Scanning Ads: Effects of Involvement and of Position of the Illustration in Printed Advertisements

Carmen Garcia, Universidad Autonoma de Madrid
Vicente Ponsoda, Universidad Autonoma de Madrid
Herminia Estebaranz, Universidad Pontificia Comillas de Madrid
ABSTRACT - Our objective was to study the effect of involvement and of the position of the ilustration on the attention received by printed advertisements. Two experiments were carried out. In the first experiment involvement with the advertisement was manipulated through the instructions; in the second, through a product test. In order to measure attention, subjects’ eye movements were recorded. We found an interaction between the position of the illustation and involvement. Position of the illustration had no effect on either number or duration of fixations on it when involvement was high, but it had when involvement was low. This result indicates the need to allow subjects to have low involvement during advertisement tests.
[ to cite ]:
Carmen Garcia, Vicente Ponsoda, and Herminia Estebaranz (2000) ,"Scanning Ads: Effects of Involvement and of Position of the Illustration in Printed Advertisements", in NA - Advances in Consumer Research Volume 27, eds. Stephen J. Hoch and Robert J. Meyer, Provo, UT : Association for Consumer Research, Pages: 104-109.

Advances in Consumer Research Volume 27, 2000      Pages 104-109

SCANNING ADS: EFFECTS OF INVOLVEMENT AND OF POSITION OF THE ILLUSTRATION IN PRINTED ADVERTISEMENTS

Carmen Garcia, Universidad Autonoma de Madrid

Vicente Ponsoda, Universidad Autonoma de Madrid

Herminia Estebaranz, Universidad Pontificia Comillas de Madrid

ABSTRACT -

Our objective was to study the effect of involvement and of the position of the ilustration on the attention received by printed advertisements. Two experiments were carried out. In the first experiment involvement with the advertisement was manipulated through the instructions; in the second, through a product test. In order to measure attention, subjects’ eye movements were recorded. We found an interaction between the position of the illustation and involvement. Position of the illustration had no effect on either number or duration of fixations on it when involvement was high, but it had when involvement was low. This result indicates the need to allow subjects to have low involvement during advertisement tests.

The aim of advertising is to make the public aware of a product or service, and to induce its purchase or use. Effective advertising would therefore be that which achieves this objective. However, although influencing purchase is the final objective, it is very difficult to obtain a direct measure of the influence of an advertisement on sales, since advertising is just one of the numerous factors that influence buyers. For these reasons, other indicators of the capacity of the advertisement for producing the desired change have been sought. One of these indicators is the capacity of the advertisement for capturing the attention of the audience or readers. The current saturation of the communications media has turned this quality of advertising messages into a fundamental one.

The measures of attention usually employed in this context are recall and recognition of the advertisements, but these measures involve not inconsiderable problems of validity (Garcfa and Ponsoda, 1997a). A more valid indicator of visual attention is provided by the recording of eye movements. It is assumed that eye fixations, that is, the periods between optokinetic movements during which the eye remains more or less still, coincide with the points of the visual field that are being attended (Viviani, 1990).

Various authors have used this measure as an indicator of the amount of attention given to an advertisement (Bogart and Tolley ,1988; Kroeber-Riel (1987); Krugman et al (1994); von Keitz (1989); Lohse (1997); Garcfa and Ponsoda, 1997a, 1997b; Rosbergen, Pieters and Wedel, 1997;Treistman and Gregg (1979); Young, 1984). A detailed review can be found in Garcfa and Ponsoda (1997b). Here we shall discuss only those aspects that directly concern this research.

Kroeber-Riel and colleagues at the Institut fnr Konsum- und Verhaltensforschung of the Saarland UniversitSt in Germany studied the influence of the disposition of the text and illustrations in printed advertisements. Bernhard (cited in Kroeber-Riel and Barton, 1980) studied the effect of text position. To this end this researcher designed four advertisements for unknown brands (two per condition) in which the text could occupy (a) the top left part or the bottom right part, and (b) the top right part or the bottom left part. As expected, the text placed in the upper half of the page on the left received more fixations than that situated in the lower half on the right, and the text in the top half on the right received more fixations than that placed in the bottom half on the left. These hypotheses were based on the results of Brandt (1941) and Niekamp (cited in Kroeber-Riel and Barton, 1980), who, on using simple stimuli, such as nonsense syllables, found that certain quadrants of the stimulus received more attention. Specifically, the syllables in the top left quadrant were those that received fewest fixations, and those of the top right quadrant those that received most. This differential distribution of resources was attributed to the western cultural reading convention of starting at top left and finishing at bottom right.

