Effects of Emotional Pictorial Elements in Ads Analyzed By Means of Eye Movement Monitoring

Werner Kroeber-Riel, University of the Saarland (W. Germany)
ABSTRACT - In addition to the product, many ads present emotional pictorial elements. In an experiment, the effects of such pictorial elements are studied using also eye movement monitoring which serves to establish intervening cognitive responses occurring during the persuasion process.
[ to cite ]:
Werner Kroeber-Riel (1984) ,"Effects of Emotional Pictorial Elements in Ads Analyzed By Means of Eye Movement Monitoring", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 591-596.

Advances in Consumer Research Volume 11, 1984      Pages 591-596


Werner Kroeber-Riel, University of the Saarland (W. Germany)


In addition to the product, many ads present emotional pictorial elements. In an experiment, the effects of such pictorial elements are studied using also eye movement monitoring which serves to establish intervening cognitive responses occurring during the persuasion process.

Eye movement monitoring informs on the extent of the effected cognitive activities. Furthermore, it indicates the selective information acquisition and processing and it delivers suggestions why product evaluation changes more or less strongly and which directions such changes take.


Beyond words, pictorial motives are better suitable to influence the recipient emotionally.

In advertising, we differentiate between two types of application of emotional pictorial motives. To clarify these types, we refer to ads for consumer goods:

1. type: The emotional picture assumes a  d o m i n a t i n g   role in the ad, thus being the center of attraction and mainly serving to create emotional brand experience as prestige, freshness or eroticism.

2. type: The picture of the product and verbal information determine the ad's content. The picture motive is used to arouse an additional emotional impression to improve the perception of the product thus playing a  p e r i p h e r i c a l   role.

In an experimental study, we have considered this second type of the - peripherical - use of pictures, our central theme being: How effective are additional emotional picture elements in such ads which do their persuasive work on the strength of verbal information and pictures of the product?

The basic hypotheses of the experiment refer firstly to the selective effects of the perception climate on the product judgement created by the emotional pictorial elements; secondly to the effects of the visual organization of the ad. We expect that a pictorial element placed close to the product influences product judgement more than a picture element which is farther away.

In the course of this experiment, results of which are shortly outlined in the next paragraph, we have also pondered a special question: Is it possible by means of eye movement registration to receive information during the process of looking at the ad which cognitive processes are effected through the ad?

Generally, effects of an ad are established through verbal measurements especially through ratings with which perception, attitude, intention to buy, etc. are measured. Thus results of the persuasion processes are procured but nothing what happens during the persuasion in the brain. To learn something about this process, only methods of process analysis are available. Some of these methods are protocols of thinking aloud and thought protocols, with which cognitive responses during processing of ads are obtained (see for example Wright 1980; Olson, Toy, Dover 1982; Kaas, Hofacker 1982). Eye movement recording has rarely been used.

The relationship between variables as considered by us look as follows on the operational level:

presentation of the ad = release of the independent variables "persuasion stimuli"

eye movement monitoring = measurement of intervening cognitive variable

interviewing = measurement of the dependent variable "persuasion success"

Eye movement recording is a psychobiological procedure: Conclusions are drawn from the movements of the eyes to cognitive activity during the visual stimulus processing and during imagery processes. It serves to measure the acquisition of the visual information (Loftus 1982, pp. 258 ff.; see also the overview or Sheikhian 1982). In addition it shows patterns of information processing: "The most general assumption of the current research is that the locus of the eye fixation can indicate what symbol is currently being processed" (Just, Carpenter 1976, p. 471; see also Just, Carpenter 1980; Gaarder 1975).

In marketing research, this procedure was systematically used, above all, by Russo (1978; Russo, Rasen 1975); van Raaij (1977); Leven (1983; Leven, Steffes 1982); Kroeber-Riel (1984; Kroeber-Riel, Bar ton 1980).

Eye monitoring is especially qualified to analyze pictorial information processing (Spoehr, Lehmkuhle 1982, p. 163 ff.). Pictorial information processing observes different rules than verbal information processing. These are rules of analog information processing: The visual organization of the picture plays an important role, especially the spatial relationship between the pictorial elements (Posner, McLead 1982, p. 490; Haber 1981, pp. 6).

The spatial relationship, for example the spatial distance of an emotional picture element from the shown product, determines to a considerable extent. which information chunks are formed and processed. The spatial relationship of the picture elements is, among others, responsible for the manner in which pictorial elements are stored in the brain and which associations are developed among them.


