Behavioural Economics for Market Researchers

Behavioural Economics is a key current topic. However, I feel the term itself is generally misused and its implications glossed over. This post is intended as homework and pre-reading for a workshop I am running at this year‰’s MRS Conference in London. The intention of this post and the workshop is to provide a grounding in what behavioural economics is, and what it is not, and to set up a review of its implications for market research.

Behavioural Economics represents a major change of direction for economics, with a wide range of implications for other disciplines. Classical economics was based on an assumption of rational and informed consumers, often referred to as ‰’econs‰’. Behavioural Economics suggests that psychological factors distort the predicted results of choices. The rise of Behavioural Economics was highlighted by the Nobel Prize for Economics that was awarded to psychologist Daniel Kahneman, and further popularised by books such as Dan Ariely‰’s Predictably Irrational and Richard Thaler and Cass Sunstein‰’s Nudge.

Behavioural Economics is re-shaping economics, business strategy, public policy, and is very slowly beginning to impact the way market research is conducted. This post looks at a few of the important findings in Behavioural Economics from a market research perspective and seeks to identify the key questions that need to be answered.

The Ultimatum Game
The ultimatum game, originally proposed by GÌ_th, Schmittberger, and Schwarze, provides a striking example of the difference between the real world and classical economics. In the ultimatum game there are two players, the first is given a sum, for example $10. The player with the money then makes an offer to the second player, offering them some of the $10. If the second player accepts, then the money is divided in those proportions, if the second player rejects the split, neither party receives anything.

In classical economics the person offering the money would tend to offer as little as they thought they could get away with, reasoning that the second person is better off with $1 than nothing, so offering them $1 should result in $9 for the person making the offer.

But experiments with this game, and a wide variety of alternatives, have shown that most people do not offer a very small amount (figures of 20% are often quoted), and when a small amount is offered the second player quite often rejects it (meaning both players receive nothing).

It has been suggested that there are at least three factors at play.

  1. The person making the offer knows what it feels like to be offered a bad deal (writers such as Mark Earls have highlighted the role of so called mirror neurons), making it harder to offer an unfair deal
  2. The person receiving the offer has to weigh the small monetary value of the offer against the disutility of feeling cheated.
  3. The person making the offer knows that if the recipient rejects a poor offer (for example because they feel it is insulting) they will both lose out.

It should be noted that the existence of the effects revealed by the ultimatum game were detected and quantified by carefully designed and implemented quantitative experiments. This is a pattern that is regularly repeated in most of the work that has been done with Behavioural Economics.

System 1 and System 2
One of the metaphors most closely associated with Daniel Kahneman is that of System 1 and System 2 thinking. Kahneman suggests that it is helpful to pretend that the brain operates with two systems. System 1 is fast, efficient, relatively hard working and gets us through most situations. If I ask you to solve the maths problem what is 2 times 3 you will use your system 1 brain to solve it.

System 2 is slower, more cognitive, capable of solving harder problems, but is lazy and tends to consume energy. If I ask you to tell me what is the answer to 17 times 24 you will engage your system 2 thinking. (Or, you will if you try to solve it, alternatively System 1 may simply encourage you to skip the problem).

Kahneman and others suggest that for every System 2 decision there are thousands of System 1 decisions. System 2 tends to be enlisted when System 1 calls it in (as with the 17 times 24 problem) or where System 2 spots that something is going wrong, for example when you suddenly find yourself lost when walking from A to B in a slightly familiar setting.

The concern in market research is that we might be talking to the wrong system. When a shopper selects a tub of margarine or butter from their regular store they are typically on auto-pilot, i.e. System 1 is in charge and making the decisions. If a market research study makes the shopper engage System 2 we may obtain a quite different answer to the problem ‰ – for example we might see in the survey that they are willing to pay more for a healthier or more ethical product when System 2 responds, but in the store System 1 keeps picking the regular brand.

However, Kahneman suggests that when a shopper compares the value of two washing machines they are engaging System 2. If a market research study is organised as a set of attitude grids we may find that the shopper is responding with System 1, which is fast and efficient, as opposed to System 2 which is slow and tiring. Kahneman describes the halo effect as being created largely by System 1, if I like X, I like everything about X, if I don‰’t like Y, I don‰’t like anything about Y.

Dan Ariely reported the impact of state of mind on ratings with his famous (infamous) arousal experiment. Ariely took two groups of male students and gave one group a series of questions under normal circumstances. The second set were asked to watch adult material and to self-stimulate themselves to be aroused and then answer the questions. The questions were the same for both cells, and were reported using a 0 to 100 scale.

