The Effects of Conceptual Priming on Stimulus-Based Choice

Antonia Mantonakis

BrockUniversity, Faculty of Business

500 Glenridge Avenue

St. Catharines, Ontario, CanadaL2S 3A1

Email:

Phone: (905) 688-5550 ext. 5383; Fax: (905) 378-5716

Carolyn Yoon

University of Michigan

StephenM.RossSchool of Business, University of Michigan

701 Tappan St

Ann Arbor, MI, USA, 48109

Email: .

Phone: (734) 764-6355; Fax: (734) 936-8716

Using a DRM paradigm (Deese 1959; Roediger and McDermott 1995), we explore the role of conceptual priming, by exposure to meaningful brand-related cues, in stimulus-based choice. We demonstrate an effect on choice and argue that it is associated with increased accessibility, revealed by illusory recognition of competing, non-exposed brands. A beneficial choice effect occurs for strong brands of the primed category, with an accompanying detriment to weak brands. This result is in contrast to the idea that conceptual priming only impacts memory-based choice (Lee 2002), calling into question the need to differentiate effects of conceptual versus perceptual priming on choice.

Conceptual priming occurs when a meaningful stimulus (the “prime”)prepares a person to perceive a target that is semantically related to the prime. Flashing the prime “Mr. Sub” on a computer screen may subsequently make a person read the target “Quizno’s” (semantically-related to “Mr. Sub”) faster than if the prime “Pantene” were flashed (there is no semantic relation between “Pantene” and “Mr. Sub”). Conceptual priming of brands increases memory-based choice probability (Nedungadi 1990).Flashing “Mr. Sub” makes a person more likely to include “Quizno’s” in the consideration set and choose it in a memory-based choice task. Perceptual priming occurs when a prime that is a close perceptual match to the target (e.g., in form or color) prepares a person to perceive the target. Flashing the prime “Burger King” on a computer screen may make a person read a perceptually-identical target “Burger King” faster than if the prime “Pantene” were flashed. Perceptual priming of brands increases stimulus-based choice probability (Lee 2002).

Lee (2002) observed that a brand that was conceptually primed (e.g., by prior exposure to meaningful ad copy) benefitted in a memory-based choice setting, because the brand would be more likely to be brought to mind and included in a consideration set. A brand that was perceptually primed (e.g., by prior exposure to the brand) benefitted in a stimulus-based choice setting. Taken together, Lee showed that perceptual (conceptual) priming of brands increases stimulus-based(memory-based) choice probability, and argued that conceptual and perceptual knowledge stores represent distinct constructs within implicit memory.

The effects of priming on choice depend on the match in cognitive operations between the requirements of the choice task and the type of priming that occurs (Lee 2002). Stimulus-based choice settings (e.g., choice of cereal brands in a grocery store aisle) demand identification and recognition. Thus, prior visual exposure to a brand that involves passive perception (e.g., an image of the product) would benefit stimulus-based choice of the brand. In contrast, memory-based choices are sensitive only to interventions involving conceptual processing; there is an overlap in cognitive operations between processing the prime, which involves elaborative processing about the meaning, and generating the target for inclusion in the consideration set (i.e., retrieval based on associations). The differential priming effects for memory- and stimulus-based choices demonstrate encoding specificity, which is the idea that the effects of priming on later tasks depends on the matchin cognitive operations between the requirements of the task and the type of priming that took place (Tulving and Thompson 1973).

We concur with Lee’s (2002) assertion about the relevance of the encoding specificity principle to the effects of priming on choice. However,we note that consumer choices in memory-based settings are also more effortful than in stimulus-based settings. In Lee’s studies elaboration and effort were confounded. The level of effort required for the target (high for memory-based, low for stimulus-based) matched the level of effort required to encode the prime (high for memory-based, low for stimulus-based).Effort is high in memory-based choice because a person has to generate alternatives to be included in the consideration set. Thus,Lee’s finding that a prior exposure to a prime involving relatively passive perception benefits only stimulus-based choice of the target can also be accounted for by the relatively low level of effort required by stimulus-based choice in her experiments.

We proposed that conceptual priming can have an effect on stimulus-based choice if the levels of effort required to process the prime and the effort required to process the target in the choice task are similarly low. We suggested that conceptual priming effects on stimulus-based choice will be observed when the level of effort required during exposure to the prime matches the level of effort required to process the target during choice. We predicted that if conceptual priming requires relatively effortless processing (e.g., a billboard slogan), and the choice settingrequires minimal cognitive effort (as is the case for stimulus-based choices), then effects of conceptual priming should be observed in stimulus-based choices which are less effortful than memory-based choices.

