Memory For Scripts in Advertisements

Christopher P. Puto, University of Michigan
ABSTRACT - Scripts are proposed as an effective method for improving consumers' memories for advertising messages. Research on scripts is reviewed and then related to potential applications for advertisements. The results of a pilot study are given. The pilot study replicated previous findings in psychology regarding memory for scripts but found no effects on measures of beliefs, attitudes, or behavioral intentions. Reasons for these findings are discussed, and suggestions for additional research are given.
[ to cite ]:
Christopher P. Puto (1985) ,"Memory For Scripts in Advertisements", in NA - Advances in Consumer Research Volume 12, eds. Elizabeth C. Hirschman and Moris B. Holbrook, Provo, UT : Association for Consumer Research, Pages: 404-409.

Advances in Consumer Research Volume 12, 1985      Pages 404-409


Christopher P. Puto, University of Michigan


Scripts are proposed as an effective method for improving consumers' memories for advertising messages. Research on scripts is reviewed and then related to potential applications for advertisements. The results of a pilot study are given. The pilot study replicated previous findings in psychology regarding memory for scripts but found no effects on measures of beliefs, attitudes, or behavioral intentions. Reasons for these findings are discussed, and suggestions for additional research are given.


Research in human memory has been a topic of interest for psychologists throughout this century, and the literature in the field bears witness to the multitude of approaches used to study it. Advertising researchers are also very much interested in memory because it is a widely accepted belief among many advertising practitioners that memory for an advertising message is an essential element governing the effectiveness of the advertisement. This paper proposes a relatively new approach to memory research - script theory (Schank and Abelson 1977) - as a potential method for improving consumers' memories for material presented in advertisements.

The first portion of the paper consists of a description and a brief review of script theory. The second section presents the development of a series of hypotheses relating script theory to memory for advertising messages. Finally, the results of a pilot study are presented and discussed, followed by a summary outlining the areas and issues open to additional research.



The early research in human memory was primarily concerned with measuring and understanding individuals' memory processes through the use of verbatim recalls for lists of individual words or, in some cases, letter strings forming groups of nonsense syllables (Ebbinghaus 1885).

Much of what is currently known about human memory processes has come from research studying this facet of memory, termed "simple-event" memory by Wood (1972). Memory for connected discourse as a subject of interest to psychologists is generally credited to pioneering work by Bartlett (1932), who proposed the concept of the schema 88 an organizing framework for the memory of material such as stories and general knowledge. Interest in such concepts as schemas has increased considerably in recent tines due to work in the development of artificial intelligence (AI) systems.

The development of artificial intelligence devices capable of comprehending narratives has long been a goal for researchers working in the AI field, and one of the major problem areas inhibiting progress has been the elliptical nature of everyday language. Typically, such of the detail associated with a communication is omitted by the sender of the message and supplied by the receiver in the form of mental elaborations to the verbal test. Thus, few readers would have difficulty understanding this brief passage:

John came inside, shook the snow from his coat, and sat down by the fire.

While most readers would be able to infer from this one sentence that: (1) it was snowing outside, (2) it was the winter season, and (3) John was cold, most AI systems would not be able to generate these inferences. If, however, it could be shown that much of normal human knowledge regarding everyday events and activities could be reduced to a reasonable set of basic schemas or stereotypical situations, then these schemas could be programmed into the system's memory, thus facilitating its inferential capabilities.

Scripts Defined

Schank and Abelson (1977) developed script theory as one method for accomplishing this. They have extended the schema concept to include the ability to understand behaviors associated with typically encountered, routine situation/action sequences such as going to a restaurant, traveling in an airplane, going to a store, and so forth, which they call scripts. Schank and Abelson refer to a script as the memory structure an individual has for encoding each of these stereotypical activities along with any idiosyncratic variations peculiar to his or her own experiences. Abelson (1976) defines a script as ". . . a coherent sequence of events expected by the individual, involving him either as a participant or as an observer (p. 33)."

Thus, a script is a sequentially ordered schema specifying the events that can normally be expected to occur given a set of specific situational cues. Scripts are quite closely related to the concept of episodic memory (Tulving 1972), and they can be thought of as conceptual mechanisms for facilitating the encoding and retrieval of episodic memory events (Abelson 1981; 1976).