Our experimental data are taken from an experiment which was conducted at the Institute for Consumer and Behavioral Research at the University of the Saarland, examining the interaction of verbal and pictorial effects of ads. Meyer-Hentschel (1983) reports here about another partial experiment.

Among others, ads in four different versions were presented in this experiment. The basic version shows the picture of the product as central point. Furthermore, the ad contains a headline in bold face six words - a text and the brand name. This was an actual, already published ad for an automobile brand (Opel Manta) and another one for a ball point pen (Lamy). These products were chosen with the intention to represent both ads from the durable goods market and ads from the convenience goods market. These different product categories explain considerable differences in perception behavior, for example the objective (functional) judgement of the product auto is more extreme on the average than the corresponding judgement of ball point pens (see below).

Through manipulation of these ads, we received four ad versions per product:

A. Basic version of the ad with picture of the product and text (Group l)

B. Version with an additional pictorial element placed close to the product (Group 2)

C. Version with an additional pictorial element farther away from the product (Group 3)

D. Version with an additional pictorial element presented in an unusual form (Group 4).

In the auto ad, the pictorial element presents a fashionable girl, in the ad for the ball point pen the head of a gentleman is shown. A separate association test proved that the girl has a positive, the head of the gentleman a negative effect ("arrogant type of man"). The pictures were professionally and homogeneously integrated in the ad (experimental control). Only in version D, the same pictorial element was presented in an extravagant and unusual manner: it was cut to pieces and refitted in a wrong way. This pictorial element was supposed to effect cognitive surprise in the subjects and should effect a conscious and critical reflection of the ad

Each version was presented to a group of 40 Ss (N = 160). The Ss saw the ads, for 3 sec. each, on a projection screen. This is the average time for looking at ads. During this time, eye monitoring was carried out. Interviews followed, among others, for recall measurement. The ads were exposed for another 3 sec., then afterwards product judgement was established through ratings.

As concerns measurement of the dependent variables, the following may be said: Eye monitoring was executed with a Japanese NAC-IV-recorder. For the evaluation, the number (frequency) of fixations for the ads and for the ad elements as well as the order of the fixations were established (the duration of fixations was registered as well, but not yet evaluated).

Product judgement was measured with monopolar seven-point rating scales. For once, the used items referred to beliefs on objective (functional) product attributes (as susceptibility to repairs or roadability of the car) and, on the other hand, to the emotional impressions of the product (as "good taste" or "this makes a good impression on other people"). Based on a factor analysis, multi item scores were formed for the functional and for the emotional product perceptions.

Results in short: By a two-factor analysis of variance for the considered groups G1 to G3 the following values are obtained (Table l): Factor l (within Ss) reflects the division into the emotional and the objective product judgement: As concerns the automobile ads, this factor creates significant differences. Objective judgement of the autos is in all groups more extreme than the emotional judgement. In the ads for the ball point pen, no influences of this factor can be ascertained. It is surprising that the objective judgement of this product is not subject to fluctuations in all groups.



Factor 2 (between Ss) reflects the division into groups. The auto ads do not show any significant results. However the ads for ball point pens effect significant differences among groups. Since these are only observable concerning the emotional product judgement, an interaction effect is developed at the same time.

All in all, the differences in product judgement obtained on the average are not consistent and comparatively low. Only few ratings show stronger and significant differences among groups.

In any case, the effects of the auto ads are lower than in preliminary experiments which showed girls close to the cars (Smith, Engel 1968; Kanungo, Pang 1973). This may be attributed to the fact that the emotional pictorial elements in our ads do not dominate as in other experiments but are only of secondary importance. To obtain stronger effects, we are dependent upon repetition effects!

A prolongation of the exposure time of the ad above 3 sec. did not bring any considerably different results. We replicated the experiment with the same stimuli and with the same group sizes. D ring the product judgement, the subjects could look at the ads as long as they wished. The three-factor analysis of variance with the additional factor "experiment" proves that this factor had no significant effect.


The verbally measured product judgement is the end of many cognitive and emotional processes aroused through the ads. It is difficult to ascertain these intervening emotional and cognitive processes. Following, some attempts are described to shed some light in these intervening processes through eye movement recording.


In the line for the sums, Table 2 shows the fixation frequency for the four ad versions. The first fixation of the ad is not implemented, since it is to be regarded as unspecific orienting reaction. Upon introduction of the emotional pictorial element, the fixation frequency per ad increases (same exposure time). Concerning the auto ad, the number of fixations of Group l (without additional element) compared with Group 2 (with additional element) increases from 347 totally to 375 fixations; during exposure of the ad for ball point pens, it increases from 362 to 400. If the additional element is farther away from the product as in ad version C - fixation frequency decreases. But fixation frequency is highest for the last ad version showing the additional pictorial element in a surprising form (the higher fixation frequency of pictures with such additional unfamiliar elements is also proved, among others, by Loftus and Mackworth 1978).



