Many advertisements appear to rely on higher-order respondent conditioning to alter preference for a product brand. Given that higher-order conditioning is difficult to achieve, there may be other processes, such as those involved in stimulus equivalence, at work. The current experiment paired six “brand” stimuli with three different types of pictures. Three brands were paired directly with the pictures and the other three paired indirectly, using a stimulus equivalence procedure. The three types of pictures were “negative” (typical consequences of ingesting excessive sugar, such as heart disease, tooth decay, and diabetes), “neutral” (pictures of geometrical shapes) and “positive” pictures (pictures related to health and fitness). This research used a respondent procedure, which is more similar to real marketing methods than operant procedures, as used in previous research in this area. A preference assessment was conducted to evaluate transfer of function from the picture to the brands by investigating whether the pairings influenced participants’ preference for branded soft drinks. The results showed that negative pictures shifted preference more than positive pictures and direct pairing was more effective than an indirect pairing procedure.