Scripts as Perceptual Aids

Scripts would appear to be ideal mechanisms for aiding individuals in the monitoring and perceiving of multiple incoming stimuli without having to devote full attentional capacity to each incoming stimulus. In his discussion of perceptual readiness, Bruner states that ". . . perception is a process of categorization in which organisms move inferentially from cues to categorial identity and that in many cases . . . the process is a silent one (1957, p. 129)."

It is reasonable for some scripts to be of such a well-learned, routine nature that an individual can engage in (or observe) much script-dictated behavior without necessarily engaging in active cognitive processing at every point in the script. This is conceptually similar to the oft-cited cocktail party situation (Cherry 1957; Kahneman 1973) in which a person attending to one conversation is able to monitor other nearby conversations for such items as the mention of his/her own name. Thus, people are hypothesized to possess scripts for routine behaviors, and furthermore, engaging in or observing routine scripted behavior does not require extensive cognitive processing.

Memory for Scripts

In an important contribution to script research, Bower, Black and Turner (1979) empirically investigated the existence of scripts, and they concluded that people do have general agreement on the content and sequence of events for such activities as eating in a restaurant or visiting a doctor, just to name two of the several activities explored in their research. Similar results were reported by John and Whitney (1982a) for visiting a McDonald's fast food restaurant.

In addition to collecting script norms, Bower, et al. (1979) conducted several experiments investigating people's memories for scripts and for script related activities. In their first memory related experiment (1979, Experiment 3), they had subjects read nine basic scripts, and after twenty minutes of distractor tasks, the subjects' correct recall of eight specific items from the test scripts ranged from 28: to 382. Intrusion, e.g., false recall of normal script actions which had been deliberately excluded from the test scripts, ranged from 7% to 11%.

In another script memory experiment, Bower, et al. (1979, Experiment 7) explored the effect deviations from scripts (e.g., interruptions ant/or irrelevancies) have on memory for a passage describing scripted behavior. Interruptions can be further subdivided into obstacles, errors, and distractions (Bower, et al. 1979). Obstacles occur when some enabling condition for an imminent action is missing (e.g., you are in a fast food restaurant, and you can't see well enough to read the menu board); so a corrective action is taken (e.g., you ask the counterperson to tell you what they have). An error occurs when a script action leads to an unexpected or inappropriate outcome (e.g., you order a large chocolate milkshake, and you are served a medium soft drink instead). A distraction is an unexpected event or state which sets up new goals for the script perceiver, taking him or her temporarily or permanently outside of the script (e.g., instead of asking for your order, the counterperson may inform you that you have "ring-around-the-collar," necessitating that you change your original action plan and come up with an appropriate rejoinder). The other form of nonscript item is an irrelevancy, which is defined as ". . . something that can occur in parallel with essential actions without impeding the flow of events (Bower, et al. 1979, p. 210)."

In the experiment testing the memory for script deviations, subjects read six script-based stories, engaged in a ten minute distractor task, and then responded to a surprise recall test of the six stories, using the story title as the only recall cue. The results were as follows: interruptions were recalled at 53: correct; script actions were recalled at 38: correct; and irrelevancies were recalled at 322 correct. The interruption scores were further broken down into obstacles (60% correct), distractions (56% correct), and errors (42% correct).

It is unfortunate that examples of the stimulus scripts for this particular experiment were not given because the results as presented do not enable the reader to discern the extent of the memory for the content of the interruption. For example, if an interruption involved a conversation, it is important to know whether the memory trace only indicated that a conversation took place, or whether it also included the specific content of the conversation. The significance of this for the use of scripts in advertising is discussed below.


Memory for Advertisements

An advertising message is a persuasive communication containing information about a company, product, service, together with the benefits or reasons why the consumer should purchase the advertised product or service. While the memory for an advertising message is only one element in the overall consumer decision making process, it can nonetheless have an effect on the buying decision. For example, decision factors such as the degree to which a consumer engages in external information search, the number and identities of the brands comprising the consumer's evoked set, and the salience and valence of the product attributes are all subject to influence by advertising. Thus, a realistic set of advertising goals might include the goals of (1) increasing consumers' memories for the factors which differentiate the advertised brand from its competitors And (2) having consumers retain a favorable evaluation and purchase disposition toward the brand.

the research reviewed in the preceding section suggests that scripts may be instrumental in enabling advertisers So achieve these objectives. For example, one way to enhance consumers memories for a message is to invoke a script and then interrupt it with the particular information that the advertiser wishes the consumer to remember.

