How does stress affect decision making?

You have no appetite. Your palms are sweaty and your heart is thumping. Your anxiety rises. In short, you’re stressed.

This scenario probably feels familiar. We all routinely experience stress, particularly ahead of significant events, when we’re working under tight deadlines or when faced with complex decisions.

But how does stress affect us when we’re making decisions?

There’s already been some research that shows stress has negative effects on our basic cognitive abilities, like our memory.

But how this effect then ‘scales up’ to more complex behaviours, like decision-making or problem solving, wasn’t clear.

Multiple cognitive processes are important and deficits in some areas, like memory, may be compensated for elsewhere. For example, we may pay closer attention to make up for any problems recalling information.

Our study, published in Communications Psychology, ran several controlled experiments to explore the role of stress in our decision-making.

We looked at the decision-making of 42 participants under the kind of stress you feel before a job interview – that acute, in-the-moment kind.

The surprising finding? Stress alone had little to no effect on decision quality until it was paired with time pressure.

The mock interview

To make our study participants feel stressed, we put them through a tried and tested three-part task, called the Trier Social Stress Test.

In the first part, we give them five minutes to take notes and prepare a speech pitching for their dream job.

Then, we took their notes off them, sent them into another room and gave them five minutes to deliver their speech to two ‘experts’ (these were, in fact, colleagues).

These ‘experts’ are wearing lab coats, filming the participants on a camera and providing zero feedback.

They don’t smile. They don’t nod. They don’t even answer questions.

Finally, in part three, the experts gave our participant five minutes to do a tricky maths task aloud where, whenever they make a mistake, they need to start all over again.

To compare their behaviour under stress to a baseline, the study participants also completed a control session on a different day without the stressful elements.

No observers, no filming and no penalties for mistakes.

How we measure stress

To determine whether inducing stress worked, we recorded several measures of stress.

To measure the concentration of cortisol, the body’s main stress hormone, we took multiple saliva samples. Other measures included perspiration and pupil diameter – as well as simply asking participants how they were feeling.

Study participants wore sensors on their fingers to measure their perspiration. 

As expected, when participants were under stress, their cortisol levels shot up, pupils dilated, and negative emotions spiked.

Studying complex decision-making

We also had our study participants make 72 ‘packing’ decisions within a time limit.

Imagine you’re packing items, each with a weight and value, into a backpack, but there are two rules.

Firstly, the combined weight of the packed items can’t exceed the backpack’s maximum capacity. Secondly, the combined value of the packed items needs to hit a minimum dollar target.

Your goal is to decide, yes or no, whether there’s a combination of these items that meets the rules.

This task simulates real-world decisions involving constraints: from shopping with a budget constraint to task scheduling with a time constraint.

It also helps us study two types of decisions: finding where benefits outweigh costs (the correct answer is ‘yes’) and where costs outweigh benefits (the correct answer is ‘no’).

We expected that stress would reduce decision quality, particularly on the harder levels.

The packing task. In this case, is there a combination of items that add up to at least $158, but whose weights add up to at most 80kg? 

Interestingly, stress alone had little to no effect on decision quality and any potential effect didn’t seem to depend on a level’s difficulty.

So, we went back to the drawing board.

Noticing that participants ran out of time on around a third of their decisions, we wondered whether decisions made under time pressure would be those where stress’ effects were strongest.

Sure enough, without time pressure, stress had no effect on decision quality. But with time pressure, we saw three key changes under stress.

First, eye-tracking hinted at changes in attention: people noticed more of the available items but spent less time looking at them before shifting their attention elsewhere.

Second, there was a marked increase in ‘no’ responses, suggesting people grew more pessimistic that a combination of items meeting the rules actually existed.

Finally, this heightened pessimism saw average decision quality drop by five percentage points.

On decisions when the target value was achievable within the weight limit, stress and time pressure caused decision quality to plummet from a strong 83 per cent down to 57 per cent – barely better than guessing.

So, our research revealed when stress is most damaging (when it’s paired with time pressure), how it may potentially mislead us (increase in pessimism with less focused attention), and the types of decisions that may become more challenging (identifying genuinely beneficial opportunities).

Stress is a complex phenomenon

An important caveat is that the drop in performance relied on the unplanned combination of stress and time pressure.

We’re currently following up this study with others to replicate these findings and identify ways to help improve decision-making in this context.

Until then, what should you keep in mind before your next exam or job interview? Stress is a complex phenomenon – it can improve some cognitive abilities and impair others.

But be warned: when stress is paired with a tight deadline, the risk of mistakes may rise significantly.

This article was published by Pursuit. The authors would like to thank their colleagues who posed as stress intervention actors: Chaoyi Chen, Dr Elizabeth Bowman, Michelle Lee, Hassan Andrabi, Declan Grindrod and Zoe Su.