What is the effect of advertising on memories of consumers? - Chapter 3

What does human memory entail?

Human memory is a system that not only allows us to record, store and retrieve the information we acquire, but influences the way this information is perceived, stored and encoded as well. Encoding refers to the processes involved in getting the information into the system by transforming an external stimulus into an internal representation, which allows us to retain it in the cognitive system. Storage involves information retention over time, which can be short term or relatively permanent. You can also be able to retrieve this information from your memory, or forget it. Retrieval refers to the processes that enable us to find the information which is stored in our memory, while forgetting refers to the processes that keep us from finding it.

The model of Atkinson and Shiffrin

The Atkinson and Shiffrin model consists of three parts: sensory memory, working/short-term memory, and long-term memory. Information is first held briefly in the sensory memory, then a selection is transferred to the short-term memory, and afterwards an even smaller selection is stored in the long-term memory.

Sensory memory stores should be considered part of the perception process, because the stored information is not yet encoded, but stored in the sensory modality in which it has been perceived.

In the working memory, input from the different sensory memories is integrated with information from long-term memory to be briefly held in conscious awareness and manipulated. New information can only access when old information is moved out. Verbal rehearsal lengthens the period of time for which information stays in the short-term store and at the same time builds up the trace in the long-term memory. When rehearsing information and paying attention to the information, the likelihood of storage in the long-term memory increases.

Long-term memory stores nearly unlimited amounts of information for an unlimited period of time. The model of Atkinson and Shiffrin proposes that information that has entered the processing system through modality-specific sensory stores and which has gone through the short-term of working memory, needs to be encoded semantically before entering the long-term memory.

According to Baddeley there is major evidence for the multi-systems view of memory, namely:

• Speed of retrieval: retrieval from short-term memory is faster than from long-term memory.

• Capacity: short-term memory is more limited; it can only hold 5-7 pieces of unrelated information.

• Serial position effects: items presented at the beginning (primacy) and at the end (recency) of a list are recalled earlier and more often than items in the middle. Primacy items can be rehearsed and stored in the long-term memory, while the last items still reside within the short-term memory.

• Memory code: long-term memory relies mainly on semantic codes, while short-term memory uses acoustic/phonological coding.

• Neuropsychology: patients who suffer from amnesia may have perfect short- or long-term memories while the long- or short-term memory is impaired.

Although evidence supports the conception of short-term and long-term memory as separate systems, these findings are inconsistent with the assumption that information can only reach the long-term memory by passing through short-term memory. In addition, people sometimes do not remember objects clearly even though they are likely to have seen those thousands of times. This seems to be inconsistent with the assumption that information that is frequently rehearsed is likely to be stored in long-term memory. Two approaches to resolve these inconsistencies are (1) to abandon the idea of different memory systems, and (2) to reformulate the distinction between short-term and long-term memory.

Levels of processing

The levels of processing approach assumes that items are remembered better the more we pay attention to them and the more deeply they are processed.

The models of working memory of Baddeley and Hitch

Baddeley and Hitch replaced the model of working memory as a unitary system by a multi-component working memory model. A controlling attentional system, the central executive, which has no storage capacity of its own, supervises and coordinates a number of subsystems. Two subsystems are the phonological loop and the visuo-spatial sketchpad.

The phonological loop is responsible for the short-term storage and manipulation of speck-based information. It consists of a phonological store that briefly holds sounds and an articulatory rehearsal system that uses sub-vocal inner speech. Unless speech-based/sound information is sub-vocally rehearsed, it will become irretrievable within 2 seconds. Articulatory rehearsal process also translates written material into phonological code to allow storage in the loop.

The visuo-spatial sketchpad is responsible for the short-term storage and manipulation of visual information. It can operate at the same time as the phonological loop, e.g. explaining someone the way to a nearby street and at the same time visualizing the route to that street.

The central executive allocates attention and coordinates the two subsystems. Its main task is planning sequences of activities. It cannot store information, so there is an episodic buffer, which is a place where information from long-term memory and the subsystems of working memory can be temporarily stored, integrated and manipulated. This is where the consumer consideration set – the number of brands from a given product category that a consumer is actively choosing from at a given occasion – can be found.