In their discussion of script interactions, Schank and Abelson (1977) suggest that deviations from the standard Script require an individual to decide whether the script has ended and a new script has begun, or whether he or she h s only encountered a deviation from the standard script which can then be reused. This requirement for a decision heightens the individual's attention level and increases his or her perceptual awareness until the issue of the appropriate script has been satisfactorily resolved. The increased attention and perceptual awareness will result in better memory for the events that occur during that time. It follows from this that advertisements containing a script interruption will produce higher recall levels for the information presented in the interruption than for the related script material.

Another question concerns the strength of the interruption and the effect it will have on recall. The preceding discussion suggests that the more intense the interruption, the more attention it will command, and hence, strong interruptions should produce stronger recall effects than moderate interruptions. Strong interruptions are represented by behaviors which obviously have no place in the script (e.g., walking into a restaurant and having the head waiter announce that people with dandruff are not permitted to eat in that establishment). Moderate interruptions take the form of conversational asides between the participants in the script.

A third issue which the current theory does not specifically address concerns whether it is more effective to invoke a script which is typically associated with the given product class and then interrupt that script with the message about the advertised brand, or to instead invoke an entirely unrelated script and then interrupt it with the message about the advertised brand. An example should clarify this point.

One way to advertise a fast food restaurant is to begin the ad with a script set in a fast food restaurant and then interrupt that (typical) script with the counterperson singing a song about the quality of the food or some other issue of perceived importance to the fast food customer. An example of the second approach would be to begin the at with a script not at all related to fast food, such as flying in an airplane, and then interrupt that (atypical) script with the message about the food quality, etc., at the fast food restaurant. Since script theory is mute with respect to the application of these two types of scripts, a hypothesis of no difference will be tested.

Scripts and Attitudes

A final question concerns the effect of scripts on consumers' attitudes toward the advertised brand. In script theory, Abelson approaches the issue by defining an attitude toward an object as an ". . . ensemble of scripts concerning that object (1976, p. 41)." He interprets this to mean that an individual's attitude toward an object is the result of a series of episodic scripts relating that person's experience either in actual or vicarious terms with the attitude object. When prompted for an attitudinal response, the individual recalls these episodic scripts and responds according to the positive or negative nature of the memories evoked. A series of slice-of-life commercials in a television advertising campaign would be one way that consumers' attitudes could be influenced by scripts. Each different slice-of-life commercial could represent a favorable, vicarious experience with the advertised brand. Testing this longitudinal approach is, however, beyond the scope of this brief study.

John and Whitney (1982b) have taken a slightly different approach to the role of scripts- in attitude formation. They hypothesized that plausible script actions placed out of their normal sequence would tend to be remembered as having occurred twice, once where they were actually placed in the stimulus script and again in the form of an intrusion into their proper place in the normal script during recall. Based on an availability-valence interpretation, the increased frequency of these recalled actions would give them more influence in the attitude determination process, and the positive or negative valence would determine the direction of the attitude shift.

One of their tests of this hypothesis involved the script action of waiting in line at Mcdonald's. The standard script calls for people to wait in line to place their order. John and Whitney moved the position of this action in their test script so that it was presented as waiting in line after returning for condiments. When subjects recalled the script, they recalled waiting in line for condiments, and they inserted another waiting in line at its normal place in the script. As expected, the negative valence of the action combined with its higher frequency to produce less favorable attitudes toward McDonald's.

While the above reasoning is theoretically sound, it is unlikely that an advertiser would ever use scripts in the manner employed by John and Whitney (1982b), and they in fact make no inferences or claims regarding such usage in their presentation. However, their results suggest the hypothesis that, in conjunction with the hypothesized effect of greater memory for script interruptions, increasing the recall of a positively valued attribute by including it in a script interruption will result in an increased belief that the advertised brand possesses that attribute. This should produce a more favorable attitude toward the brand (cf. Fishbein and Azjen 1975).

