Pictorial Image Measurement - Applied to Countries

PICTORIAL IMAGE MEASUREMENT - APPLIED TO COUNTRIES

Guenter C. Schweiger, Wirtschaftsuniversitat Wien

[The author would like to acknowledge the helpful and competent support provided by the following researchers: Z.S.Demirdjian, California State University; Gordon E. Miracle, Michigan State University; Jerry C. Olson, Pennsylvania State University; Jagdish N. Sheth, University of Southern California; Brown Whittington Jr., Emory University Atlanta; Helmut Kurz, Vienna School of Business; Wolfgang Wiklicky, Vienna School of Business.]

ABSTRACT -

This paper examines the suitability of picture stimuli for image measurement. For that purpose, the image of Austria in the USA was investigated in comparison to Western Germany and Switzerland by using visual stimuli along with the usual verbal ones. Simultaneously, the pictures used in the main study were evaluated by a separate sample of respondents. Evidence was found that picture stimuli have a higher power of discrimination between the countries.

INTRODUCTION

There is no doubt that advertising relies heavily on verbal as well as visual information. Considering advertising effectiveness, it can be stated that pictures exceed textual information in importance. There are a great number of reasons for this:

- A human being's capacity to perceive and store visual information is nearly unlimited and by far exceeds human capacity to perceive and store cognitive information.

- Usually, pictures catch the eye before words do.

- In general, the "activation potential" of pictures is higher than that of written information.

- Imagery researchers have shown that pictures rank first whilst words rank second in terms of recall.

- Picture stimuli, as well as word stimuli, are appropriate to build up positive images of products.

- Attitudes and images can be built up and changed without written information.

When accepting the assumption that the processing of information is based on pictures and visual information, a question arises: are the purely verbal methods of image and attitude research (multiattribute models derived from social psychology) appropriate for researching images?

In this paper the integration of non-verbal stimuli is postulated. The author proposes Visual and acoustic stimuli, the former being more important.

PRAGMATIC ASPECTS OF VERBAL VERSUS PICTORIAL IMAGE MEASUREMENT

In addition to the fundamental advantages of image measurement with use of visual stimuli, further aspects have to be taken into consideration.

The encoding-decoding problem

In general, briefings sent to an advertising agency are verbal descriptions. These briefings have to be encoded by the "creative" of the receiving agency. Sucess or failure of a message depends on that encoding process. This is clearly one of the weakest points in the message conceptualization process. Figure 1 shows the typical planning and execution process of communication:

FIGURE 1

THE ENCODING-DECODING PROBLEM IN THE PLANNING AND EXECUTION PROCESS OF COMMUNICATION

The market researcher, too, faces an encoding-decoding problem. When using the traditional instruments of attitude and image research, he must try to verbally describe the visual appearance of a brand. In this case the encoding-decoding problem arises: the market researcher has to encode visual stimuli into verbal stimuli to form the basis for his research. Then, the results he obtains (e.g. semantic differential) supply the foundation for the briefing sent to the advertising agency. The creative then has to translate these verbal stimuli into Visual stimuli, i.e. pictures.

Of course, this encoding-decoding process involves problems and makes the execution of advertising campaigns more difficult. Hence a question arises: is this encoding-decoding process really necessary? This potential source of problems could be eliminated by considering the visual aspect in image research.

Images as entire visions ("Ganzheit")

Images are entire visions. In the case of verbal image research, the investigation of images is done by presenting separate, disunited verbal stimuli. To be able to deduce images, the reactions to the separate word stimuli have to be combined afterwards. But first splitting up and then combining the reactions contradicts the entirety of images (German Word "Ganzheit")!

International research projects and campaigns

Translation problems arise in the execution of international research projects and campaigns. This includes the problem of inter-cultural misunderstanding. By using pictorial stimuli, translation problems could be reduced.

The object of this paper is to examine the suitability of picture stimuli for image measurement. For this purpose, an empirical study was conducted. The study investigated the image of Austria in comparison to Western Germany and Switzerland by using not only verbal statements, but also pictures.

SELECTION OF THE PICTURES

Around 6.000 pictures (provided by the Austrian Tourist Board) served as a basis for the pre-selection of visual stimuli. Additionally, some pictures with negative contents (e.g. "polluted") were used. From this basis, a set of 70 pictures was Put together. This set was numerically reduced by group discussions with American respondents, eliminating pictures typical for other countries (e.g. France).

The next step was to check, which of these pictures communicated specific attributes (e.g. "friendly", "pleasurable", "romantic", "cultural", "beautiful landscape", "Polluted"). 21 respondents were asked to assign these attributes to the remaining 50 pictures. Only those pictures, to which at least two thirds of the respondents assigned one or more of the attributes, constituted the final list. Thus, a set of 28 pictures emerged.

With the help of these pictures, two studies were executed simultaneously:

1) Investigation of the image of Austria in comparison to Western Germany and Switzerland in the view of 292 American respondents ("main study").

2) Evaluation of the pictures by an American sample of 31 respondents.

