Editor's note: Felipe Korzenny, Ph.D., Betty Ann Korzenny, Ph.D., and Sandra M.J. Wong, Ph.D., are all senior research staffers at Hispanic & Asian Marketing Communication Research Inc., Belmont, Calif.
Imagery investigation - uncovering the images consumers have of products - has become an important aspect of marketing research. It's a technique that seems to work well with U.S. consumers, most of whom speak English and share social and cultural reference points. Studying product imagery across cultures, on the other hand, is complicated.
There is an unfortunate tendency to oversimplify imagery research when it is conducted with consumers who don't speak English. Simply translating a general market instrument from English into another language may not make the instrument ready to use with consumers in another culture. In addition, it is unwise to compare the data collected from an English instrument with that from a translated instrument.
For example, a product's image of health or power may have different meanings across cultures. One culture may view a car as powerful because it is associated with powerful people. Another culture may perceive it as powerful because of the engine's horsepower rating. Both cultures rate both cars high on power but for critically different reasons.
Etic vs. Emic approaches
Cross-cultural marketing research often involves a difficult, but important, choice between sacrificing cultural understanding and sacrificing comparability.
Etic1 research strives for comparability across diverse cultures, seeking data about specific items - data that can be contrasted across cultures. Conducting etic research involves translating and adapting one culture's terminology and structures so they are understandable and useful in another culture.
For example, etic research allows the comparison of Hispanics' perceptions of artificiality with the general public's perceptions of same, or perceptions of corporate character across sub-populations with various cultural backgrounds.
While there have been many attempts to produce cross-cultural data from an etic research perspective, the data sets such comparisons are based on are rarely equivalent.
A common assumption in the etic approach is that aside from language differences, cultures are similar and equivalent. People seem to feel that if we all spoke the same language, we would all be alike.
But the simplicity sought in the etic approach is elusive. Mostly that's because cultures and languages are responses to environment. Cultures consist of adaptive tool sets that humans create in the interest of survival. These tool sets are particular to a group, and are passed on and maintained over time as traditions.
There are three universal dimensions of meaning: power, evaluation and activity. Nevertheless, even the ways in which these universal dimensions are used vary across cultures. For example, in the car research previously mentioned, power was important across cultures, but it was perceived as having a technological association in one culture and an interpersonal association in another. Wealth can be commonly sought across cultures, but one culture may measure wealth in terms of the number of children one has, and in terms of gold in another. Happiness can be seen as a common pursuit, but may have very different end goals.
Language and culture are intrinsically related
Language is an intrinsic part of culture. It is a tool and a means toward important ends. Language is not just a way to speak; it shapes the way we process information.
Cultures create words to reflect new experiences so those experiences can be shared. Since different cultures have different experiences, different languages are not just different sets of words, they are reflections of specific experiences.
If language influences how we think, then language also affects the imagery we associate with different products. Our language and culture affect our perceptions and abilities to generate metaphors.
Imagery and language in contrast
General market or mainstream U.S. meanings contrast with the meanings understood by many of the country's non-Western cultures, including Hispanics and several Asian groups. Figure 1 contains some examples of cultural differences in imagery gathered while doing research. The image attributes listed are instrument items that can be rated on a Likert-type scale ranging from "strongly agree" to "strongly disagree" for a particular type of product, e.g., a pizza brand or a soft drink.
The reader should keep in mind a product or service close to his or her heart while looking at these examples, and ask the following question about it: How would the disparities of interpretation shown affect research conclusions?
These examples suggest how an etic approach to cross-cultural research may backfire. It becomes evident: making comparisons while working across cultures can be complicated and results obtained are often misleading. This is particularly true when words and their translations are taken literally.
Some readers may argue that it is naive to deal with concepts the way we have. Some people believe that it is possible to adapt a concept in a way that allows for culturally relevant conclusions. But attempting to do so consistently gives rise to a specific problem - the so-called the paradox of equivalence - that is not easily resolved.
The paradox of equivalence
When conducting multicultural research, one can nobly attempt to make various translations of a concept as equivalent as possible. For example, a researcher could culturally adapt and translate the connotation of "tradition" in English as "the glory of your ancestors" so it made sense to a certain cultural group.
The more culturally equivalent a concept is, the more relevant the translations will be to their respective cultures. But the translations will likely fail to uncover cultural differences precisely because they were well adapted to each culture. What a headache!
On the other hand, a more literal and unadapted translation will be less relevant to a second culture. This approach will tend to uncover cultural differences, but these cultural differences may be complicated and have implications that render the translation worthless.
For example, a product may be judged to be quite natural by the U.S. mainstream, while Hispanics may feel it is not natural. But mainstream U.S. consumers may think the product is natural because it does not have preservatives, while Hispanics may feel it is not natural because it is not fresh from the farm.
Going native: the emic approach
Multicultural research can also be undertaken from an emic perspective, wherein concepts are examined from the perspective of the culture being studied. This approach can lead to a deeper understanding of the culture, but may preclude comparisons between cultures.
Still the emic perspective can be quite useful. After all, cross-cultural comparisons can be used to generate global approaches to selling products or services, but they may not do much to ensure the success of a specific product.