However, the differences in fixation frequencies as quoted above are not yet significant. Nevertheless, they are ascertained in the ads for both products in an entirely consistent manner. We are in a position to interpret these differences, since they correspond with the results of activation research: Fixation frequency can be regarded as indicator for activation effected by the ads. Activation, in short, is the state of inner arousal determining the cognitive performance of the individual. The stronger the stimulated activation, the more efficiently information is acquired, processed, and stored ( Kroeber-Riel 1979).

The dependent increased cognitive activities are expressed in a more frequent fixation behavior. Therefore, fixation frequency of visual elements may generally be regarded as an indicator for the attention and activation induced by these elements. Experimental studies on advertising proved that strongly activating ads were fixated more frequently (Witt 1977). Under certain conditions, not only the activating ad element itself is fixated more frequently, but also other ad elements (Kroeber-Riel 1979, p. 246).

Thus eye movement recording can inform whether or not stronger or weaker persuasion effects originate in activation effects, i.e. in a more or less cognitive performance. Thus, a cause for the persuasion success can be controlled, which is most important then, when there is uncertainty as to the activation potential of the used visual stimulus.

The Distribution of Fixations as Indicator for the Selective Information Acquisition

It was mentioned already that fixation of a visual element is an indicator that this element has been acquired and is processed in the primary memory.

The distribution of fixations to the single elements in Table 2 reflects the distribution of attention when looking at the ad and thus the selective information acquisition.

In the examined ads, Ss in all the groups fixated the product approximately equally frequent - even when the emotional pictorial element is placed close to the product. The pictorial element does not distract attention from the product. This, in turn, refers to the peripherical role of the additional pictorial element.

Very often, persuasion of the attitude cowards the product is decreased or even prevented in so far as the emotional pictorial elements distract from the product. This process can be proved but insufficiently through verbal reports (especially gained by interviewing). Eye movement recording is the only way to prove visual distraction and thus a link in the persuasion process.

The superior effect, which can be derived from the differences in fixation frequency as stated in Table 2, is the holding power, which pictures effect when Ss are looking at a picture-text-mix. Picture elements exceeding a certain minimum size are fixated b e f o r e the text elements in 90 % of all cases. If the fixation time is invariable, the text elements suffer receiving less fixations.

However the preference of pictorial elements to text elements of an ad can also be observed at unlimited looking time because recipients avoid, if possible, the cognitive effort of reading the verbal elements when skimming ads.

The presented eye movement recording clearly shows that the introduction of the picture of the girl in the auto ad reduces the fixations of the entire verbal part of this ad (headline and text) by nearly one third (p = 0.01). In the ad for the ball point pen, the reduction is less drastical. Here the headline does not have to compensate, alone the text (copy) suffers. The holding power of the surprising pictorial element is especially strong and significant in both ads in Group 4: Now text elements are fixated by 40 % less.

Such shiftings of the fixation frequency within an ad deliver suggestions of the relationship between persuasion success and selective information acquisition:

With consideration to the visual organization of ads the neglect or preference of the offered pictorial and text elements can have considerable effects for product judgement: Our experimental results show for example that persons fixating the picture of the girl above average tend to a more emotional product judgement than other persons (contingency coefficient of 0.3, Chi test: p < 0.05). Different - more emotional or more cognitive - judgement processes can thus be explained by the fixation behavior of the Ss at least to some extent.

Fixation Patterns as Indicators of Cognitive Information Processing

Increased cognitive activities do not suffice to develop an efficient persuasion which is made evident in our examination of the auto ad comparing the persuasion effects of Group 2 with those of Group 3. Although we can assume that activation in Group 2 is higher than in Group 3 (please compare section on Fixation Frequency as Indicator for Released Activation) the persuasion effects in Group 3 are stronger, but this is not the case in the ad for the ball point pen.

To obtain criterions for the explanation for these highly indistinct differences of the persuasion effects. we carry out the following calculation:

At first, we determine the extent of the persuasion obtained all in all. Actually the extent of the persuasion can only be ascertained by a before-after-design. However we suppose the following assistance hypothesis: On the average, a stronger persuasion is expressed by higher values of the product judgement in group comparison.

