Articles & Blogs

Varying Use of Response Scales, Globally

When collecting and analyzing global data, one of the most important questions we need to ask, as researchers, is whether or not observed differences are a true finding or due to response bias.  Response bias may be defined as a “systematic tendency to respond to a range of questionnaire items on some basis other than the specific item content” and when this bias is applied consistently across time and situations, the bias is said to be a response style[1].  When conducting global market research, we may observe one or more response styles as a function of the degree of differentiation between the cultures being studied. This differentiation may be characterized along any number of dimensions; however, the individualism-collectivism dimension continually stands out with respect to the research industry due to the impact these stances have on larger personal constructs related to communication and information processing.  Research has shown that one of the net results of these stances is that individualist cultures (traditionally Western cultures) tend to be characterized by an extreme response style (i.e., tendency to use only the extremities of the scale) while collectivist cultures (traditionally Eastern cultures) tend to be characterized by a midpoint response style (i.e., tendency to use predominantly the middle of the scale)[2]. As discussed, these results are due to larger processes related to how information is communicated, stored, and retrieved and are therefore, correlated with questionnaire and scale design.

It is generally agreed that the use of contextual information, when understanding and forming a response to a question, is universal; however, the extent to which this information is used varies from one culture to the next.  Along the individualism-collectivism continuum, this impact would follow as less sensitive to more sensitive and consequently, have a significant impact on pragmatic interpretation – the respondent’s understanding of what we, as researchers, are ultimately interested in (this is decidedly different from the literal interpretation). In general, this has implications for how we communicate intent not only in the questions we’re asking but also in the anchors we use on scales – a simple “strongly disagree” to “strongly agree” scale may seem universal; however, the concept of “agreement” is categorically different for a respondent from an individualist culture as compared with a respondent from a collectivist culture.  Further, the amount of information being considered, contextually or otherwise, directly impacts response formation.  This is especially true when considering the formation of attitude objects and standards with individualist’s being more likely to form attitudes through objective contrasts and collectivist’s being more likely to form attitudes through assimilation[3]; responses to any form of attitudinal question will be biased. In totality, there is a commonality within these predispositions that speaks to differences in autobiographical memory. The frame of reference from which a respondent will think about their behaviors, attitudes, and opinions, as a function of the recall of their life’s experiences, varies by cultural disposition and sets a differential locus of response.

Global market research often hinges on the belief that cultures are comparable to one another with respect to the instrument being used to study them; response biases (construct, method, and item) do not vary systematically with culture and a single, translated survey instrument is applicable to all markets.  However, when considering culture along the individualism-collectivism dimension and the subsequent impact on conversational maxims we can imagine some deleterious effects of this line of thinking.  Whether a culture is individualist or collectivist will necessarily impact how they the communicate, encode, and retrieve information; significant response bias in a survey setting can exist between cultures with different response styles and threaten the validity of empirical findings. As researchers, it is important to consider where each country being studied falls along this continuum and make appropriate accommodations both in design of research and analysis of data.

[1] Paulhus, D.L. (1991). Measurement and control of response bias. In: Measures of Personality and Social Psychological Attitudes, Volume 1. San Diego, CA: Academic Press Inc.

[2] Ayse K. Uskul, Daphna Oyserman, Norbert Schwarz, Spike W. S. Lee, and Alison Jing Xu (2013). How Successful You Have Been in Life Depends on the Response Scale Used: The Role of Cultural Mindsets in Pragmatic Inferences Drawn from Question Format. Social Cognition: Vol. 31, Situated Social Cognition, pp. 222-236.

[3] Schwarz, N., Oyserman, D., & Peytcheva, E. (2010). Cognition, communication, and culture: Implications for the survey response process. In J. A. Harkness, M. Braun, B. Edwards, T.P. Johnson, L. Lyberg, P. Ph. Mohler, B.E. Pennell, & T.W. Smith (eds.), Survey methods in multinational, multiregional and multicultural contexts (pp. 177-190). New York: Wiley.