Bernhard also studied the effect of text position with respect to the illustration in the advertisement, designing four advertisements (two per condition) in which the text could appear (a) beneath the illustration or above it, (b) to the right of the illustration or to the left of it. As expected, when the text appeared beneath the illustration it received more fixations than when it appeared above it; the superiority of right position versus left position did not reach significance.

Bernhard¦s hypotheses were based on work by Guttman (1972), who, in a tachistoscopic study, found that the illustration was the first element of an advertisement to be processed. Thus, the image would first attract attention, and the transition to the text would prove esier when the movement was culturally facilitated.

The results related to recall were not always the expected ones:

a) if the text was situated at top left it was recognised better than if it was at bottom left, but the differences did not reach statistical significance in all the advertisements.

b) no differences were found between the positions top left and bottom right.

c) as expected, in terms of number of fixations received, only when the text was situated beneath the illustration was it significantly better recognised.

The divergence between the results for recognition and those for number of fixations can be explained in various ways. Bernhard attributes it to the fact that the standard exposition time of 5 seconds for all advertisements may have cancelled out some differences, given that they were very simple advertisements. It is also possible that the sensitivity of this variable for detecting this type of effect is not as high as implicitly assumed, or that the relationship between recall and fixations is not monotone and increasing, but rather that there is a minimum number of fixations after which everything is remembered, or after which the relationship between number of fixations and recall begins to increase.

In sum, it was found that the position of the text in the advertisement affected the attention it received. Text placed in the top left quadrant and beneath the illustration received most attention, but although when the text occupied these positions it was recalled better, the differences were not always significant. Meanwhile, text in the bottom left quadrant and placed above the illustration received least attention.

In our own research we also proposed to study the effect of position of illustrations in advertisements on attention received, but with two important differences with respect to previous studies: the first difference was that, in addition to recording the number of fixations, we also noted their duration. Number of fixations is considered as an indicator of the quantity of information acquired, whilst duration indicates depth of processing (Tversky, 1974). The second, and more important difference concerned our interest in how the position of the illustrations and involvement with the advertisements interacted in order to produce effects on attention. Involvement is defined as the personal relevance an advertisement has for its audience (Zaichkowsky, 1985). Several authors (Celsi and Olso, 1988; Garcfa and Ponsoda, 1997a; Pratkanis and Greenwald, 1993; Rosbergen, Pieters and Wedel, 1997 and Thorson and Page, 1988) have found a significant effect of involvement on attention. High involvement produces a higher level of attention. A detailed review can be found in Garcfa and Ponsoda (1996).

In line with the above, it would be expected that, independently of its position, the illustration in an advertisement will always receive the first fixations. When involvement with an advertisement is high the person will search actively for information, and will presumably take notice of all information available, be it text or image. The text will be more accessible when it is to the right of or beneath the illustration, but if there is an interest in obtaining information, any disadvantages in this respect should not matter. However, when involvement is low there will be no active search for information, so that it would be expected that the eyes will sweep over the stimulus guided by the inertia derived from the reading pattern, so that, when the text position does not make it easily accessible (on the left or above the illustration) the exploration will be concentrated on the image. In accordance with this idea our hypotheses are as follows:

HYPOTHESIS 1: There will be an interaction between involvement and position of the illustration (above versus below). In the high involvement condition, the number and duration of fixations produced on the image, independently of its position, will be similar. Nevertheless, in the low involvement condition the number and duration of fixations will be higher when the image occupies the lower half of the advertisement, given that movement towards the text will be more improbable.

HYPOTHESIS 2: There will be an interaction between involvement and position of the illustration (right versus left), so that, in the high involvement condition, a ceiling effect of attention will prevent us from finding differences related to position of the illustration. However, in the low involvement condition, we shall find a higher number and duration of fixations when the image is on the right side of the advertisement.

EXPERIMENT 1

Method

Subjects

Pilot Experiment. The sample was composed of 10 Psychology students, 5 males and 5 females, aged between 20 and 23 years.

Principal Experiment. The sample consisted of 32 subjects, 15 males and 17 females. None of the subjects had any visual defects that could affect the recording of their eye movements.

Materials

Eye movements were recorded by means of an EYE TRACK CAMERA MODEL 210 A.S.L. For the recording and analysis of this data we used the software developed by Garcfa, Garcfa & Ponsoda (1993).

Stimuli. 38 colour slides were projected. 16 were adverts and 12 showed gardens, whilst the remaining 10 showed the two elements together.