To be in a position to compare the persuasion effects between groups with consideration to this assistance hypothesis, we determine the absolute differences in product judgement of one group compared with the product judgement of our reference group 1 (presentation of the product without emotional additional element). We interpret every deviation of the product judgement of the reference group 1 as result of the ad manipulation (persuasion success) independently whether or not this deviation is positive or negative.





Table 3a shows the results of this calculation: With regard to the automobile ad, Group 3 (girl farther away from the car) scores higher values of persuasion than Group 2. This is reversed in the ad for the ball point pen.

Next, we determine an indicator which we name (relative) "fixation change". It is defined by the number of transitions from one visual element to another one in a fixation sequence. For example, if the fixation sequence is

1, 2, 1, 2, 4, 4, 4, 5

the value for the fixation change is 5. However, this value is only relevant if put in relation to the entire number of fixations in a fixation order. Therefore it is standardized; in the example the value is, 5 of 8 = 63%.

Table 3a also shows the differences in fixation change between groups related to reference group 1. For example, the presented value of 9.3 means: On the average, the fixation change has increased from 39.4 % in Group 1 to 48.7 % in Group 2.

The analysis of variance concerning fixation change as shown in Table 3b establishes a significant group effect. Comparing the group differences of fixation change and persuasion effects suggests the conclusion that fewer fixation changes support persuasion effects!

With other words, increased cognitive activity resulting in additional fixation frequency are only converted into persuasion effects, if information acquisition is not too negligent and unsteady. If elements are only fixated one time shortly and the eye jumps currently from element to element, the acquired visual elements (information) have less effects than those being fixated twice or more - i.e. in longer sequences.

This result is underlined by the fact that we did not find any relations between fixation frequency of the ads and the extent of persuasion stimulated by the ads. Von Keitz (1983, p. 105 ff.) has found corresponding results in an examination on the effects of TV spots: In fact, short cuts in TV spots do activate, however, the more frequent fixation change induced thereby reduces the persuasion success. Von Keitz points out that this connection depends to a great extent on the fact, how difficult the processing of the presented information is.

That the fixation change is a convenient indicator for the extent of the persuasion effect can also be proved on a low aggregation level: Extreme values on the individual ratings of a group for the emotional and for the objective (functional) product judgement are linked to lower values for the fixation change. We found strong relations with significant Chi values. A contingency table is shown in Table 4 as example with the values of the emotional auto judgement (multi item score) in relation to the values of fixation change (rt = 0.62, p = 0.025.

However, we do not yet have detailed knowledge of the conditions, under which the "fixation change" can be used as reliable predictor for the efficiency of visual communication.

Finally, we want to point out a further fixation parameter, which is connected with the fixation change. This is the number of fixations for selected elements appearing at least in a sequence of two. For example, for the elements 1, 7 and 4 in the fixation sequence 3, 3, 1, 2, 2, 1, 7, 4, 4, 4, the value 3 is received. According to our first experiences, a value remaining steady between groups with regard to certain visual elements means a relatively stable (between group-) judgement of the objects represented by these elements.



In addition, there is a positive relationship between the gaze frequency - this is the frequency, by which the same pictorial element (to be exact: the same pictorial information chunk) gets a single fixation or a sequence of fixations - and the memory of this element (Leven 1982, p. 22).


Through different parameters of eye movement recording, it is possible to get insight into the processes of information acquisition and processing during visual persuasion.

For example, the measured fixation behavior reveals the extent of cognitive activities; it indicates the selective information acquisition and it shows why a perception (product judgement) is stable and in which direction it would change.

Research in this field has just started. The presented results are affected by considerable uncertainty. Therefore, our results may be regarded in the context of discovery.

An especially interesting field seems to be the interpretation of patterns of fixation. Obviously elements which are fixated first and are fixated repeatedly in a sequence. play an important role in persuasion effects. Here fascinating possibilities open up to learn more about cognitive processes than that what can be compiled through verbal reports.


Gaarder, K. R. (1975), Eye Movements, Vis:on and Behavior. A Hierarchical Visual Information Processing Model Washington: Hemisphere Publishing Corporation.

Haber, R. N. ( 1981 ), "The Power of Visual Perceiving", Journal of Mental Imagery, 5, 1-40.

Just, M. A. and P. A. Carpenter (1976), "Eye Fixations and Cognitive Processes", Cognitive Psychology, 8, 441-480.

Just, M. A. and P. A. Carpenter (1980), "A Theory of Reading from Eye Fixations to Comprehension", Psychological Review, 87, 329-354.