FIGURE 2

DESIGN OF THE STUDY

METHOD

Main study

In the main study the respondents were asked to spontaneously assign both pictures and word stimuli to the three countries. They were allowed to mention none, or one, or more countries. The data were obtained from a standardized questionnaire, which additionally contained the following facts:

- spontaneous knowledge of all Europeancountries

- European countries actually visited; intention to visit Europe

- knowledge of Austria (e.g. famous Austrians, famous cities, states and regions ... )

The word stimuli included a set of 12 attributes which was also used in the evaluation study (see below).

Evaluation of the pictures

In a separate study, the 28 pictures were evaluated. This evaluation study was based on 31 interviews in the United States of America.

The pictures were judged on the basis of a 4-point rating scale ("applies very well"; "applies well"; "applies a little"; "does not apply at all") with reference to twelve attributes (e.g. "friendly", "pleasurable", "depressing", "romantic", "exciting", "boring").

HYPOTHESIS AND EMPIRICAL TEST

The basic hypothesis is: "Picture stimuli (visual stimuli) are more suitable for measuring images - in this study the image of countries -than word stimuli. Visual stimuli discriminate better between objects (e.g. countries)."

For the operationalization of this hypothesis different methods can be applied. The basis is formed by 28 picture stimuli and 44 word stimuli. In the main study both pictures and words had to be assigned to the three countries (nominal level of measurement).

The hypothesis now was operationalized and tested on the basis of both nominal and ordinal dasta.

First operationalization of the hypothesis (nominal data)

The assignment of picture stimuli was compared to the assignment of corresponding word stimuli by applying the Wilcoxon test. This test checks on an individual level the extent of multiple assignments of corresponding picture and word stimuli to the three objects under evaluation (Austria, Western Germany, Switzerland).

For this purpose, it was necessary to formulate a corresping null hypothesis which says that "The extent of multiple assignments to objects is the same in the case of both picture and word stimuli".

The Wilcoxon-Test-procedure tests the null hypothesis, that the sum of the multiple assignments of a picture to the three countries does not differ from the sum of the multiple assignment of a corresponding word stimulus. The following example illustrates this procedure.

For instance picture # 17, which shows a river and a high mountain is compared with the word stimulus "beautiful landscape". The sum of multiple assignments of this picture and word stimulus was as follows:

TABLE

Under Ho, the + and - differences between the two sums should be zero. This hypothesis was rejected at the 0.0001 level of significance in the Wilcoxon-Test: There were significantly less multiple assignments of picture # 17 to the three countries than multiple assignments of the word stimulis "beautiful landscape".

This procedure was repeated for another 42 comparisons of corresponding pictures and word stimuli. 28 times the results were the same as in the example shown above (significantly less multiple assignments of pictures than words), 10 times the H could not be rejected, and in 4 comparisons ~he pictures provoked significantly more multiple assignments than their word counterparts.

The results of the 4 comparisons showing more multiple assignments of pictures than comparable word stimuli demonstrate a further advantage of the pictorial image measurement procedure. The following comparison shows the percentages of assignments of the word Stimulus "polluted" and pictures, which were evaluated as being polluted to the three countries (see table 1).

TABLE 1

RELATIVE FREQUENCIES OF ASSIGNMENTS OF POLLUTED STIMULI

As can be seen from this desired answers seem to be cases of picture assignment assignment.

Second operationalization  of the hypothesis (ordinal data)

A second hypothesis with regard to discrimination power was deduced in a further step. The operationalization refers to the question whether direct assignment of emotional attributes (with use of a rating scale) or indirect assignment (with use of pictures) has more power to discriminate between objects (countries).

The results of direct assignment are shown in figure 3, (see below); differences between Switzerland and Austria are only slight. In the case of the items "friendly", "romantic" and "exciting", differences between these two countries are not significant (Wilcoxon test).

Indirect assignment of the attributes was effected by finding out which of the pictures were assigned significantly most often to the countries. Table 2 shows that 5 pictures were assigned significantly most often to Austria. 11 pictures were assigned significantly most often to Western Germany and 6 pictures were assigned significantly most often to Switzerland (CL!!~ 0.05). 22 of the 28 pictures were assigned significantly more often to one of the countries.

TABLE 2

To minimize the time for the interview, a split run was effected. 15 pictures were used in split run A, 14 in split run B, one picture (No. 18/23 Girl drinks wine) being included in both samples to check possible interaction effects. NO such effects were found.

With the help of the pictures significantly assigned to one of the countries, it is now possible to compute a mean for each country and each attribute used in the evaluation study. This calculation process is illustrated below for "Austria" and the attribute "pleasurable".

TABLE

This calculated mean (2.13) can be found in figure 3. The means for all other attributes and countries were computed analogously.

Comparing both charts in figure 3, the following conclusions can be drawn:

1) When assigning the attributes "pleasurable", "exciting" and "depressing" both directly and indirectly to the countries, Switzerland ranks first, Austria ranks second and Western Germany ranks third.

