012 min
Action bias in a penalty kick
A penalty kick gives the goalkeeper a problem with almost no information. The ball is struck from 11 metres away, so a keeper who waits to see where it goes is too late. The keeper has to pick a direction before the kick, and almost every keeper picks a side.
In 2007 the researchers Michael Bar-Eli, Ofer Azar and their colleagues counted what top keepers do. They studied 286 penalty kicks from top leagues and championships. Given where kickers tend to aim, the best choice for the keeper was to stay in the centre of the goal. The keepers did not do that. In 93.7 percent of the kicks they jumped to the right or to the left.
Staying put is not only a theory. Goalkeepers who jumped left stopped 20 of 141 kicks, a rate of 14.2 percent. Out of the three choices, the centre had the highest chance of stopping the ball.
A keeper who dives left while the ball goes down the middle has left the middle of the goal open. That keeper looks busy and still concedes the goal. Why would a person paid to stop goals choose the option that stops fewer of them?
Why staying in the centre is not a sure save
Staying in the centre is the best average choice, not a guarantee. A kick that the researchers labelled centre did not have to hit the exact middle, so a keeper who jumped to one side could still reach some of them. Keepers who jumped left stopped 9.8 percent of the kicks aimed at the centre. The advantage of staying comes from where kickers aim, and it would shrink if kickers aimed differently.
023 min
Why action bias happens
The answer starts with regret, the bad feeling of wishing a choice had been different. Daniel Kahneman and Dale Miller proposed norm theory in 1986. It says people react more strongly to an outcome that follows a departure from what is normal, because the normal alternative is easy to imagine. In most areas of life doing nothing is the normal choice, so a bad result after acting hurts more than the same result after waiting. That pattern is called omission bias.
Penalty kicks reverse the norm. A keeper is expected to jump. A goal conceded after staying in the centre therefore feels worse than the same goal conceded after a dive, and the keeper avoids the worse feeling by diving. Bar-Eli's team argued that this is the reason for the 93.7 percent.
The older finding behind it
Kahneman and Amos Tversky had shown the feeling in 1982 with two investors. Paul considered switching his shares to another company and decided against it. George did switch, from the other company's shares into the first. Each ended up $1,200 worse off than if he had made the other choice. A large majority of the people asked said George, who acted, would feel more regret.
What the keepers said
The researchers also asked 32 top professional goalkeepers how bad they would feel after a goal in three cases: a jump left, a jump right and staying in the centre. Fifteen gave the worst rating to all three, so their answers could not be compared. Of the other fifteen, eleven said a goal after staying in the centre would feel worse than after a jump. Four said the opposite. The group is small and the gap is not large by usual statistical standards, so it supports the idea without settling it.
When the norm flips, the regret flips
Later experiments tested the norm idea directly. When social norms favour inaction, people regret an action more than an inaction. When norms favour action, that gap becomes smaller in two experiments and reverses in two others. The feeling follows what is expected, not what is active.
A simpler explanation that fails
One simple guess is probability matching, where people pick each option about as often as it pays off. Keepers who did that would stay in the centre about 28.7 percent of the time, the share of kicks aimed there. They stayed in the centre in only 6.3 percent of kicks. Matching the odds does not explain the dives.
Other forces behind the pull
Regret is probably not the only cause. Acting gives a person a sense of control and an answer when someone asks what was done. Effort is easy to see, so acting also makes a person look capable. Colleagues and managers often read waiting as indifference. Some writers add an evolutionary story, that quick action paid off more often than waiting in dangerous places, but that story is hard to test. These forces are harder to measure than the keeper data, so treat them as likely contributors and not as proven causes.
031 min
Where action bias comes from
The pull toward action had been described before it had a name. The label action bias entered the journals in 2000, in "Action Bias and Environmental Decisions" by the economists Anthony Patt and Richard Zeckhauser. They did not treat acting as wrong. Their point was narrower: people often have good reasons to prefer action, and the bias begins when that preference is carried into situations where the reasons do not apply.
The two-site experiment
Patt and Zeckhauser tested the idea with surveys of imagined environmental decisions. In one, respondents could help only one of two sites: improve one, or stop the other from getting worse. The framing made it arbitrary which site was which, yet respondents chose to improve rather than to prevent decline. They also found what the authors call strong action bias, where people chose to gain something even though it meant causing a loss elsewhere. The authors tied these results to loss aversion, status quo bias and omission bias.
Bar-Eli's penalty study, published seven years later, took the idea out of surveys and into a setting with real money, real careers and many repeated choices. Other names for the same pattern are the action effect and commission bias.
041 min
How bad news strengthens action bias
A keeper's norm is fixed by the sport. In other settings events decide the norm, and bad news is the strongest event. Zeelenberg, van Dijk, van den Bos and Pieters named this the inaction effect in 2002. One of their scenarios compared two soccer coaches. When a team had just lost, people rated the coach who changed nothing as more regretful than the coach who changed the team. Before the loss, most had expected the opposite.
The result matters outside sport, because a failed plan makes doing something feel necessary, whatever that something is. In four experiments, people deciding whether to keep funding a failing project leaned toward action when the action was framed as continuing. When withdrawing was framed as the action, they leaned toward withdrawing.
Escalation of commitment
The second finding connects action bias to a well-known error. Researchers had explained why people keep funding failing projects by pointing to the money already spent, the sunk cost fallacy. The experiments add a second cause: negative feedback by itself pushes people to act. If the easiest act on offer is to add more money, they add more money. The effect size across the experiments was modest, with a Cohen's d of 0.37.