Results of experiment 1 supported our hypotheses about the effects of conceptual priming on stimulus-based choice; however, it was not entirely clear how the increase in stimulus-based choice likelihood for the brands occurred. Results of experiment 2 supported the hypothesisthat conceptual priming increased accessibility of the category, which then increased accessibility of the more prototypical brands of the category because they are also strongly associated with the category. Weak brands within the category also received increases in accessibility, although the increases were not as large as those for the strong brands. Thus weak brands were at a competitive disadvantage when in the same choice set as strong brands whose accessibility was enhanced. In summary, the relationship between accessibility strength and probability of stimulus-based choice is in the same direction for strong brands (increased accessibility from conceptual priming is associated with an increase in stimulus-based choice probability) and in opposite directions for weak brands (increased accessibility from conceptual priming is associated with a decrease in stimulus-based choice probability).

We introduce the idea that specificity not only involves elaboration (Lee 2002), but also effort. We reasoned that the level of effort involved can predict an effect of conceptual priming on stimulus-based choice; however, it is not predicted by the assumptions of the “separate systems” account of implicit memory, which suggests that conceptual and perceptual stores of knowledge exist which account for separate influences of conceptual (perceptual) priming on memory-based (stimulus-based) choice (Lee 2002). Thus, we have outlined predictions about choices that can be outcomes of both conceptual and perceptual priming, through a single set of underlying principles: matching effort required during priming and choice decision (Mantonakis, Whittlesea and Yoon 2008). Our finding of conceptual priming effects on stimulus-based choice is at least challenging for proponents of separate-systems accounts of memory as it requires them to provide functional explanations at a deeper level than the notion that there are two kinds of knowledge, and therefore two memory systems specialized to deal with them.

Theimplicit/explicitmemory distinction emphasizes the differential role of consciousness in performance, distinguishing between tasks in which awareness of prior exposure is important (tasks that cue the “explicit” system), versus tasks that depend on associative priming (tasks that engage the “implicit” system; Graf and Schacter 1985). Priming occurs when a stimulus (the “prime”) prepares a person to perceive or retrieve a target. Priming is indexed by an increase in probability of reporting (or identifying) a target,relative to a baseline ofreporting (or identifying) “non-primed” targets (Jacoby and Dallas 1981). A common method used in priming studies involves exposing participants to a prime, and in a subsequent test having participants read the target; the experimenter records the reaction time to reading a target. Another type of test entails having participants report an item, such as listing a brand to be included in a consideration set. Under many conditions, priming tasks may be better measures of advertising effectiveness than explicit measures such as recall (Lee 2002). Although priming itself is based on specific prior exposures, performing the tasks that demonstrate priming has taken place does not require the person to report a conscious reference to prior episodes in memory.

Within the implicit system, there are two kinds of knowledge, conceptual and perceptual, and there are two sub-systems specialized to deal with them (Lee 2002). That is, conceptual knowledge and perceptual knowledge each have their own specialized processing systems.

Conceptual priming occurs when a meaningful prime prepares a person to perceive a target that is semantically related to the prime. For example, flashing the prime “Mr. Sub” on a computer screen may subsequently make a person read the target “Quizno’s” (semantically-related to “Mr. Sub”) faster than if the prime “Pantene” were flashed (there is no semantic relation between “Pantene” and “Mr. Sub”). Conceptual priming of brands increases memory-based choice probability (Nedungadi 1990). That is, flashing “Mr. Sub” makes a person more likely to include “Quizno’s” in the consideration set and to choose it in a memory-based choice task.Perceptual priming occurs when a prime that is a close perceptual match to the target (e.g., in form, color, label-font, or other physical characteristics) prepares a person to perceive the target. For example, flashing the prime “Burger King” on a computer screen may make a person read a perceptually-identical target “Burger King” faster than if the prime “Pantene” were flashed. Perceptual priming of brands increases stimulus-based choice probability (Lee 2002): Flashing “Burger King” makes a person more likely to choose “Burger King” in a stimulus-based choice setting.

Lee (2002) proposed that whereas perceptual priming of brands increases stimulus-basedbut not memory-based choice probability, conceptual priming of brands increases memory-basedbut not stimulus-based choice probability. Lee showed that a brand that was conceptually primed (for example by prior exposure to meaningful ad copy) would benefit in a memory-based choice setting, because the brand would be more likely to be brought to mind and included in a consideration set. On the other hand, a brand that was perceptually primed (for example by prior exposure to the product’s brand name) would benefit in a stimulus-based choice setting. This imposes a constraint on the types of actions (e.g., ad copy) that may be used effectively by marketers of offerings in stimulus-based choice situations.

The effects of priming on choice depend on the match in cognitive operations between the requirements of the choice task and the type of priming that occurs (Lee 2002). To illustrate, stimulus-based choice settings (e.g., choice of cereal brands in a grocery store aisle) demand identification and recognition. Thus, prior visual exposure to a brand that involves passive perception (e.g., an image of the product) would benefit stimulus-based choice of the brand. In contrast, memory-based choices are sensitive only to interventions involving conceptual processing; there is an overlap in cognitive operations between processing the prime, which involves elaborative processing about the meaning, and generating the target for inclusion in the consideration set (i.e., retrieval based on associations). The differential priming effects for memory- and stimulus-based choices(Lee 2002) demonstrate encoding specificity, which is the idea that the effects of priming on later tasks depends on the matchin cognitive operations between the requirements of the task and the type of priming that took place (Tulving and Thompson 1973).

