Editor's note: John McKay is senior vice president at Data & Management Counsel, Inc., a Chadds Ford, Pa., research firm.

The research process is key to understanding Asian consumers. The tools for multicultural research are the same as those used in mainstream market research: focus groups, telephone surveys, one-on-one interviews, etc. However, adjustments are sometimes necessary in designing research for Asian-American markets. Research assumptions and techniques that work in researching the general public are not always applicable in the Asian-American market.

The best approach is to stick to the core strategy of understanding consumer needs, attitudes and behaviors, identifying the cultural hot buttons, and using the established research techniques that have proven successful in mainstream marketing. The key is in the research process: using good, fundamental research and intelligent, culturally sensitive people with insights into Asian culture to help interpret the findings.

This article will focus on several key aspects of designing market research for the Asian-American market, including:

  • the impact of Asian culture on conducting market research;
  • conducting quantitative research among Asian Americans;
  • sampling issues in targeting Asian Americans.

The impact of Asian culture on conducting market research

Asian cultures can have a significant impact on how marketing research is conducted within the Asian-American community in at least three key areas:

1) selecting appropriate research sample/respondents;

2) designing questions to effectively gather meaningful information; and

3) establishing a productive relationship between researcher and respondent.

U.S. researchers need to consider cultural subtleties or complexities in the Asian-American market that are not always evident in research conducted among the general market. To avoid mistakes that are commonly made by researchers unfamiliar with Asian cultures, it may be helpful to have a culturally sensitive researcher involved in the research planning stages.

1. Respondent selection
Determining the most appropriate respondent for an Asian-American research project requires an awareness of the various roles, responsibilities and authority levels in an Asian-American family. A researcher needs to have an understanding of and sensitivity to the Asian family structure in order to design the proper framework for finding appropriate respondents. Without this knowledge, researchers risk gathering information from the "wrong" respondents and developing conclusions based on invalid data.

It is therefore important for researchers to understand how Asian family relationships add complexity to the selection of research respondents. Most Asian-American households tend to be multi-generational as well as multi-family. There is an intricate web of Asian family ties which makes selecting appropriate research respondents especially challenging for the researcher.

In the typical Asian-American family structure, the designation "head of household" does not necessarily mean "decision maker." That designation clearly depends on the type of decision to be made and/or the topic under investigation. The context in which decisions are made for the Asian household influences whether the researcher should treat the occupants of a multi-family household as one extended family or as separate families.

2. Question design
Asian culture also affects the development of the questions used to gather information from Asian respondents that will be useful and meaningful to the researcher. For example, where a respondent might be asked to give a Very Good or Excellent rating in a survey designed for the general market, research shows that native or first-generation Asian respondents do not verbally understand such subtle distinctions that are more apparent in English, and are therefore more responsive to numeric scales.

Asians prefer to be noncommittal when asked pointed questions - and this lack of definitiveness can skew research. They frequently tell the interviewer what they believe he wants to hear rather than what they really think. When asking Asian respondents to answer a survey question using a rating scale, an even-numbered scale (four points, six points, 10 points) will encourage dichotomy and discourage their tendency to give noncommittal responses.

In some Asian cultures, people find it difficult or uncomfortable to demean a product or a company. If Asian respondents are presented with a top-box question on customer satisfaction using a scale ranging from Very Satisfied to Very Dissatisfied, Asian respondents will seldom give a response lower than Satisfied. It is therefore necessary to take a different approach with Asian respondents, even though the process is the same. This is where it is important to use the expertise of an Asian cultural expert.

If the researcher fails to provide a context for the question that Asian Americans recognize in terms of their own cultural reality, Asian-American respondents by either:

  • redirecting the question,

  • giving a noncommittal response, or

  • creating a context in order to respond.

The danger in this process is that respondents may provide information that reflects issues other than those under investigation. To avoid such cultural pitfalls, it is important to pre-test a survey prior to full-scale implementation to help the researcher understand word/concept interpretations and contextual clues.

3. Researcher-respondent relationship
The researcher-respondent relationship consists of the researcher's own assumptions and meanings; the respondent's perception of the researcher's expectations; and the rapport built between the two people. While this applies to all research, it is especially true in Asian marketing research. Establishing a good rapport with Asian-American respondents is crucial to successful Asian market research. The researcher needs to demonstrate a sincere desire to understand the respondents' values and opinions, so the respondents can teach the researcher about important dimensions of their world.

If it is obvious to an Asian person that the researcher cannot relate to or is critical of the respondent's world view, the respondent often will answer briefly with a response most likely to terminate the questioning. An Asian respondent who senses that the researcher is sincerely open to learning about the complexities and nuances of his or her world is more likely to provide a more complete, valid and useful response to a question.