Procedure

Pilot Experiment. Before carrying out the Principal Experiment, we carried out a Pilot Experiment in which subjects were shown, with no time limit for viewing, the 10 slides showing the advert and the garden together. The object of this experiment was to decide exposition time for the Principal Experiment. Mean viewing time was 5.3 seconds, with no significant differences between the slides.

Principal Experiment. This began with the positioning and calibration of the sensors for recording the eye movements. Once the calibration was complete, subjects heard the instructions they were to follow during the showing of the 10 slides, and the slides were then projected. The instructions were designed to manipulate involvement. In the high involvement condition subjects were asked to pay special attention to the adverts, as they would be asked a series of questions after the projection, 90% of them about the adverts. In the low involvement condition they were asked to pay special attention to the gardens, and informed that 90% of the questions would be about them.

After the reading of the instructions subjects viewed, for 5 seconds, each of the slides, whilst their eye movements were recorded.

Immediately after this they were given a booklet containing the rest of the measures for the experiment. These questions, along with the results deriving from them, can be found in Garcfa and Ponsoda (1997a and 1997b).

Results

We present the results of the two advertisements separately, since they differ in an important aspect: knowledge of the brand.

1. Advertisement for the beauty cream "Valeris".

As predicted, we found a sigificant interaction between involvement and position of the illustration in the expected direction. The results can be seen in Table 1.

The illustration of the advertisement receives a significantly higher number of fixations in the low involvement condition with the image below [F (1, 28)=6.46, p<0.02]. As regards the total duration of fixations we found the same pattern of results [F (1, 28)=4.99, p<0.03]. On considering mean duration of the fixations we found only a principal effect of involvement. The mean duration was significantly higher in the high involvement condition [F (1, 28)=5.76, p<0.02].

We carried out a sequential analysis, taking as the unit of analysis the pair of fixations image-image; more specifically, we considered that there had been an image-image sequence when in the course of the recording there appeared two consecutive fixations on the illustration of the advertisement. There are more image-image sequences when involvement with the advertisement is low, particularly when the image occupies the lower half, though statistical significance is not reached.

2. Advertisement for the cava (Spanish champagne) "Codorni·".

In this case no significant effects were found for any of the variables studied. However, although the differences failed to reach significance, all of the effects were in the expected direction. The number and total duration of the fixations on the image were higher when involvement was low and the illustration occupied the lower half of the advertisement. The sequential analysis of image-image fixations shows a pattern of results analogous to that of the previous advertisement. It is in the low involvement condition with image below that the highest number of sequences is produced, though statistical significance is not attained. Table 2 shows all of the results.

Discussion

With this first experiment we tested the first of our hypotheses. The results, though not always reaching statistical significance, were in the expected direction for both advertisements. However, in our opinion this experiment has an important limitation. The procedure used for manipulating involvement is quite artificial, and would be very difficult to use in applied contexts. We therefore designed a second experiment with greater ecological validity, and which, moreover, tested our second hypothesis.

TABLE 1

MEAN OF EYE FIXATIONS ON THE IMAGE OF THE ADVERTISEMENT "VALERIS BEAUTY CREAM" AS A FUNCTION OF ITS POSITION AND OF INVOLVEMENT LEVEL (I.)

TABLE 2

MEAN OF EYE FIXATIONS ON THE IMAGE OF THE ADVERTISEMENT "CAVA CODORNIU" AS A FUNCTION OF ITS POSITION AND OF INVOLVEMENT LEVEL (I.)

EXPERIMENT 2

Method

Subjects

First Pilot Experiment. The sample was made up of 10 female Psychology students, aged between 21 and 25 years.

Second Pilot Experiment. The sample was made up of 16 female Psychology students, aged between 21 and 25 years.

Principal Experiment. The sample was made up of 16 female Psychology students, aged between 21 and 24 years. None of the subjects had any visual defects that could affect the recording of their eye movements. Eight women regularly drank coffee, and were assigned to the "high involvement with coffee" condition, whilst the remaining eight all confessed to having a favourite perfume, and described themselves as being interested in cosmetics and perfumes. These subjects were therefore assigned to the "high involvement with perfume" condition.

Materials

For the recording and analysis of eye movements and the slides projection, the same material as that for Experiment 1 was used.

Stimuli. 20 colour slides showing adverts were projected: 10 for the experimental session, and these plus 10 others for the recognition tas.

