'Rational' decisions and when to avoid them

Table of Contents:

Being fully instrumentally rational means choosing an option that maximizes out subjective utility. However, there are limits on our ability to do so, such as our limited WM capacity, so we constrain the information we process. There are also uncertainties involved in analyzing one option over another. The kind of reasoning this entails is generally an attempt to weigh up the pros and cons of a decision, bearing in mind that there may well be unconscious as well as conscious influences on what constitutes the various pros and cons. As a result, we rely on heuristics and biases in the ways we think. However, if one were to make the conscious effort to be as rational as possible in our decision-making, what would that involve? Suppose you have to make a decision about something important such as asking someone to marry you or buying a car. Given their relative importance, it would be reassuring to think that such decisions, although subjective, are based on some analysis of the pros and cons or the costs and benefits to us of a particular outcome. A rational model of decision-making assumes we can rank various preferences. For example, in buying a house, we might rank the room sizes as being probably more important than the garden size, or the kitchen layout as more important than the size of the main bedroom, or whatever. Another assumption is that we try to look at all the available information and use it to work out the probability of success -in this case, choosing the best house from the ones on offer -given that there is a degree of uncertainty about much of the information we use.

One view of human rationality from economics, the Homo Economicus model, was based on the idea that people try to maximize their expected utility. According to this, people use all the information they have access to. Furthermore, we have consistent preferences and we tend to avoid biases in processing the information available (for more details see Corr & Plagnol, 2019). There is even an equation that can be used to calculate the expected utility of an outcome:

where U is the utility of a particular state and EU is the Expected Utility of an action or decision (A) where the outcome is uncertain. X is the state reached after performing the action A and has a number of values (x, to x_n) and a distribution (p! to p_n).

So, every time you decide to buy a sofa, you sit down with a scientific calculator and work out the expected utility of buying this sofa. Don’t you? Well, perhaps not. You could, however, do something simpler such as list the pros and cons of buying the sofa and giving each one a weight, positive and negative. Add them up and voila. Although it appears reasonable, in real life, things don't quite work like that. Have a look at the scenario in Box 4.2 where this mechanism is applied. It contains an example of what Corr and Plagnol (p. 89) referred to as ‘more Homer (Simpson) than homo economicus’.

So, a rational approach to decision-making in situations like this is probably not going to work. Indeed, in the early stages of a romantic relationship, we can be very attuned to the benefits but very blind to the costs. Juliet didn't seem to pay all that much attention to the cons of running off with Romeo, nor did Paris perform a cost-benefit analysis of abducting Helen and carting her off to Troy. This isn't to say that we don't make good choices, but rather that we don't use all available information and we don't compute a rational choice from that information because we don't have time to do all that thinking. And most of the time, we don't need to. What we do instead is make a quick decision based on limited criteria and choose the first one that meets those criteria. This is called satisficing (Simon, 1956). It means choosing an outcome (a restaurant, or a car, or a partner) that is good enough. Although the word is often attributed to him, it is a Northumbrian word reflecting the fact that our decisions are often made based on the degree to which they satisfy some criteria and on the extent to which they, in turn, suffice in allowing us to achieve a goal. Another example would be when you're on holiday in an unfamiliar town. It’s the evening, you’re hungry, and you want something to eat. You wander along the town centre looking

BOX 4.2 HOW NOT TO SWEET TALK YOUR PARTNER

"Hi, Jennifer?”

"Yes?”

"Hi, it's George."

"Oh, hello love, I didn't recognize your voice for a moment. How are you?"

"Great. Listen, I had a really great time the other night.”

"Mmmm. me too.”

"So, what I’ve done is ... I’ve been trying to work out how profitable this relationship is likely to be.”

“Huh?”

"What I’ve done is I’ve drawn up a list of all the rewards I can get out of our relationship and put a rough figure beside each one and then I've done the same with the costs. But listen, the great thing is when I subtract the costs from the rewards I'm still left with quite a high figure. Isn't that great?"

"Are you having me on George?”

"No really. I'll read out some of my list. You should do one of your own, too, to see what you think. Ok, so. Rewards first. No particular order here: You look good, 10 points; you've got really good taste, 6 points; you seem to like the same activities as me, 6 points; you're good in bed, 15 points; you do that thing I like, you know the one ... anyway that has to be worth a good 12 points. You’ve ..."

"George ..."

"No wait, there's more. On the costs side, right? I've got: My mum doesn't like you, 7 points; you don’t earn much money, 8 points; you spend a fortune on clothes - that’ll have to be 10 points, I'm afraid."

“George!”

"Yeah?"

[Click. Dialling tone]

"Jennifer? Jennifer?” at the numerous restaurants available. It would take a very long time to examine and compare all the menus and the ambiance and so on. You might compare just a few and settle on one that meets some criteria of your choosing (price, dishes on offer, background music not too loud, etc.). Satisficing can be regarded as a rational strategy in a world where we don't have access to all the information or time to process it all. We can make the best use of what we've got instead. Our rationality is limited and spending time and effort working things out ‘rationally’ is often unnecessary and sometimes counterproductive. Most of the time, it is enough to rely on heuristics and habits. Some heuristics allow us to make fast and reasonably accurate judgements. Others can take the form of ‘biases' in thinking where our thinking is suboptimal. This is the subject of the next chapter.

Summary

There are different ways of defining rationality, the main ones being normative rationality where human reasoning can be judged by the criteria of logical norms, and instrumental rationality judged by the extent to which our thinking allows us to achieve our goals. Both are linked to our beliefs about the world (epistemic rationality).

Syllogisms involve a number of assertions (premises) leading to a conclusion based on the structure of the argument - its syntax - rather than its content - its semantics. The simplest structure that we tend to get correct is modus ponens - if A is the case and A implies B, then B must follow otherwise A would not imply B. Similarly, if B is not the case and A implies B, then A can't be the case either (modus tollens). We are less accurate at detecting the invalidity in such cases.

Induction is the process by which we learn about the world. Once a rule or category is learned, it can be applied to new instances through deductive inference. Abductive reasoning is based on the likelihood of A given B even if A is not a strictly valid conclusion.

The New Paradigm takes human beliefs and updating evidence into account based on Bayesian probabilities and the extent to which people take prior and posterior probabilities into account in their thinking. If people do not update their beliefs based on new evidence, no matter how overwhelming that evidence is, then they may end up deriving false conclusions, and making errors of judgement.

A rational model of human decision-making was proposed by classical economists based on the assumption that people tried to maximize their expected utility. However, humans are not always very good at doing so. There was, therefore, a shift to behavioural economics and the heuristics and biases approach to thinking.

'Rational' decisions and when to avoid them

Summary

Suggested Further Reading