Quantitative research

There are four key aspects of designing quantitative market research studies in the Asian-American market:

  • project timelines;
  • Asian language versus English-language interviewing;
  • translation;
  • survey design.

1. Project timelines
First and foremost, dealing with multiple Asian cultures in a study takes more time than a general market study. Clients should know this so that they can adjust their expectations and marketing accordingly. A quantitative study with several ethnic groups and languages usually requires multiple pilot tests to examine length, flow, and comprehension. In a general market English-language study, the researcher typically is dealing with one questionnaire, one set of crosstabs, and one analysis. In a study among 800 Asian consumers, for example, the researcher might be dealing with 200 interviews in each of four (or more) languages. This can mean almost four times the analytical work, along with managing multiple questionnaires, multiple sets of crosstabs, and multiple teams of in-language interviewers, translators, and supervisors.

2. Asian-language versus English-language interviewing
Telephone interviews conducted with Asian-Americans should generally be conducted in the native language of the respondent, because, unless noted otherwise, it is their preference. However, using bilingual (English/Asian) interviewers is recommended, as many second-generation Asians primarily speak English. For best results, the respondent should be given the choice of being interviewed in either English or the respective Asian language. This approach will provide the most representative data for the targeted segment. The researcher will then be able to track any differences in responses from foreign-born versus U.S.-born Asians, and English-dominant versus Asian language-dominant Asians.

3. Translation issues
Translation of the survey document is a vital factor in the success of the study execution and data analysis. It is therefore critical to give plenty of time and attention to this stage of the research, so that the effectiveness of the translation does not become a victim of project deadlines. Too many researchers are in such a hurry to rush a project into the field that they do not allow enough time for this important stage in the project schedule.

The more people involved in the translation, the better. The translator and the researcher must review the survey question-by-question so the translator understands the concepts and the intent of the questions, not just the words. People involved in the translation must be close to the market and understand how the concepts are articulated in the marketplace. One person who understands the purpose and meaning of the English survey should be responsible for guiding all language translations to ensure consistency.

It is very important to pre-test the in-language versions of the survey with actual respondents to make sure the wording is easy and comfortable for the interviewer to articulate, and the concepts and intent of each question are phrased in a way that is easily understood by and meaningful to the respondents.

Some multicultural researchers insist that a back translation be performed on the survey by a different translator once it has been translated into the Asian language. Their reasoning is that the client will be able to use the English back translation to determine whether the Asian language translation has been done correctly. The fallacy in this logic is that translation is not an exact science. No two translators will always use the same word or phrase to interpret the same concept. The Asian translation may be perfectly accurate, but the translator performing the back translation into English may not understand the context of the questions or may interpret the Asian concepts slightly differently and thus choose a different English word or phrase than what was used in the original English version. In addition, the time and money spent in performing the back translation is not always worth the effort and delay in fielding the study. Rather than waste time and money with back translations, it is more valuable to follow the two steps outlined below, which have already been discussed:

  • have the researcher or client's Asian staff review the translation;
  • pre-test the translated survey with actual respondents.

4. Survey design issues
Researchers need to be aware that most new Asian immigrants are not accustomed to telephone research. In their native countries, the predominant method of data collection is in-person interviews, as the penetration of telephones in consumer households is far lower than in the U.S. The in-person method lets the interviewer develop a level of rapport and personal trust with the respondent which is not as easily accomplished over the telephone. Asian consumers are not accustomed to being asked to give their individual, honest opinions, and this lack of familiarity with research methods can impact data collection.

Because of the nature of the political regimes in many Asian countries, a telephone call from an unfamiliar voice (even an Asian one) asking for opinions may often be met with a high level of suspicion. It is therefore very important that the wording of the initial contact with the respondent be deferential and polite while briefly explaining the purpose of the call.

It is important to note that Asian-Americans typically have a lower survey refusal rate than the general public -- about half of the national average of 45 percent. This lower rate of survey refusals demonstrates an eagerness on the part of Asian-Americans to have their opinions heard. It may also be a reflection of the politeness of the Asian cultures and a reluctance to say no. Researchers should be careful not to take advantage of these factors by keeping the respondent on the phone for 30 minutes or longer. A maximum of 20 minutes is recommended for Asian telephone surveys, with a 10 to 15 minutes preferred.