Amongst the results that were significant at the 99% level were:

  • Are women‰’s shoes erotic? Control=42, Aroused 65.
  • Is a woman sexy when she‰’s sweating? Control=56, Aroused 72.
  • Can you imagine having sex with a 50-year-old woman? Control 28, Aroused 55.
  • Is kissing just frustrating? Control 41, Aroused 69.
  • Would you keep trying to have sex after your date says ‰’no‰’? Control 20, Aroused 45.

The most immediate implications from the research is that men who are reasoned when not aroused, may be less reasonable when aroused. One wonders what the results might have been like if Ariely had also created a cell that were aroused and had been drinking alcohol?

For market researchers the question is whether asking decontextualised questions invalidates the results of the research? For example, can a fast food study capture the choices that people will make when in a restaurant, experiencing hunger, smelling the food, and seeing others order large meals with extra fries?

The Impact of the Default
One of the key themes in the book Nudge is the impact of the default and the power of inertia. One often quoted example is Germany and Austria in the context of organ donation. Germany operates an opt-in model and has a donor rate of 12%. Neighbouring and German speaking Austria has an opt-out system and the donor rate is 99%. In the US state of Illinois when they changed from making the organ donor questions on licence renewals optional to compulsory organ donations jumped from 38% to 60%. Note, in Illinois there was no opt-in or opt-out, the change was that drivers had to pick one or the other, a procedure often referred to as forced choice.

In most market research situations the choices are forced. The respondent does not have to say yes, they do not have to say no, they do not have to say don‰’t know, but they do (usually) have to say something.

The Rating Problem
In ‰’How We Decide‰’, Jonah Lehrer reported on several experiments that showed how asking people to rate things altered their choices. For example, in one study two cells of respondents were asked to choose which strawberry jam they preferred from a choice of jams. The first cell simply made their choices, and these choices turned out to be predictive of market related assessments. The second cell were asked to fill out questionnaires about the jams they were testing. This cell made different choices, and indeed their most preferred jam was the worst performing jam according to Consumer Reports.

Lehrer posited that asking people to think too much about strawberry jam resulted in them focusing on attributes that don‰’t really matter. Kahneman would probably say the questionnaires were asking System 2 to do a job that in the real world is done by System 1.

To what extent are market research questionnaires asking people to rate things that they should not be rating?

Priming, Anchoring, and Framing
Both Kahneman and Ariely have a lot to say how the context of a task (in terms of things like priming, anchoring, and framing) impacts the results. Kahneman describes how exposing people to a high number tends to result in them selecting higher numbers later in the experiment, and exposing them to lower numbers tends to result in lower numbers later on. Ariely shows how showing something in a set with weak comparators improves its score and showing it a set with strong competitors reduces its score.

Some of these effects have been known by researchers for many years. For example, Kahneman highlights the order effects of describing two people, Alan and Ben.

  • Alan is intelligent ‰ – industrious ‰ – impulsive ‰ – critical ‰ – stubborn ‰ – envious
  • Ben is envious ‰ – stubborn ‰ – critical ‰ – impulsive ‰ – industrious ‰ – intelligent

Our perceptions of Alan and Ben are going to be shaped by the order of the things we heard about them.
Similarly, Kahneman reports on the impact of asking people to do something with the word EAT, and later asking them what is the missing letter in SO_P? A frequent response would be Soup. However, if people had been exposed (primed) to the word WASH then SO_P would have produced many more occurrences of Soap.

I‰’ll Have What She‰’s Having
Mark Earl‰’s new book ‰”I‰’ll have what she‰’s having‰” makes the point that we can probably estimate what hem length and colour most of our respondents‰’ new dresses will be next year, what colour most of the cars our respondents will buy in two years, and what some of the Christmas presents most of our respondents will be buying in 9 months. However, the answer is less than useful in that it is simply ‰’The same as everybody else!‰”. In many areas of consumption and behaviour respondents can‰’t tell us specifically what they will buy and do because they are programmed to do what other people do. Respondents aren‰’t exercising some form of free will and making their own executive choices, they are letting System 1 apply the heuristic of herd activity.

The Implications for Market Research?
I think some of the key questions for market research are:

  1. How much of this is really new news, and how much is stuff that has been known for years, but is perhaps being addressed more formally and holistically now?
  2. Which bits of behavioural economics matter, in the sense that they should change what we do or how we interpret results?
  3. How do we think market research currently works, and does that need to change?
  4. What new opportunities does Behavioural Economics create for market research?
  5. Most of the work on Behavioural Economics has been by creating tightly defined and controlled quantitative experiments. Does Behavioural Economics have any implications for qualitative research?

Subscribe to the Vision Critical blog

Get free customer intelligence tips and resources delivered weekly to your inbox.

By completing this form you consent to receive emails from Vision Critical. You can unsubscribe at any time. Learn more in our privacy policy.