Yusof Ahmad | Head of Methodology
Why cultural differences matter
Market Researchers have long understood the challenges of undertaking research across national boundaries. Not only were actual cultural conditions acknowledged to be different enough to require differences in questionnaires, but sometimes the subject matter of questions would not translate well across national boundaries. These problems are receiving renewed emphasis and scrutiny today with the emergence of cultural groups with vastly different approaches to consumption, responses to authority, and managerial styles within most of the advanced economies of the West.
At the same time, the vast increase in the number of multinational corporations and the broader availability of consumer products across cultures, along with the use of web-based technologies for quantitative and qualitative data collection, have made issues of cross-cultural comparability of survey instruments and discussion guides even more prominent. Intersectional studies have also shown that people may respond differently to life situations and the world around them based on intersections of gender identity, race, age, and sexual orientation. This means differences in managerial styles, organizational behavior, and consumption behavior. How can market researchers take account of these differences to arrive at optimized survey designs and valid cross-cultural comparisons?
Accounting for cultural differences
Assuming we have taken care of the most important cultural differences in content, there remains the well-documented issue of dealing with differences in how respondents from different cultures use measurement scales, with respondents from some cultures more prone to using the extreme ends of the scale and those from others more likely to steer clear of extremes. So, how do we deal with such differences?
At Phronesis Partners, our approach to dealing with these issues of cross-cultural research rotates around the following pivots:
1. Ethnographic research: Use background desk research and qualitative in-depth interviews to understand differences in cultural content and build them into questionnaire items (i.e., questions, battery items, and selection options):
-
- If part of our cross-national study requires understanding how respondents spend their leisure time, we need to build options that reflect the most common ways people across the cultures under study distribute their leisure time. In South Korea, for example, work-leisure preference and the structure of leisure time activities changed significantly after the reduction of the working hours per week to 52.
- Firms in developing economies tend to show a greater tendency toward one-way communication from firm to a supplier, especially for smaller suppliers. In more developed economies, two-way communication is more common. A cross-national channel partner satisfaction study may include questions about communication style if such cultural differences are known to exist.
2. Be cognizant of cultural biases in response styles:
-
- Acquiescence bias (“yea”- saying behavior) – if this bias is known from prior research, questions need to be framed in a way that controls for this bias. This means using questions that do not include batteries of affirmative responses to be answered using scales (e.g., “We are always cognizant of the experience our retailers have when working with us”). In consumer studies, such bias has been associated with extraversion and impulsiveness and also with lower education and income levels.
- Social Desirability Bias – This is where respondents tend to respond in ways that they consider socially desirable by their estimate of their society’s belief system (e.g., “It is important for me that the products I buy are environmentally friendly”). Again, where this is suspected, it is always more optimal to use neutral questions rather than socially desirable affirmations. Another option is to use the more complex strategy of asking respondents to choose between alternatives that may differ on the attribute of importance (in our example, we can use two products that vary only in price and the nature of raw materials sourced).
3. Control for differences in cultural contexts when phrasing questions:
-
- In both B2B and B2C research, we need to understand that choices and interactions of economic agents occur in different cultural contexts. Where the context is absent from some questions across different cultural groups, the responses would be based on mental contexts that are unknown to the researcher, making comparison difficult. The use of anchoring vignettes tries to resolve this issue by first establishing the cut-off points that different groups use to assess a situation. For example, before taking respondent self-ratings on health consciousness, we may want to describe the routine of a certain hypothetical individual and then ask respondents to rate the respondent on their level of health consciousness. Later self-ratings can then be anchored around this vignette.
4. Control for scale use bias:
-
- Studies have shown strong differences in the way respondents use market research scales. For example, Latin Americans have been shown to tend to use the more extreme ends of Likert-type scales. Scale use bias makes comparability of results from different sub-groups or different country-level data unreliable. Analytical tools can detect and correct for such differences in the scale used to make the results more comparable. It has also been found that such differences in scale use can be minimized by using longer scales like 10-point scales anchored at the endpoints.
5. Identify and use prior research that has studied general differences in national research:
-
- At Phronesis, we have found it useful to look at international differences in culture using the framework of Hofstede’s seminal work. While some of the results of this work would have become outdated due to changes in cultural behavior, especially in Eastern economies like Japan, South Korea, and China, the theoretical framework of understanding cultural differences through the lens of dimensions like collectivism (vs. individualism), power-distance, uncertainty avoidance, etc., remain valid and helpful in understanding and interpreting the results of market research (MR) studies.
6. Use MR scales that have been cross-validated across cultures
-
- MR scales that have been validated across cultures and are robust in cross-cultural research are safer than ad-hoc scales when dealing with muti-cultural research. However, such cross-culturally cross-validated scales are sometimes unavailable, and researchers must exercise care to develop and test scales that reflect such cultural differences and work well across cultures.
Investing in design, analysis and interpretation
In conclusion, cross-cultural market research presents issues of design as well as analysis where comparisons across cultures have to be made. Researchers may often need to interpret whether differences are actual and can be interpreted in terms of actual cultural differences or whether they reflect problems of comprehensibility or response style and scale-use bias. Adequate time must be invested at the design stage to allow for cultural differences, while analysis and interpretation of results also require detecting and correcting biases.