TABLE 3

MEAN OF EYE FIXATIONS ON THE IMAGE OF THE COFFEE ADVERTISEMENT AS A FUNCTION OF ITS POSITION AND OF INVOLVEMENT LEVEL (I.)

Procedure

Prior to the Principal Experiment two pilot experiments were carried out.

First Pilot Experiment. The object of this experiment was to decide the exposition time that would be included in the instructions. To this end, 10 slides were presented to a sample of 10 women, and they were asked to look at them as though they were leafing through a magazine. Mean viewing time for the 10 slides was 24.86 seconds, with a standard deviation of 6.26. Thus, it was decided to include in the instructions a presentation time of 25 seconds.

Second Pilot Experiment. The objective of this experiment was to confirm whether involvement was manipulated. Subjects were informed that their task would be to test a new brand of a product (perfume or coffee), give their opinion of it and, subsequently, whilst their eye movements were recorded, watch a series of 10 adverts on slides, among which would appear the advert of the product they tested. The object of the product test was to manipulate the involvement of the subjects with the tested products. The test (or not) of the product produced high (low) involvement with the advert. After the presentation, subjects selected the four adverts that had most interested them. Although the results obtained were as expected, we decided to introduce certain modifications in the procedure in order to increase the effect.

Principal Experiment. In order to increase the differences between the high and low involvement conditions, the following modifications were introduced:

1. For increasing involvement with the tested product, subjects in the coffee testing condition were regular consumers of this product, and subjects in the perfume testing condition had a favourite brand, which they used habitually.

2. For increasing the association between the product test and the subsequent viewing of the advert, after the testing and assessment of the product we included a question about any printed adverts for coffee (perfume) that they remembered. Also, before the product test, subjects were told that their task would be "to test a new brand of perfume (coffee) and then give an opinion on an advert".

The calibration procedure for the recording of eye movements was identical to that used in Experiment 1. Immediately after the presentation of the adverts subjects were given a booklet containing a series of questions in order to check the manipulation of involvement and measure recall and recognition of the advertisements. These questions, as well as the results derived from them, can be found in Garcfa and Ponsoda (1997a and 1997b)

Results

1. Coffee advertisement

We used a 2 x 2 ANOVA, in which the factors were involvement (high/low) and position of the illustration (in the upper or lower part of the advertisement). Involvement and position of the illustration interacted to produce their effects on the total duration of fixations on the image [F (1, 12)=4.6, p<0.05]. Shortest duration was found when involvement was low and the image occupied the lower half of the advertisement. On examining the number of fixations on the image a principal effect of position was observed. The image received more fixations when it occupied the upper half of the advertisement [F (1, 12)=6.3, p<0.03]. Note that when involvement is high, the differences in number and total duration total of fixations produced on the image as a function of its position are much smaller than when involvement is low. With regard to mean duration, we again found a significantly longer duration when involvement was high [F (1, 12)=5.28, p<0.04]. The results can be seen in Table 3.

We found also that, independently of involvement level, when the image occupied the upper half of the advertisement, the first fixtion was on it in 100% of subjects, whilst the first fixation was never on the image when it occupied the lower half.

2. Perfume advertisement

We used a 2 x 2 ANOVA, in which the factors were involvement and position of the perfume illustration (on right side or on left side of the advertisement) and the dependent variables number, total duration and mean duration of fixations on the illustration of the advertisement. The results are presented in Table 4.

No significant effect was found for number, nor for total duration of fixations, though as it can be observed: (a) when the image occupies the left position it receives more fixations, and (b) the differences between right position and left position are greater when involvement is low.

With respect to mean duration of the fixations, this is once again significantly superior in the high involvement condition [F (1, 12)=5.17, p<0.042].

We also found that, when the image was situated on the left side of the advertisement, first fixation was on it in 100% of subjects; however, first fixation was never on the image when it was situated on the right side.

TABLE 4

MEAN OF EYE FIXATIONS ON THE IMAGE OF THE PERFUME ADVERTISEMENT AS A FUNCTION OF ITS POSITION AND OF INVOLVEMENT LEVEL (I.)

Discussion

In the second experiment we succeeded in manipulating involvement using an easily-applied procedure of proven validity, that of the product test. The effects of involvement on attention were identical to those obtained in Experiment 1. Subjects in the high involvement condition looked at the advertisement for longer, and made more fixations and of longer duration, which indicates that they extracted more information from the advertisement and process that information more deeply.

