Every day when we wake up and make choices about what to have for breakfast, what clothes to wear or which friends to reach out to. When we make these choices what does our brain do? Does it calculate using a complex model to predict our decisions or does it just use some kind of shortcuts to make decisions? The latest study of neuroscience lays some evidence to solve this mystery. Long story short, the study says that we use a combination of these two strategies, but there’s more.
Before we dive into the practicalities of the study here is a brief background about human decision making in the brain. Humans decision making goes through two different ways, one which is patient and calculative called model-based system and the other called model- free which is fast and easy but relies on shortcuts.
So what is this model free?
Model-free is sometimes popularly known as heuristics. Heuristic refers to any approach to problem-solving that uses rules of thumb not guaranteed to be perfect (or optimal in- science speak), but enough for present goals (Korn and Bach, 2018). A good example of a heuristic is availability heuristic (Tversky and Kahneman, 1973): We tend to value things more than we remember. This is perhaps the reason people were spending a huge premium (Kahneman, 2011) on Terrorism insurance. Even though the probability of a terrorism-related event is far less compared to car accidents or obesity.
And what’s the deal with Model-Based?
Our brain makes a model of the world for us. Decisions using this model which is based on probabilities of things happening. This method of decision making makes optimal or right decisions based on the information you might have. The main caveats for using this best system is that it is slow and consumes a lot of energy (Korn and Bach 2018).
Well, there could be many reasons why multiple decisions making process could be useful. It might save time, effort or perhaps help in specialisation. However, one thing remains both have costs and benefits in using them. So how does our brain decide which one to use? This is what this exciting piece of research by Korn and Bach (2018) tried to answer. The result they found was quite intriguing yet intuitive.
To understand how we make decisions the researchers developed a multi-step task where the participants had to choose whether they would forage or rest in a forest. This task was accompanied by varying energy levels and weather uncertainties. There were 5 different steps that one to uncover to finishes this experiment.
The main logic was that if the participants choose an optimal model-based decision making then, in the end, they would get higher payoffs than if they choose something like a heuristic based action profile.
Some Cool Results
1 Using Heuristics lead to faster choices: Well not something we didn’t know already. Interestingly when there two options which contradicted the model based answer and model-free answer, the model-free answer was more quickly taken.
2 More time was taken when the advice contradicted: Do you remember the last time when you were coming home from work, and then on the way, you forget that you wanted to buy milk on the way? What do you do? You wait and then slowly make decisions. This is similar to what participants. They became slower as their advice from the first impulse was being contradicted by their more rational model based mind.
3 Time allotted matters: The researchers also found from the seeing the reaction times that as less time is given to decision making the decision became more and more heuristic-based. Well we knew already that heuristic decision making is quick and easy, but this experiment shows that alter the choice mechanism by allotting time.
4 Increasing Uncertainties increases the time taken: This is more interesting because as participants found they were under more uncertainty they started to take more time to make choices. Perhaps this is where their heuristic, model-free part started to communicate with the model based part.
What this means for us?
You might ask why does all this matter? So what is the deal that we use some combination of these two models? I believe what matters is not that when we are making normal decisions we use two systems. What is more important that these two systems can be switched by changing the perceived risk or time allotted. This is crucial because in our attention-grabbing digital era all the marketers want us to choose their product. So how do they do it? Perhaps by playing with these two parameters of risk and time.
What is more important that these two systems can be switched by changing the perceived risk or time allotted.
I am sure you must have seen these two parameters being played out in real life by big companies. Take, for example, Ebay’s countdown timer. It is an excellent way to make sure that when we are buying something on auction we are time constraint and perhaps rely on our heuristic based decision making and make some cognitive mistakes.
To conclude the research by Korn and Bach sheds more light on how we make decisions. Their final result is that we use both models based and model free systems. They, however, do not tell us what is the how much do we rely on one system to make a decision. This is difficult to ascertain as every decision we make changes with dynamic contexts. What they interestingly showed that our decision-making process can change with changes in some particular contexts such as time and riskiness. Something to keep in mind when making the next purchase at high street or Amazon!
What do you feel about our situation? Do you think two is better than one?
Kahneman, D. (2011). Thinking, fast and slow. Macmillan.
Korn, C. W., & Bach, D. R. (2018). Heuristic and optimal policy computations in the human brain during sequential decision-making. Nature communications, 9(1), 325. Retrieved April 14, 2018 https://www.nature.com/articles/s41467-017-02750-3
Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive psychology, 5(2), 207-232. Retrieved April 14, 2018 https://msu.edu/~ema/803/Ch11-JDM/2/TverskyKahneman73.pdf
Acknowledgement: I thank Dr Korn for useful comments to improve this blog post