Summary of Hypotheses

Stated in research form, the hypotheses for applying script theory to advertisements are as follows:

H1: The content of interruptions to scripts describing a typical product usage situation will be recalled at a higher rate than will the standard elements of the scripts.

H2: Recall for the content of strong script interruptions will be higher than recall for the content of moderate script interruptions.

H3: Recall for product attributes presented in interruptions to scripts typically associated with a product will be the sane as recall for product attributes given in interruptions to scripts in which the product does not typically belong.

H4: Script interruptions featuring a positive attribute will produce favorable attitude scores for the brand.



The basic design of the pilot study was a two (script type) by two (interruption) factorial. The goal of the test scripts was to promote a new fast food restaurant. Script type refers to the setting used in the script, which was either typical (e.g., a script about going to a fast food restaurant) or atypical (e.g., a script about attending a course lecture). The interruption was either strong or moderate and was balanced across the two script conditions. The dependent variables were the percentage of script actions and interruptions correctly recalled and responses to three belief, attitude, and behavioral intention measures.

Stimulus Materials

Because this was a pilot study, it was not deemed feasible to develop representative advertisements for each experimental condition. Rather, four written scripts were developed and presented in a story format similar to the way one might be expected to retell a slice-of-life television commercial to a friend who had not seen it. The four stimulus scripts are given in table l.

The scripts were developed in the following manner. Using the same instructions given in Bower, et al. (1979) and John and Whitney (1982a), a convenience sample of twelve undergraduate psychology students generated scripts for visiting a fast food restaurant. A basic script was then developed using the script items that achieved 30: or greater agreement across the 12 subjects. The content of the script items and the relative frequencies of the items across subjects were remarkably similar to those reported by John and Whitney (1982a), and in the interest of conserving space, they are not reported here. A fast food restaurant was selected for the product class because it represents a product category and consumption behavior familiar to a majority of the members of the pilot study subject pool.



Script 1-A

Fast food Restaurant

(Strong Interruption)

Two friends were driving along at lunchtime when they decided to eat at a new fast food restaurant called WUV's Hamburgers. They parked the car and went inside. They stood in line and read the menu board. They were discussing what to order when without warning both the people in front of them and the people in back of them simultaneously interrupted the two friends' conversation and told them that all the food - even the french fries was fresh, not frozen like in other fast food chains. They took their food into the dining room and sat town at an empty table. After they finished eating, they left the restaurant and trove back to the camPus.

Script 1-B

Fast Food Restaurant

(Moderate Interruption)

Two friends were driving along at lunchtime when they decided to eat at a new fast food restaurant called WUV's Hamburgers. They parked the car and went inside. They stood in line and read the menu board. They were discussing what to order when they overheard one of the other customers remark to her companion that all the food - even the french fries--was fresh, not frozen like in other fast food chains. Soon it was their turn, and they placet their order. When it was ready, they took their food into the dining room and sat down at an empty table. After they finished eating, they left the restaurant and drove back to the campus.

Script 2-A

Attending a Course Lecture

(Strong Interruption)

Two friends arrived at the classroom just in time for their 12:20 class. They went into the room, found two empty seats, sat down, took out their notebooks, and got ready to start taking notes. Just as the lecture was beginning, another student came in hurriedly and sat down nest to them. Instead of getting out her notebook, she took out her lunch and started eating it. She told them that it came from a new fast food restaurant called WUV's Hamburgers, and that all the food--even the french fries--was fresh, not frozen like in other fast food chains. When the lecture was finished, they gathered up their things and went to their nest class.

Script 2-B

Attending a Course Lecture

(Moderate Interruption)

Two friends arrived at the classroom just in time for their 12:20 class. They went into the room, found two empty seats, sat down, took out their notebooks, and got ready to start taking notes. When the lecture was finished, they closed their notebooks and began gathering up their things. While they were in the process of doing this, they overheard one of the other students remark to her companion that she had lunch at a new fast food restaurant called WUV's Hamburgers, and that all the food - even the french fries - was fresh, not frozen like in other fast food chains. The two friends finished gathering up their things and went to their nest class.