The practical implications of the results of the two experiments are clear: (a) the need to allow subjects to have low involvement with the advertisements in tests of the effectiveness of advertising, in order to obtain more valid results, and b) the importance of carrying out further research to explore the possibility of interactions between involvement during effectiveness tests and different characteristics of the advertisements (e.g., knowledge of the brand, relationship between the image and the product, position of the illustration in the advertisement, etc.). It would be interesting to look at how involvement with an advertisement affects the distribution of attention, and thus the location of fixations. Antes and Kristjanson (1993) found that the reduction of cognitive resources available for viewing a scene had significant effects on the informative value of the areas on which fixations were made. On reduction of the resources available the number of fixations on objects of low informative value decreased. As shown by the results of the two experiments carried out here, when involvement with an advertisement is low the resources assigned to its processing are fewer, so that it could be expected that the number of fixations produced on objects of little informative value would be smaller when involvement was high.

ACKNOWLEDGMENTS

This research was supported by DGICYT (Project PS94-0040).

REFERENCES

Antes, J. R., & Kristjanson, A. F. (1993). Effects of capacity demands on picture viewing. Perception & Psychophysics, 56, 808-813

Bogart, L. & Tolley, B.S. (1988). The search for information in newspaper advertising. Journl of Advertising Research, 28, 9-19.

Celsi, R. L., & Olson, J. C. (1988). The role of involvement in attention and comprehension processes. Journal of Consumer Research, 15, 210-224.

Garcfa, C., Garcfa, L. & Ponsoda, V. (1993). Software para el registro y anßlisis de moviemientos oculares. Psicothema, 5(2), 413-418.

Garcfa, C. & Ponsoda, V. (1996). Las caracterfsticas de los anuncios y la attention: una revisi=n. Revista de Psicologfa Social Aplicada, 6(2), 5-25.

Garcfa, C. and Ponsoda, V. (1997a). Involvement in advert testing: The effects on attention. Proceedings of The XXII International Colloquium of Economic Psychology. Valencia, 2, 748-760.

Garcfa, C., and Ponsoda, V. (1997b). Las caracterfsticas de los anuncios y la atenci=n. Estudios sobre Consumo, 41, (85-100).

Guttman, J. (1972). Tachistoscopic test of oudoors ads. Journal of Advertising Research, 12, 21-27.

von Keitz, B. (1989). New consumer’s attention processes. Esomar Seminar on How to Increase the Efficiency of Marketing Communication in a Changing Europe. (pp. 125-142), Turin.

Kroeber-Riel, W. (1987). Effects of emotional pictorial elements in ads analyzed by means of eye movement monitoring. Advances in Consumer Research, 11, 591-596.

Kroeber-Riel, W. & Barton, B. (1980). Scanning ads effects of position and arousal potential of ad elements. Current Issues and Research in Advertising, 147-163.

Krugman, D. M.; Fox, R. J.; Fletcher, J.E.; Fletcher, P.M. & Rojas, T. H. (1994). Do adolescents attend to warnings in cigarette advertising? An eye tracking approach. Journal of Advertising Research, 34, 39-59.

Lohse, G. L. (1997). Consumer eye movement patterns on yellow pages advertising. Journal of Advertising, 26(1), 61-73.

Pratkanis, A. R. & Greenwald, A. G. (1993). Consumer involvement, message attention and the persistence of persuasive impact in a message dense environment. Psychology and Marketing, 10, 321-332.

Rosbergen, E.; Pieters, R. & Wedel, M. (1997) Visual attention to advertising: a segment level analysis. Journal of Consumer Research, 24, 305-314.

Thorson, E. & Page, Th. J. (1988). Effects of product involvement and emotional commercials on consumer’s recall and attitudes. En S. Hecker & D. W. Stewart (Eds.), Nonverbal Communication in Advertising (pp. 111-126). Lexington, MA: Lexington Books.

Tversky, B. (1974). Eye fixations in prediction of recognition and recall. Memory & Cognition, 2, 275-278.

Treistman, J. & Gregg, J.P. (1979) Visual, verbal and sales responses to print ads. Journal of Advertising Research, 19(4), 41-47.

Viviani, P. (1990). Eye movement in visual search: Cognitive, perceptual and motor control aspects. En E. Kowler (Ed.) Eye movements and their role in visual cognitive processes (pp.353-393). Elsevier Science Publishers BV.

Young, E. (1984). Visibility achieved by outdoor advertising. Journal of Advertising Research, 24, 19-21.

Zaichkowsky, J. L. (1985a). Measuring the involvement construct. Journal of Consumer Research, 12, 341-352.

----------------------------------------