Forms of long-term memory

The most important distinction of components of the long-term memory is between forms of conscious and non-conscious memory. Conscious forms are declarative, explicit, or recollective. Non-conscious forms has been referred to as implicit memory.

Declarative or explicit memory is characterized by a person’s conscious recollection of facts or events. Subcategories are episodic and semantic memory.

Episodic memory refers to a memory about a specific event that occurred at a particular place and time. For example, do you remember whether you have ever been to Paris?

Semantic memory refers to the mental thesaurus, organized knowledge a person possesses about words and other verbal symbols, their meaning and referents, about relations among them, and about rules, formulas, and algorithms. For example, do you know what the capital of France is?

Free recall tests are the standard tests of explicit memory. Respondents who previously learned a list of words, are asked to recall as many of them possible. In recognition tests respondents are presented with a list with words they learned before, and some that have not presented before.

Implicit memory effects occur when previous experiences facilitate our performance of subsequent tasks without us remembering the previous experience or being aware of its influence on our performance. In measures of implicit memory – such as word stem completion tests, word fragment identification tests, perceptual identification tests, lexical decision tasks and category instance generation – instructions refer only to the task at hand and make no reference to prior experiences. In word stem completion tests participants are presented with the first letters of each, earlier presented and asked to present the first word coming to mind, word: e.g. com____ for computer. In word fragment identification tests participants are presented with a few letters of the word and asked to name a word that fits: e.g. c_mp_t_r.

In perceptual identification tests participants who are previously presented with a list of words, are presented with another list, comprised of the same and new words, for 35 seconds. They will recall more old than new words. In lexical decision tasks participants are either presented with words or non-word letter strings and are asked to decide as quickly as possible whether the presented item was a word or not.

Category instance generation are more conceptual measures of implicit memory, where participants who have previously learned a list of words containing among others a set of animals, are asked to recall the animals.

Incidental ad exposure should result in implicit memory, whereas the content of ads to which we pay attention should be stored in explicit memory is reminiscent of the levels of processing assumption.

Priming is the most widely used research strategy to demonstrate distinction between explicit and implicit memory. It refers to the phenomenon that exposure to an object or a word in one context increases the accessibility of the mental representation of that object/word in a person’s mind.

As a result, the activated concept exerts for some time unintended influence on the individual’s responses in subsequent unrelated contexts without the individual being aware of this influence. This conception of priming implies the existence of concepts within memory that represent familiar objects/words. When the object/word is presented, this representation will be activated/primed.

Priming is considerably reduced or even absent when the target material is in presented auditory form or when participants study pictorial equivalents of words. Priming cannot only increase the accessibility of single concrete lexical memory location corresponding to the stimulus word, but can activate abstract trait concepts too. Example: when you present respondents with negative terms and then ask them to perceive someone’s behavior, they will tend to be more negative than when they have been presented with positive terms upfront.

One of the most important findings of research on priming that illustrates the distinction between explicit and implicit memory is the differential impact of manipulations of processing depth on explicit and implicit memory measures. These findings showed that a person is less likely to remember a word if he/she has only seen it once than if he/she had to think of the word itself. What was surprising, was that the opposite was true for implicit memory; here performance was best when the word had been presented without a context and worst when it had been generated by participants themselves. Research has also shown that priming can not only increase the accessibility of single concrete lexical memory location corresponding to the stimulus word, but can also activate abstract trait concepts.

In supraliminal priming someone is exposed to the priming stimuli as part of a conscious task and is aware of the stimulus, but not its influence. If a researcher really wants to be certain that participants are unaware of the connection between priming stimuli and subsequent task, they can use subliminal priming. In this case the stimuli are presented at such a brief exposure that participants remain unaware that any stimulus has been presented. To ensure that the prime stimulus does not linger on in the short-term memory and thus be recognized despite the brief exposure, presentation of the priming stimulus is immediately followed by a letter string, a so-called post mask. The finding that subliminal presentation of words or pictures can prime mental representations of words of objects in a person’s mind raises the possibility that such procedures could be used effectively in advertising.

