In an excellent paper that won the silver award for best overall paper in the 2011 Esomar congress, Behaving Economically With the Truth, Orlando Wood, Peter Harrison and Alain Samson show the importance of behavioural economics for market research.

They created a framework of how people make decisions based on the latest thinking from cognitive, consumer and social psychology. The framework works very well for ethnographic and qualitative research, and a few of the principles described are very relevant for traditional quantitative research. The list below has been picked from their work; it can give directions for the organization of most quantitative research.

1.    Personal factors

• Brain processing is a scarce resource, so we tend to be save on it:

• Our brain only processes salient information, and it uses mental shortcuts.
• We behave with as much routine as possible to avoid thinking (autopilot).

• Our rational side (system 2, as described by Stanovich and West, and as opposed to our intuitive/emotional system 1) over-estimates our ability to make rational decisions in the future and is a rather poor predictor of behaviour.
• Ego and image are key motivators in decision making; we look for esteem (self-esteem and esteem from others). This may explain why we sometimes give inaccurate descriptions of our own behaviours; we want to display a positive image rather than sharing personal beliefs if these are not social-proof (e.g. one respondent may not be willing to say “I couldn’t care less about the environment” in a research on ecological sustainability).

2.    Environmental factors

• Context can influence our choices even unconsciously (with music, light, atmosphere). The impact of environmental conditions is demonstrated in The tipping point, in which M. Gladwell describes how improvements in the subway environment in New-York effectively reduced criminality in the subway.
• The choice, and the way it is presented influences what we choose. This is largely discussed in Nudge (Thaler and Sunstine 2009). For the sake of quantitative research we should cite:

• Primacy/recency effect, which makes us forget items in the middle of a list. Anchoring is similar to primacy.
• Framing, which means that questions can highlight different aspects of the outcome and have an impact on our choice.
• Finally, asymmetrically dominated alternatives bias our choices. One experiment cited by Wood, Harrison and Samson illustrates this well: Participants were given the option of choosing an elegant pen or $6. The pen was chosen by 36 percent. When the choice of a less elegant pen was added as a decoy, however, making a third option available, 46 percent chose the elegant pen.

• We are influenced by others: we copy each other and make decisions at least to some extent based on others. This human characteristic is the basis of innovation diffusion models (Bass 1969).

These insights on how our brain works and how we make choices have huge implications for quantitative consumer research. There are things we should start doing, and others we may want to stop doing if we want to get the best insights from quantitative research. Here are a few things that come to mind when applying the knowledge from behavioural economics to quantitative research:

1. The questionnaire flow (order of questions), the wording of questions as well as short item lists are proven to be a no-miss for quantitative research.
2. Further to these classical best approaches, we should put particular focus on the available choice for respondents to each question, as each alternative has an impact on respondents’ answers. For instance we should avoid choices that include a clearly dominated alternative.
3. Asking for an absolute measure of interest given a specific stimulus is risky. History has witnessed a considerable amount of concept evaluation on a purchase intent scale. According to behavioural economics, this approach is not optimal. Because it is heavily dependent on the concept used and the way the question is asked (framing). Because a 5-points scale does not give respondents with a choice between alternatives; this results in significant over-claim making interpretation difficult (limited variability in the data, need for a benchmark or a data-base to interpret the results). Because it lacks the relevant context; no consumer will ever see the new product alone on a store shelf.
4. Rather than capturing only rational views (brain system 2) through declarative answers which we know are not 100% predictive of future behaviours, we should strive to capture quick emotional responses from respondents (brain system 1) through experiment:  large scale observation, tachitoscope, eye-tracking, time-constrained choices or implicit association tests.
5. The choice context should be set in a way that matches people reality. This sounds trivial, yet many quantitative tests are set-up without providing the right context. Paired comparison product usage tests are somewhat flawed, as only very few consumers will eventually use two products in a short period of time and think about the two products relative to each other. And this can sometimes lead to surprises in the innovation process: think of a pack claim that was winning a MaxDiff test with written claims as stimulus; now that the pack is being developed, the claim cannot be integrated because it is too long… tough luck.
6. Another context element that is sometimes forgotten is budget. In economic theory, people make choices to maximize their utility under budget constraint. We know that “maximizing utility” is a difficult notion to handle, but we can assume that respondents do that when they answer a quantitative questionnaire. However, the budget constraint is sometimes forgotten when research is designed and it heavily influences respondents’ choice. Do you really want to conclude from research that budget-constrained consumers want to buy the most premium product? We may want to adapt purchase intent questions to reflect respondents’ budget in addition to the price of the alternatives.
7. When consumers purchase and use products, they tend to do so following a routine. A quantitative research is unfortunately breaking that routine, which does not let respondents act as usual. Standards seem scarce in this area (In-Vivo BVA may have a lead for store dynamics); the closer research can be to actual purchase or usage occasion the better.
8. The environment in which the research is conducted should be optimized, or at least monitored. As an example, does it really make sense to have a door-to-door, interviewer-administered questionnaire if we want to capture in-store purchase decisions?
9. Actually, does it make sense to have an interviewer at all, considering respondents are influenced by others? This bias cannot be controlled. What’s more, if we believe an interviewer is needed to ask pre-coded open questions or to clarify the questionnaire, we may want to think whether the questions are the right ones to generate insights.
10. How can we capture human interactions and impact on behaviours as we do research? There might not yet be an answer to this. Predictive markets are good to predict success/failure of different ideas, but they assume limited interactions amongst respondents. Social networks may be the future in understanding individuals’ interactions about new products, and provide further insights into innovation diffusion dynamics.

Of course, this list of ten considerations is not exhaustive, but it should already give a hint as to how quantitative research should evolve to generate deeper and more accurate insights based on the latest knowledge of how the brain works. Behavioural economics is a young science, and with advances in neuroscience and psychology it is rapidly developing. The ability of researchers to adapt their approaches based on discoveries in behavioural economics will be critical to offer best-in-class quantitative research in the future.