To best position a product or service, one should be most interested in knowing how a particular culture perceives the product or service. For example, brands that have positioned themselves uniquely within the Hispanic or Asian markets have claimed more loyalty and market share than others. To assess imagery correctly across cultures, researchers must take emic approach.
It's best to start by conducting qualitative research within each of the target cultures to gain an understanding of the particular meanings associated with a brand or product. The specific meanings uncovered at the qualitative stage can then be used in a second quantitative step.
In the quantitative phase, the attributes uncovered are checked for prevalence and importance with a representative sample. The results of the quantitative phase must be tempered by the meanings discovered at the qualitative phase.
Neither phase would produce results that are necessarily equivalent across cultural groups. Qualitative procedures may differ, and the quantitative imagery attributes can differ even more. Different instrument scales and structures would have to be used to accommodate cultural differences.
Imagery in qualitative research across cultures
Product sorts and personifications are techniques routinely used in the general market for understanding product imagery, but these approaches may not make sense to some cultural groups.
Sometimes it's not just cultural differences that restrict consumers' ability to participate in such qualitative approaches. Many times it is a lack of knowledge of products in a category that precludes the generation of imagery.
Think of yourself - assuming you are not completely familiar with the Japanese culture and its products - in a focus group in Japan. Imagine a set of products placed in front of you with Japanese labels. How would you sort these products?
Logically enough, consumers from other cultures tend to classify products by size, color, shape and familiarity. But researchers often expect consumers from other cultures to have an image of all products normally found on the shelves of supermarkets. To the surprise of many marketers, products that have not been directly advertised to culturally different consumers have almost no imagery among those consumers.
When asked to sort products, consumers typically cluster familiar products together simply because they know and trust those products. Likewise, they will group other products together "because we don't know them. They are not advertised enough in our language."
Personification exercises and other projective approaches that ask respondents to attach attributes normally associated with people, animals or objects to products tend to flop if not planned with the culture in mind. A researcher might ask a question such as, What type of animal does each of these cans of beans remind you of? And respondents could respond with total silence.
In other categories, such as deodorants or dairy products, respondents may only know one or two brands. They would likely say they like the smell or taste of their favorite brand, and would not know how else to differentiate the products.
In such a situation, a qualitative researcher who is well anchored in his or her culture can get consumers in the same culture to discuss attributes, perceptions and associations relevant to the culture. Avoid preconceptions about approaches that may have worked in the general U.S. market. Such approaches can perplex respondents and fail to elicit important findings.
Quantitative research variants
Once the researcher has successfully identified themes and attributes in qualitative research, the project can proceed to a usage and attitude study. There are several issues to consider when striving to make the research approach culturally sensitive.
- Allow for ignorance. In imagery research, typical quantitative approaches to asking questions do not allow the respondent to say "I don't know." The researcher usually says, "Please tell me how much you agree or disagree with each of the following statements. Please tell me if you strongly agree, agree, disagree or strongly disagree, or if you neither agree nor disagree."
Soda X is natural.
Soda X is good for you, etc.
The respondent has not been given the option to say that he or she does not know if the soda is natural or good for you. A "neither agree nor disagree" answer does not indicate that the respondent doesn't know.
Semantically, "neither agree nor disagree" could indicate indifference or simply the belief that the product has an intermediate rating on an attribute. The respondent should be given the option to say that she or he does not know if the attribute applies to the product.
- Make scales culturally compatible. Some cultures find scales easier to use than others. Mexicans, for example, are quite familiar with the notion that children in school are graded on a 0 to 10 scale. We have found out that 0-to-10 scales are easier for Mexican respondents to understand and use because of this familiarity.
- Pre-coded open-ended questions. The use of a long battery of items to detect perceptions and opinions may not be as useful in some cultures as they are in the general market. When a long list of attributes is read to respondents who share a specific cultural background, many respondents are likely to give the same kind of answer.
Hispanic respondents, for example, are known for using extremes of scales. After a few attributes are rated, they tend to settle on a high or low value and continue that way. In contrast, Asian respondents tend to avoid the extremes. They tend to settle on the more moderate value or to seek a yes/no or agree/disagree type of response.
When researchers need to present a battery of items, it best to break the items into small subsets to avoid getting the same answer to each question from respondents.
A better approach is to have pre-coded categories of responses. The attributes discovered in the qualitative phase are listed and pre-coded, but the respondent is asked to freely talk about a brand or product. Other response options are allowed.
Each mention that matches a pre-coded category is coded. This approach is more likely to validly reflect the way in which respondents look at the brand or product. But again, the results are not easily comparable across cultures.
Conclusion
An emic approach to research across cultures is recommended when product and brand imagery are important. Sensitivity in both qualitative and quantitative phases of research is fundamental to identifying valid cultural trends.
Looking at a culture from the inside is more likely to produce information that can be used for effective positioning. Imagery is strongly tied to culture, and language and culture cannot be separated.
Notes
1 From phonetics or vocal utterances that are universal, as opposed to phonemics or vocal utterances that are specific to a culture. A discussion of this topic can be found in Kenneth Pike's "Etic and Emic Standpoints for the Description of Behavior" in the book Communication and Culture, edited by Alfred Smith, published by Holt, Rinehart, and Winston, 1966.
2 ibid.