Knowledge structures in long-term memory

To be stored in long-term memory, our perceptions of stimuli and events need to be interpreted and encoded into some form of cognitive representation. We succeed in the encoding task by relating the stimulus to previous knowledge, by recognizing that the stimulus belongs to a particular category (e.g. a product category like cars). Once we have assigned an object to a category, we make inferences about its functions, which go far beyond our perception. According to the classical view categories should be represented as a set of defining features,that apply to all category members. However, most natural categories can’t be defined in terms of such features. Therefore a prototype view of category representation has become accepted. According to this view, people abstract category prototypes from their experience with different category members and then classify exemplars on the basis of their similarity to the prototype. To be able to interpret and encode our perceptions of the world around us, we use scripts, which are abstract knowledge structures that describe standardized sequences of events and the interrelationship between different categories. For example, knowing that the waiter in a restaurant is the one who serves food and who you can ask for the bill, while relying on the restaurant script to derive these predictions.

Associative network models conceive of mental representations of each isolated piece of knowledge as a discrete node connected to other nodes by links of various types. Each concept/attribute is represented by a node. When looking at our knowledge about drinks, there will be nodes for soft drinks, juices, beers, etc. but also for attributes like sweet, tasty, etc. When a node is activated because it is actively though about, the activation of nodes spreads to other nodes via the links. If an individual is primed with the word ‘sweet’, the activation would spread to all the concepts connected to this feature. The more activated a node is, the more likely it will burst into awareness, be recalled, or be applied to incoming information.

What does advertising and its implications entail?

The role of memory in judgments: on the ineffectiveness of traditional measures of advertising effectiveness

One of the standard strategies to measure the effectiveness of an advertising campaign is to contact consumers who were exposed to the ad and ask them what they remember about the ad or the advertised product (recall test). They can also be shown the advertisements and asked whether the recognize it (recognition test).

It is false to assume that people’s product judgments should be related to their recollection of the evidence on which those judgments are presumably based. Using only recall or recognition measures might result in an underestimation of the impact of a campaign on memory because exposure to advertisements can impact on implicit memory and these effects may not be detectable with measures of explicit memory. Also, recollection of arguments made in a campaign may be unrelated to consumers’ attitude towards the advertised product, because the targets of advertising campaigns may change their product judgments online, while being exposed to the campaign. Then they might be influenced by arguments, even if they cannot remember them, or they might still remember arguments, even though they considered them invalid.

Incidental exposure is likely to leave a trace on implicit memory. Assessing the impact of a campaign on memory, whether implicit or explicit, is not an adequate measure of the effectiveness of an advertisement campaign. The aim of such campaigns is to change consumer attitude towards the product in order to increase the probability that they will buy it.

Thus, attitude change achieved by a campaign, rather than memory for arguments contained in the advertisements should be the indicator on which an evaluation of campaign effectiveness should be based. The change in attitudes achieved by a communication is often unrelated to the arguments that are remembered.

Memory-based versus online judgments

In many situations, we are presented with information, which motivates us to form our judgment at the time we are taking in the information. People integrate the information they receive immediately into an overall evaluation. Once the impression has been formed, there is little purpose for them to remember all the traits on which their impression has been based. Similarly, when reading advertisements about brands of a product we are planning to buy, we will probably form an opinion while reading the information. We will not remember all the advantages and disadvantages later on because we were focused on certain attributes. Thus, when we make online judgments, integrating the information while we are exposed to it, our evaluations can be relatively unrelated to the product claims, even though we will still be able to recall them at some later time. If we had no intention of buying the product, but a friend would ask us to advise him, our judgment would be more memory-based.