FIGURE 1

Effects of Conceptual versus Perceptual Priming on Memory- versus Stimulus-Based Choice

Note: This diagram summarizes our explanation of Lee’s (2002) observation of a double dissociation (whereby conceptual priming affected only memory-based choice, whereas perceptual priming affected only stimulus-based choice). We argue that the conceptual priming manipulation (and memory-based choice task) required high elaboration and high effort, whereas the perceptual priming manipulation (and stimulus-based choice task) required low elaboration and low effort. We argue that elaboration and effort were confounded.

The level of processing effort required for the target (high for memory-based, low for stimulus-based) matched the level of effort required to encode the prime (high for memory-based, low for stimulus-based).Effort is high in memory-based choice because a person has to generate alternatives to be included in the consideration set. Prior exposure to a prime that requires extensive elaboration (e.g., elaborative ad copy) also requires more effort, which would benefit memory-based choice for the target. In contrast, stimulus-based choice involves recognition, which is less effortful than the processing involved in memory-based choice.A prior exposure to a prime that does not require elaboration (e.g., straightforward ad copy) also requires less effort to process.

Thus,Lee’s finding that a prior exposure to a prime involving relatively passive perception benefits only stimulus-based choice of the target can also be accounted for by the relatively low level of effort required by stimulus-based choice in her experiments (cf. Kolers 1973; Hughes and Whittlesea 2003). We argue that, under some conditions, for example when consumers are bombarded with advertisements, processing of conceptual primes may also become virtually automatic and involve little effort such that they facilitate stimulus-based choice of the target.

We propose that conceptual priming can have an effect on stimulus-based choice if the levels of effort required to process the prime and the effort required to process the target in the choice task are similarly low (Figure 2). We suggest the idea that facilitation in cognitive processing of a target from exposure to a prime depends on the match in effort between what is required in the choice task and during exposure to the prime. More specifically, we suggest that conceptual priming effects on stimulus-based choice will be observed when the level of effort required during exposure to the prime matches the level of effort required to process the target during choice. Based on this reasoning, we predict that if conceptual priming requires relatively effortless processing (e.g., a billboard slogan), and the choice settingrequires minimal cognitive effort (as is the case for stimulus-based choices), then effects of conceptual priming should be observed in stimulus-based choices which are less effortful than memory-based choices. This finding would imply that marketers should explicitly consider the amount of effort involved in the typical choice situations for the brands that they manage. The level of effort required by the choice should influence the types of marketing interventions employed to increase sales.

FIGURE 2

Effects of Conceptual Priming on Stimulus-Based Choice

Note: We use a low effort conceptual priming manipulation to demonstrate effects of conceptual priming on stimulus-based choice.

If conceptual priming leads to changes in stimulus-based choice as we propose, there are significant theoretical implications. Lee’s (2002) finding of a “double dissociation” (i.e., memory-based choices are sensitive to interventions that involve conceptual priming but not perceptual priming, and stimulus-based choices are sensitive to interventions that involve perceptual priming but not conceptual priming) was the basis for the conclusion that conceptual and perceptual systems are separate systems in implicit memory. If conceptual priming leads to changes in stimulus-based choice, it would undermine the argument for separate systems. It would require proponents of separate systems to argue at a deeper level as to why it would be necessary to account for a distinct knowledge processing system that is differentially sensitive to conceptual priming. That is, proponents would have to answer why it would be the case that a separate conceptual system is necessary to specialize in memory-based choices.

What is needed to test our conjecture that variations in effortfulness of processing conceptual primes will have differential effects on stimulus-based choice is prime stimuli that are low (versus high) in effort-to-process. To develop our stimulus set, we conducted two pilot studies: One group of participants generated a list of brands for several product categories; a second group rank-ordered each brand based on prototypicality within its category. Given that the most prototypical items in such lists are known also to be more fluently processedthan less prototypical items(Whittlesea 2002), our hypotheses incorporate the role of processing fluency (cf. Lee 2002). Consistent with attributional accounts of memory, we define processing fluency not as sheer speed of processing, but rather as the degree to which a stimulus, or component features of a stimulus, are consistent with current expectations of the category in which they belong (Whittlesea and Williams 1998). For example, McDonald’s, a prototypical fast food brand, would be more consistent with expectations for the product category evoked by the conceptual prime “Burger King”, versus a prime such as “Joe’s Deli”. Accordingly, McDonald’s should benefit more than a less prototypical brand (because of an advantage in fluency) in a stimulus-based choice setting.