test if results obtained when using a common scaling factor agrees with results obtained when using direct comparisons. The possibility of revealing lack of substitution by double demand curve slopes equal to zero was also addressed. Test resources used were sand, carrots, and sand with carrots. They varied according to whether the pigs could both root and eat (sand with carrots), only root (sand), or only eat (carrots). Peat was used as scaling factor. Ten ad libitum fed pigs were allowed to work concurrently, by pressing either of two panels, for two resources. Pigs were subjected to three scaling tests: peat versus sand, peat versus carrots, and peat versus sand with carrots, and three direct tests: carrots versus sand, sand with carrots versus carrots, and sand with carrots versus sand. The workload (FR, i.e. presses required per reward) alternated between FR8, FR16, FR24, FR32, and FR40 for one material, while it was kept constant for the other at FR24 (‘midpoint’). Double demand curves were constructed for all test conditions. For both the scaling tests and the direct tests, cross-points between the double demand curves showed that the pigs preferred both carrots and sand with carrots to sand, and that they did not prefer sand with carrots to carrots or vice versa. Preference quotients were calculated to measure the preference strength. Overall, results showed that a rooting material with food feedback was preferred to one without, and that non-hidden and hidden food was valued equally. Food feedback appears, thus, to be an important quality of a rooting material, even when pigs are fed ad libitum. Furthermore, results from the scaling tests overall agreed with results from the direct tests. The scaling approach therefore seems valid, at least for these particular resources. Yet, the possibility of revealing lack of substitution by slopes equal to zero is not clear.