Studies have shown that there seems to be a lack of relationship between argument recall and product judgment. Carlston gives an dual coding explanation for these findings. According to him individuals store both the original information about the attributes of the products (attribute-based representation) as well as their evaluations based on this elaboration (evaluation-based representation) in separate representations. Each of these may be used independently in subsequent judgments. An evaluation-based representation is retrieved for use in making subsequent global memory-based judgments, whereas an attribute-based representation is retrieved for use in making subsequent discrete memory-based judgments. Since global judgments predict purchase intentions, measures of memory for descriptive information are poor measures of advertising effectiveness.

Memory factors in brand choice: the role of cognitive accessibility

Consumers rarely consider all brands they are aware of, but base their purchase decision on their consideration set, which refers to the set of brands brought to mind in a particular choice situation. The focus is on cognitive accessibility of a brand name as determinant of brand consideration and brand choice.

Purchase situations can encourage more memory-based or more stimulus-based strategies. Buying a car (e.g. choosing to go to a Peugeot dealer) is typically a memory-based decision. Cognitive accessibility influences the composition of the consideration set of car dealers to go to. Shopping in a supermarket with a shopping list is more stimulus-based. Cognitive accessibility will influence the consideration set because highly accessible brands are more familiar.

Priming a brand name increases the cognitive accessibility of that brand name. With totally unfamiliar brand names, priming can increase liking by increasing processing fluency. Priming can also increase the accessibility of closely associated brands and of the product category to which the brand belongs. This is most likely to happen if the brand is prototypical for the product category and if that particular category was previously not highly accessible in the individual’s mind. By reminding people that there is an attractive but non-obvious choice alternative, some are likely to choose this alternative, which they would not have thought of unaided.

The set of brands consumers consider when deciding on a purchase is typically not only smaller than the total number of brands available in the marketplace, but also than the number of brands of which the consumer is aware. Priming that requires a greater degree of conceptual elaboration is more effective for memory-based choices but less effective when choice is stimulus-based. Increasing accessibility of a brand name through priming is likely to influence only the inclusion of a brand in the consideration set rather than the purchase decision. Priming can also influence this decision when (1) individuals have no particular preference for a specific brand; (2) their preferred brand is not available; and (3) an attractive, but less accessible subcategory of a given product category has been primed.

Forgetting the message: advertising clutter and competitive interference

The passage of time is not the only reason why we forget things. Forgetting is also influenced by our experiences in the period between exposure to information and recall, and seems to be accelerated if this period is characterized by exposure to a lot of competing information from the same content domain. Interference theories try to explain this forgetting due to competing information. Retroactive interference is an effect which occurs when interpolated learning interferes with the recall of the original material. Learning nonsense syllables after meaningful adjectives results in much less interference than learning synonyms. Proactive interference occurs when what we have learned earlier interferes with later learning. Advertising clutter refers to the extent to which messages compete for the consumer’s attention.

Several factors should moderate the impact of competitive interference on memory:

• The processing goal of the recipient of the ad information seems to be a factor: if recipients read or view the ad with the goal of potentially purchasing the advertised product, they are likely to process the information more thoroughly than if they look at the ad without any specific interest in the brand that is being advertised.

• Another factor is brand familiarity, which reflects the consumer’s level of (in)direct experience with the product: recipients will have an established knowledge structure for familiar brands. New information will be strongly linked to the brand name.

• With unfamiliar brands, recipients are less likely to link the information to the brand name and more likely to link it to the product category.

• In order to combat interference due to advertising clutter, marketers use retrieval cues, ad repetition or place the ad in an advantageous position within the block of advertisements.

Can advertising distort memory?

Advertising presented after experiencing a product can distort consumers’ memory of their experience. People experience what they expect to experience, at least with products that are not easily evaluated. The effect of advertisements shown afterwards reflects a distortion of individuals’ memories of their evaluation, involving some form of selective retrieval, with possibly a reinterpretation of the retrieved evidence in the light of the ad claims. Post-experience memory distortions are more likely to occur for (rather ambiguous) experience and credence product attributes than for search attributes. People who are familiar with a product and confident in their ability to judge the quality of that product, are unlikely to be susceptible to the distorting influence of advertising messages.