Kaas, K. P. and T. Hofacker (1983), "Informationstafeln und Denkprotokolle - Bestandsaufnahme und Entwicklungsmoglichkeiten der ProzeBverfolgungstechniken (IDM and Thought Protocols - State of the Arc and Developments), Innovacive Marktforschung, (ed.) 'Forschungsgruppe Konsum und Verhalten', Wurzburg-Wien: Physica-Verlag, 75-103.

Kanungo, R. N. and S. Pang (1973): "Effects of Human Models on Perceived Product Quality", Journal of Applied Psychology, 57, 172-178.

Von Keitz, B. (1983), Wirksame Fernsehwerbung - Die Anwendung der Aktivierungstheorie auf di R Gestaltung von spots (Effective TV Advertising - The Application of Activation Theory on the Creation of TV-Spots), Wurzburg-Wien: Physica-Verlag.

Kroeber-Riel, W. (1979), "Activation Research: Psychobiological Approaches in Consumer Research", The Journal of Consumer Research, 5, 240-250.

Kroeber-Riel, W. (1984), Konsumentenverhalten (Consumer Behavior), 3rd ed., in print, Munchen: Vahlen verlag.

Kroeber-Riel, W. and B. Barton (1980), "Scanning Ads Effects of Position and Arousal Potential of Ad Elements", Current Issues and Research in Advertising, (eds.) J. H. Leigh, C. R. Martin, Jr., Ann Arbor: Graduate School of Business, University of Michigan.

Leven, W. (1983), "Der Zusammenhang zwischen Informationsaufnahme und Informationsspeicherung beim Betrachten von Werbeanzeigen" (The Relationship Between Information Acquisition and Information Storage as Effects of Looking at Ads), Marketing, Zeitschrift fur Forschung und Praxis, 5, 15-28.

Leven, W. and H. Steffens (1982), "Automatisierte Auswertung von Blickverlaufen" (A Computer System for the Processing of Eye Movement Data), Interview und Analyses, 483-484.

Loftus, G. R. (1982), "Picture Memory Methodology", Hand-book of Research Methods in Human Memory and Cognition, (ed.) C. R. Puff. New York. London: Academic Press. 958-987.

Loftus, G. R. and N. H. Mackworth (1978), "Cognitive Determinants of Fixation Location During Picture Viewing", Journal of Experimental Psychology: Human Perception and Performance, 4, 565-572.

Meyer-Hentschel, G. (1983),"Consumer Evaluation of Informative and Non-informative Ads", paper presented to the Proceedings of the 14th ACR Annual Conference in Chicago.

Olson, J. C., D. R. Toy and P. A. Dover (1982), "Do Cognitive Responses Mediate the Effects of Advertising Content on Cognitive Structure?", The Journal of Consumer Research, 9, 245-262.

Posner, M. I. and P. McLead (1982), "Information Processing Models in Search Of Elementary Operations", Annual Review of Psychology, 33, 477-514.

Van Raaij, W. F. (1977), Consumer Choice Behavior: An Information Processing Approach, Ph. D. Thesis, Katholicke Hogeschool, Tilburg University, Tilburg, Netherlands.

Russo, J. E. (1978), "Eye Fixations Can Save the World: A Critical Evaluation and a Comparison Between Eye Fixations and Other Information Processing Methodologies", Advances in Consumer Research, (ed.) K. H. Hunt, Ann Arbor: Association for Consumer Research. 561-570.

Russo, J. E. and L. D. Rosen (1975), "An Eye Fixation Analysis of Multi-alternative Choice", Memory & Cognition, 3, 267-276.

Sheikhian, M. (1982), Pictorial Information Processing: A Selective Review, Working Paper, International Series, Institute for Consumer and Behavioral Research, University of the Saarland, Saarbrucken, W. Germany.

Smith, G. H. and R. Engel (1968), "Influence of a Female Model on Perceived Characteristics of an Automobile," Proceedings of the 76th Annual Convention of the American Psychological Association, 681-682.

Spoehr, R. T. and S. W. Lehakuhle (1982), Visual Information Processing, San Francisco: W. H. Freeman & Co.

Witt, D. (1977), Blickverhalten und Erinnerung bei emotionaler Anzeigenwerbung - eine experimentelle Untersuchung mit der Methode der Blickaufzeichnung (Emotional Advertising: The Relationships Between Eye-Movement Patterns and Memory - Empirical Study with the Eye-MovementMonitor), Ph. D. Dissertation, University of the Saarland, Saarbrucken, W. Germany.

Wright, P. L. (1980), "Message-evoked Thoughts: Persuasion Research Using Thought Verbalizations", Journal of Consumer Research, 7, 151-175.