For cost and timing estimation purposes, it is important to know that in-language interviewing typically takes about 20 percent longer to conduct than the same interview in English. If an English language version of the survey takes an average of 15 minutes to conduct, then the Asian-language version of the same survey will probably take an average of about 18 minutes to conduct. This is because it takes longer to describe the same concept in Asian languages than it does in English. The degree of formality and politeness required in the phrases used by the interviewer when speaking with Asian respondents also requires extra time.

Sampling issues

By merely using a random digit dialing approach, it is very difficult and very costly to find the small segment of the U.S. population who claim Asian ethnic heritage. Therefore, it is necessary to create more efficient and cost-effective methods to find these customers.

While there is currently an indicator available for listed telephone sample that identifies Asian-American households, the indicator does not identify households by subsegment (i.e., Chinese, Japanese, Korean, etc.). For any studies that require quotas by subsegment, this indicator provides little help to the researcher. The indicator might be useful if the telephone surveys will only be conducted in English and there are no subquotas by ethnic segment. Without having any clue about a potential respondent's language preference, the data collection vendor will not know how to manage its interviewers or its subsegment quotas.

If the study is designed to offer the respondent a choice of languages in which to be interviewed, or if the client requires certain quotas by subsegment, then employing a surname sampling methodology is critical. The surname sample can be generated by ethnic group, so that bilingual Japanese interviewers, for example, will only be given sample targeted to people with Japanese surnames. The surname sampling method helps improve interviewing efficiency in that the data collection vendor can schedule interviewers based on the amount of sample per language and the quotas required per language group. If a Chinese interviewer discovers any Korean respondents (for example) in the Chinese sample, these names and phone numbers can be passed along to a Korean interviewer for re-contacting.

In order to increase the incidence of finding Asian-American households within a surname sample, a supplemental census tract approach is recommended. With this method, the sample supplier can identify census tracts within the targeted geographic markets that have some minimum incidence of Asian-American households - for example, 35 percent - as decided by the researcher. The sample supplier would then pull random listed sample from all census tracts that meet the minimum threshold, distributed between the geographic markets according to the preferences of the client.

The higher the threshold of Asian household penetration within a census tract, however, the less representative that census tract will be of all Asian households within that geographic target. The trick is to find a threshold that is high enough to increase the incidence of finding Asian households and yet still have a relatively high degree of representation of all Asian households in that market. It may be easier to pull sample only from census tracts that are 100 percent Asian, but census tracts with 100 percent penetration of Asian households are usually representative of only a very small percentage of all Asian households in that particular market. Chances are the study results would be skewed because the sample is not representative of the majority of Asian households in that market. The best option is to decide on a penetration threshold that is high enough to find concentrations of Asian households while remaining representative of at least 50 percent of all Asian households in that market. The sample supplier can calculate this incidence for the researcher.

Limitations of surname sampling methodology

One limitation to this method is the fact that surname sampling does not provide a perfect match. For example, some common Asian surnames - such as Lee (Chinese) and Park (Korean) - are also common surnames among other Asian groups or among the general American population. Many Filipino surnames are of Hispanic origin, which makes surname sampling for Filipino populations very problematic.

For the Filipino market, the best sampling strategy may be to avoid the surname sample methodology and focus on the census tract technique. The researcher and the sample supplier can identify geographic areas with high concentrations of Filipino households. Then, rather than a surname sample database, the sample supplier can use the Asian household sample indicator to target Asian households in those census tracts, as Filipinos are counted as Asian in the census database. While some non-Filipino Asians are likely to be included in the sample, the benefit is that nearly all of the Hispanic households (as well as African-American and Caucasian) will be filtered out.

Another limitation of this method is that households with unlisted telephone numbers and households without telephones are not represented in the sample. The surname sample only works with listed sample, as unlisted sample purchased from an independent supplier does not include surnames. However, the benefit of using a targeted surname approach is to provide a very cost-effective and highly representative solution to identifying and interviewing the particular Asian segment.

Because few marketers have tried to reach the Asian market, lists are scarce. In fact, all of the available Asian targeted lists are surname lists. Asian surnames are matched against various records to produce these lists. For companies who plan to use surname matching to find possible Asian customers in their own databases, it can sometimes be helpful to run a second sort on first name matches, too - if a list of Asian first names can be found or compiled. There has been an attempt to create lists for each of the major Asian segments that incorporate both surnames and first names, but such lists are not widely available in market research circles.

Since making assumptions based on surnames can be tricky - due to marriage, immigrants who "Americanize" their last names, and names that may be similar in certain cultures - this additional criterion can help reduce the risks of finding non-Asians in the database. However, this method is not fool-proof either, as many Asians have Americanized their first names. Neither of these methods is helpful in finding households where an Asian woman has married a non-Asian and Americanized her first name as well.