Script 1-A represents the typical fast food restaurant script with a strong interruption containing a message about the advertised brand. Script 1-B represents the same script with a moderate interruption. Script 2-A -represents a nonproduct-related script (attending a course lecture) with a strong interruption containing a message about the advertised b and, and script 2-B represents the same script with a moderate interruption. the items comprising the course lecture script were adapted directly from those collected by Bower, et al. (1979, p. 182).


Subjects were 83 students in an introductory psychology course who were participating in partial fulfillment of a course requirement. The subjects were tested all at one time. They were informed that the purpose of the research was to collect script norms for future research in understanding television commercials. The experiment consisted of two phases.

In the first phase, each subject was given a booklet containing a brief description of what the experimenter meant by a script, together with the following stimulus materials: (1) one of the four experimental scripts described in table 1; (2) a "visiting a doctor" script (exactly as given in Bower, et al. 1979, p. 182); (3) a set of instructions to develop two scripts of their own (one for a typical beer commercial and one for a typical instant coffee commercial); and (4) a 25-item self-monitor scale (Snyder 1979). This last item was included primarily as a distraction task. They were instructed to work at their own pace and to use the "visiting a doctor" script as a prototype in developing their own two scripts. This first phase lasted approximately 45 minutes.

As soon as the booklets were collected, the second phase began. It consisted of a surprise recall test in which the subjects were instructed to reproduce verbatim, if possible, the first script they read at the beginning of the session. The only cue given was the title of the script, either "Fast Food Restaurant" or "Attending 8 Course Lecture," corresponding to the title of their test stimulus. Where they could not remember the exact words, the subjects were advised that it was permissible to supply the gist of the material.

After the recall protocols were completed, the subjects' final task was to complete a brief questionnaire consisting of a series of 7-point belief and attitude items and one 7-point behavioral intention measure. The three measures used in this analysis are (1) subjects' beliefs that the food at WUV's is fresh (anchored by "Very Fresh" and "Very Stale"); (2) Subjects' attitudes toward WUV's as a place to eat (anchored by "Very Good" and "Very Bat"); and (3) subjects' intentions to eat at WUV's (anchored by "Definitely Will" and "Definitely Will Not").


Although each of the scripts contained exactly 116 words, they varied somewhat in the number of propositions (basically clauses). In order to maintain consistency of measurement, the number of propositions was tabulated for each category of interest in the research (e.g., script actions and interruptions) within each script. The recall protocols were then scored for the presence of each proposition, and the raw scores were converted to percentages of each category correctly recalled within each script.

The results of a multivariate analysis of variance are reported in Table 2. The percentages of correctly recalled script actions and correctly recalled interruptions are the dependent variables, and type of script and type of interruption are the treatment variables. There were no significant differences for the main effects (largest F = .13, P > .80), and the interaction was also nonsignificant (F = .30, P > .70). The absence of a main effect for type of interruption provides no empirical support for H2, i.e., strong interruptions not better recalled in this study than are moderate interruptions. Recall that H3 was, in effect, a null hypothesis, and the absence of a main effect for script type supports the hypothesized prediction of no difference between recall of interruptions for typical and atypical scripts.



The within-subject differences between recall for script actions and script interruptions were tested separately for each of the four cells using a paired t-test. For clarity of presentation, the data in Table 2 are the raw percentages of the propositions correctly recalled for the script actions and the script interruptions. Significance tests were conducted using the arc sine transformed percentages, and the results are virtually identical to those reported in the table. In each cell, the interruptions were recalled better than the normal script actions. Moreover, the pattern is repeated across each of the four cells, thus adding a degree of robustness to the support for H1.

The responses to the belief, attitude, and behavioral intention measures are given in Table 3. Three separate analyses of variance were computed (one for each dependent variable), and they revealed no significant effects (largest F - 0.65; P > .59). Thus, there is no empirical support for H4. The means for each condition (adjusted for differences in cell size) are presented in the table to provide the reader with an indication of the directionality, which differs for each of the three variables examined.




The major finding in this pilot study was that script interruptions are recalled better than standard script actions. This replicates the earlier results of Bower, et al. (1979). However, since the protocols in the present study were coded for content, these results also indicate that the content of the interruptions is better remembered than the standard script actions. This suggests that advertisers can improve consumers' memories for key points in advertisements by inserting them as interruptions in typical scripts.