2) When assigning the remaining three attributes "friendly", "romantic" and "boring" directly, Austria and Switzerland were ranked similarly, both ahead of Western Germany. Yet, when assigning these three attributes indirectly, evaluation differences between Austria and Switzerland turned out, to be much larger. For all items, Switzerland got more positive values than Austria.

Test of the hypothesis

For the testing of the hypothesis it is important to know whether word stimuli or rather picture stimuli have a higher power of discrimination between objects. For this purpose, we calculated the difference between the rating value of the country, to which the most positive value had been assigned, and, on the other hand, the country, to which the most negative value had been assigned. This was done for all six attributes.

The next step was to compute the mean over all these differences. This mean score was calculated for both direct image measurement (using word stimuli) and indirect image measurement (using picture stimuli). A comparison of both means indicates that picture stimuli discriminate considerably better between objects than do word stimuli - a result which was supported by computing a t-test (% = 0.005).

TABLE 3

DIFFERENCES OF RATING VALUES BETWEEN THE BEST-RATED AND THE WORST-RATED COUNTRY

How to read the table: 0.46 shows the difference between the rating value of the "most pleasurable country" (Switzerland, 1.53) and the "least pleasurable country" (West Germany, 1.98) in the case of direct image measurement.

CONFORMITY OF THE RESULTS

Conformity of both methods of measurement could be found not only for the six items mentioned above, but also for the following items which were used in both the main study and the picture evaluation:

- beautiful landscape

- old fashioned

- polluted

- modern

CONCLUSION AND OUTLOOK

With respect to the underlying data, the tests provided evidence in favor of visual stimuli. Visual stimuli discriminate better between objects than do verbal stimuli. Although all pictures Used in the study also communicate cognitive contents (in addition to their emotional contents), the indirect method of assigning emotional attributes to countries is absolutely consistent with direct assignment Of such attributes.

Despite the fact that both studies are based on relatively small samples (main study 292 respondents, evaluation study 31 respondents), the high consistency between both methods gives evidence for the high reliability of non-verbal image measurement.

Besides, some further advantages of non-verbal image measurement with use of picture stimuli should be mentioned:

-Results are more comprehensible and more easily interpreted.

-For the "creatives" in the advertising agencies, visual briefings become possible.

- In the case of multi-national studies, problems can be reduced.

- Finally, and not be overlooked, the questionnaire is more interesting and respondents do not get as tired as they do answering verbal item pools.

Finally, a second modus- of non-verbal image measurement deserves at least to be mentioned: the use of acoustic stimuli. For the first time, image measurement with the help of acoustic stimuli was effected in a pioneer study very similar to the one described above, which was conducted in Mexico.

A set of 15 pieces of Austrian music provided the basis for the acoustic stimuli, e.g. the "Harry Lime" - theme, "The Blue Danube" (Donauwalzer), "Silent Night, Holy Night" (Stille Nacht), the "Radetzky Marsch" etc.

Those acoustic stimuli that were assigned significantly more often to one country than to the others are shown in table 4.

TABLE 4

SIGNIFICANT ASSIGNMENTS OF ACOUSTIC STIMULI TO THE COUNTRIES (% < 0.05)

FIGURE 3

RESULTS OF DIRECT VERSUS INDIRECT IMAGE MEASUREMENT

REFERENCES

Childers, Terry L. , Houston, Michael J., Conditions for a Picture Supriority Effect on Consumer Memory, Journal of Consumer Research, Vol. 11, 1984

Haley, R.I. 9 Richardson, J.~ and Baldwin, B.M., The Effects of Nonverbal Communications in Television Advertising, Journal of Advertising Research, Vol. 24, 1984

Hansen, Flemming, Towards an alternative Theory of the Advertising Communication Process?, Research in Marketing, Vol. 1, Nr. 1

Hofstaitter, Peter R., Sozialpsychologie, Berlin, New York, 1973

Kisielius, J., Roedder, D.L., The Effects of Imagery on Attitude-Behavior Consistency, in: Bagozzi, Richard P., Tybout, Alice M. (Ed.), Advances in Consumer Research, Vol. X, 1983

Kisielius, J., Sternthal, B., Detecting and Explaining Vividness Effects in Attitudinal Judgements, Journal of Marketing Research, 21, 1984

Kroeber-Riel, W., Konsumentenverhalten, 2. Auflage, 1980

Marks, David F., Visual Imagery Differences in the Recall of Pictures, in: British Journal of Psychology, 64, 1973

Paivio, Allan, Imagery and Verbal Processes, Hillsdale, New Jersey, 1973

Percy, L., Rossiter, J.R., Advertising Strategy - A Communication Theory Approach, New York, 1980

Rossiter, J.R., Percy, L., Visual Imaging Ability as a Mediator of Advertising Response, Advances in Consumer Research, Vol. 5, edited by H.K. Hunt, Ann Arbor, Michigan, Association for Consumer Research

Sheikh, Anees A., Imagery - Current Theory, Research, and Applic n, Milwaukee, Wisconsin, 1983

Sherman, Kulhavy et al., Cerebral Laterity and Verbal Processes, Journal of Experimental Psychology: Human Learning and Memory, 1976