Market research between worlds

Globalization implies that marketing and market research activities quite naturally take place at a global level. It often happens that a parent company in one country contracts and manages its research activities from a central location, usually with the desire of having comparable data collected in the various markets.

Market research departments in companies and institutions are thus presented with a tremendous challenge, both in terms of the amount of work, and ultimately the necessary investment, as well as the required methodological expertise.

In multinational studies, researchers are generally confronted with a wide range of tasks. This article takes a closer look at two potential stumbling blocks. What first appears to be a simple task of drawing up questions and a response scale is clouded by the difficulty of reaching the identical connotation when translating the questions to other markets. This can be illustrated in a simple example: Does the question in German “Sind Sie glücklich?” hit the same chord as “Are you happy?” when asked in the United States? And would the answer “sehr glücklich” be identical to an American saying “very happy”?

Researchers refer to this as scale equivalence. We need to ensure that translations are not only correct in terms of the wording but also that they semantically and functionally imply the same thing as to ensure they conjure up the same ideas in the participants’ minds, allowing them to be answered in the same way. A literal translation is often not the most appropriate solution, because it does not adequately take the cultural context into account.

A different challenge arises with the responses. Regardless of the actual content of a question and the exact wording of the choices of answers, there are great differences in how people respond or, more specifically, how they use a response scale. In Asia, for example in China or India, we see a strong tendency to give consent (a “yes tendency”). Japanese participants, however, seem to avoid extremes and show a pronounced tendency towards responding in the middle of the scale. Americans give very pointed responses with clear approval or rejection. And Europeans are somewhere in between. Given these patterns of response, researchers could quite accurately determine in which culture a survey might have been conducted.

Admittedly, this is not a newly discovered phenomenon. Correctly designing a study, adequately evaluating the collected data and retroactively removing the cultural effects from the result is anything but trivial.

Strategies for multinational studies

How then does one face these challenges appropriately? The first strategy looks at the phase of drawing up the questionnaire. Usually, the best starting point is looking at the language in which the survey instruments have been developed. In close consultation with the client, the questions and responses selected are then translated. As translators are usually not involved in determining the content of the survey, an untrained translator may initially just offer a literal translation of the material. It would be better to gather a multinational team of translators at an early stage when the survey questions are being determined, and this team could play a role in developing the instruments instead of translating them post hoc, as to capture the aforementioned “core meaning” in the questions as much as possible. Though this may appear to be an impossible and overly expense approach in a realistic setting, it certainly could be useful with regard to the validity of the results. The second strategy is geared towards the evaluation phase. A statistical assessment can be carried out to determine whether the response scales apply across more than one country. Other methods even allow for the collected data to be adjusted as to reduce or even eliminate cultural specific response behavior. At this point, simple arithmetic is not enough; instead fairly complex statistical models (for example, Bayesian models) and wide-ranging background knowledge are required.

The third strategy focuses on interpreting the results. Often, participants from country to country are not of the same structure; for example, Chinese participants on average are younger than in Germany, particularly in online studies. This does not necessary entail a mistake in methodology, but it can certainly be reflected in how representative groups of buyers are. If, however, additional consumer characteristics coincide with the differing age groups, for example, age is often confounded with an affinity for technology, it can be difficult to determine whether observed differences between the two countries really exist or are merely attributable to precisely these different structures among the study groups. This is somewhat like the chicken or the egg dilemma, which offers no simple answer; instead a clear definition of the respective target group in each country is required.

Measuring the client’s attitudes and preferences is anything but trivial, and this applies particularly to multinational studies. Future generations of market researchers will have to take this issue even more into account and have the corresponding methodological expertise in their toolboxes.


(Image source:

More publications