The lack of a difference between the moderate and the strong interruptions should be treated with extreme caution because the pilot study involved the reading of written scripts and used subjects with high verbal skills. "Real" consumers may process real advertisements differently than these subjects processed the experimental scripts. Additionally, in the absence of a manipulation check, these findings may be due to the lack of a true difference in "strength" for the two categories of interruptions as operationalized in this pilot study.

The finding of no memory difference between typical and atypical scripts should meet with agreement among advertising practitioners because this reflects the current state of many advertising campaigns. Some commercials are based on typical scripts (eeg., McDonald's), and others are based on atypical scripts (e.g., Wisk). These findings suggest that memorability should be comparable for both types.

The lack of significant findings for the belief, attitude, and behavioral intention measures is most likely tue to the informal design of the pilot study sat to the inadequacy of the measures themselves and would not be given much weight. For example s the study did not take premeasures on these variables, nor did it include a control group. Also, there was no effort made to balance the attitude manipulation with negatively directed stimuli. Rather, these findings point to the need for additional research to identify the relationship between the memory for key points in advertisements and the corresponding effect, if say, on attitudes toward the advertised brand.

One approach for doing this would be to content analyze existing advertisements for script related elements such as standard script actions and interruptions. These could then be used as "treatment" or predictor variables, with recall and attitude change measures as the criterion variables. Similar research investigating other forms of distraction, e.g., obstacles and errors, also represents an opportunity for increasing our understanding of the role of scripts in the marketing communications Process.


Research was reviewed which suggested that script theory might be a worthwhile approach to examining advertising effectiveness. The results of the pilot study suggest that the memorability of key points in an advertisement can be enhanced by presenting the points as interruptions to scripts. The absence of positive results for the attitude and behavioral intention measures indicates the need for further research on this aspect of scripts, but the script concept still offers much potential for researchers seeking to understand and improve the effectiveness of marketing communications.


Abelson, Robert P. (1981), "Psychological Status of the Script Concept," American Psychologist, 36, 715-29.

Abelson, Robert P. (1976), "Script Processes in Attitude Formation and Decision Making," in J. Carroll and J. Payne (eds.), Cognition and Social Behavior, Hillsdale, N.J.: Erlbaum, 33-45.

Bower, Gordon a., John B. Black, and Terrence J. Turner (1979), "Scripts in Memory for Test," Cognitive Psychology, 11, 177-220.

Bruner, Jerome S. (1957), "On Perceptual Readiness," Psychological Review, 64. 123-52.

Cherry, E C (1957), On Human Communication: A Review, A Survey. and A Criticism, Cambridge: Technology Press. MIT.

Ebbinghaus, R. (1885), Memory new York: Dover, 1965.

Fishbein, Martin and Icek Azjen (1975), Belief, Attitude. Intention. and Behavior: An Introduction to Theory and Research, Reading, Mass: Addison-Wesley.

John, George and John C. Whitney (1982a), "An Empirical Investigation of the Serial Structure of Scripts," in B.J. Walker, W.O. Bearden, W.R. Darden, P.E. Murray, J.R. Nevin, J.C. Olson, and B.A. Weitz, (eds.), An Assessment of Marketing Thought and Practice, Chicago: American Marketing Association 75-79.

John, George and John C. Whitney (1982b), "An Empirical Investigation of Memorial and Attitudinal Effects of Script Processing in a Consumer Contest," Working Paper, University of Wisconsin-Madison.

Kahneman, Daniel (1973), Attention and Effort, Englewood Cliffs, N.J.: Prentice Hall.

Schank, Roger and Robert Abelson (1977), Scripts. Plans. Goals and Understanding, Hillsdale, N.J.: Erlbaum.

Snyder, Mark (1979), "Cognitive, Behavioral, and Interpersonal Consequences of Self-Monitoring," in P. Plimer, R.R. Blankstein, and I.M. Spigel (eds.), Perceptions of Emotion in Self and Others, new York: Plenum Press, 181-201.

Tulving, Endel (1972), "Episodic and Semantic Memory," in E. Tulving and W. Donaldson (eds.), Organization of MemorY, New York: Academic, 382-403.

Wood, Gordon (1972), "Organizational Processes and Free Recall," in E. Tulving and W. Donaldson (eds.), Organization of Memory, new York: Academic, 49-91.