052 min
Action bias at work: incidents, managers and markets
Take an on-call engineer paged at night because response times have doubled. This is an illustrative case. The cause is unclear. The engineer restarts a service, then scales up the database, then rolls back the last release, and each step changes the system and hides what the previous step did. Waiting ten minutes with the dashboards open would have shown that a batch job was the cause and was about to finish. The engineer chose action for the reason the keeper did: a visible response is easier to defend than a pause, even when the pause is the better move.
The same pull appears wherever someone is judged on visible effort.
- A new manager reorganises a team that was working, to look decisive.
- A doctor orders a test or a prescription that will not change the outcome, because leaving the patient with nothing feels like neglect. The related pattern is intervention bias.
- A driver leaves a slow lane for a detour and arrives later.
- A parent steps into a problem a child would have solved alone.
- A policy-maker announces a measure that is easy to see and does little.
Investors who trade too much
The best-documented business case is trading. Brad Barber and Terrance Odean studied the accounts of 66,465 households at one large discount brokerage between 1991 and 1996. The households that traded most earned 11.4 percent a year. The market itself returned 17.9 percent. The authors link heavy trading to overconfidence, but the cost shows how a steady habit of doing something can lose to the habit of doing little.
Why the workplace rewards it
Many companies praise a bias for action. That phrase fits decisions that are cheap to reverse and slow to learn about, where acting is the fastest way to find out. It fits badly when a decision is expensive to undo, or when the situation will change by itself. Teams under deadline pressure, or in cultures that read hesitation as weakness, tend to lean further toward action. The pull is also easy to exploit: a phishing email that demands an immediate click uses the same impulse.
062 min
How to guard against action bias
The engineer's night suggests the first fix: make waiting a choice that someone has to name, not a default that feels like failure. Several methods do this, and a team can use them together.
Ask what happens if nothing is done.
Before acting, write down what the situation will probably look like in an hour without any change, and what the action costs if it is wrong. If the first answer is that things stay fine and the second is that the action is hard to undo, wait.
Add a delay to big moves.
A rule such as no hiring, purchase or reorganisation until 48 hours after the idea gives the urge time to fade. It costs nothing for small decisions, so apply it only above a size you set in advance. The same habit helps with the busy-but-unproductive afternoon, where many small actions replace the one that matters.
Break the action into a test.
Replace "restart everything" with one change and a time to look at the result. Changing one thing at a time keeps each effect visible.
Name the problem first.
Write the problem in one sentence before acting. Teams often act on a symptom and leave the cause alone.
Check the result.
Set a date to see whether the action changed anything. If it did not, count it as activity and not as progress.
Ask a second person.
A short review by a colleague who does not feel the pressure often catches an action nobody needed.
Decide the rule before the pressure arrives.
People make more patient choices about the future than about the present. A binding choice made in advance, such as a trading limit or an incident runbook, takes the decision out of the moment.
Precommitment in the evidence
A study of bank customers in the Philippines offered some people a savings account that held their money until a goal they chose was reached. Those who were offered it saved on average 81 percent more than people with only an ordinary account. The lesson for action bias is that a rule set in advance can do what willpower cannot.
Why knowing the bias is not enough
Reading about action bias will not remove it. People tend to see a bias in other people more easily than in themselves. For that reason, a written rule or a colleague who asks what happens if the team waits works better than a resolution to be careful. Training also has limits: it helps most when it is specific to a task done soon and often, and it does not carry over easily to new settings. Treat the methods above as habits of the team, not as a lesson each person learns once.
071 min
When acting is the right choice
Acting is often the right call, and a rule that always says wait is as risky as one that always says act. Patt and Zeckhauser stated it directly: the preference for action has good reasons behind it. In a fire, a medical emergency or a security breach already in progress, delay has a cost, and a response plan that was rehearsed beats a debate.
Two limits are worth knowing.
- The keeper may be right to dive. Economists have pointed out that a penalty kick is a game, and a keeper who always stood in the centre would soon be punished by kickers who aimed for the sides. The 93.7 percent shows a bias only if kickers keep aiming where they did in the sample. The study counts as evidence of a pull toward action, but it does not prove every dive is a mistake.
- Overcorrecting is also a bias. A person who has learned about action bias may start to wait in every case. That trap is close to analysis paralysis, where thinking replaces doing. The test is the one in the earlier section: compare the cost of waiting with the cost of a wrong action.
081 min
Action bias vs. nearby concepts
Action bias is easiest to confuse with its mirror image, so the table puts the axis next to each neighbour.
| Concept | What it prefers | What separates it from action bias |
|---|---|---|
| Not in the library yetAction bias | ||
| PsychologyOmission bias | ||
| ResearchIntervention bias | ||
| PsychologyAnalysis paralysis | ||
| Not in the library yetStatus quo bias |
Omission bias and action bias are two outcomes of one rule: the choice that breaks the local norm carries the heavier regret.
?5 questions
Questions people ask
Is action bias the same as a bias for action?
How do you overcome action bias?
Why do goalkeepers dive when staying is better?
When is action bias useful?
Does action bias affect investing?
§6 sources
Sources
Bar-Eli, M., Azar, O. H., Ritov, I., Keidar-Levin, Y., & Schein, G. (2007). Action bias among elite soccer goalkeepers: The case of penalty kicks. Journal of Economic Psychology.
Patt, A., & Zeckhauser, R. (2000). Action bias and environmental decisions. Journal of Risk and Uncertainty, 21(1), 45-72.
Norm theory and the action-effect: the role of social norms in regret following action and inaction.
When action-inaction framing leads to higher escalation of commitment: a new inaction-effect perspective on the sunk-cost fallacy.
Show all 6 sourcesShow fewer sources
Barber, B. M., & Odean, T. (2000). Trading is hazardous to your wealth. Journal of Finance, 55(2), 773-806.
Soll, J. B., Milkman, K. L., & Payne, J. W. (2015). A user's guide to debiasing, in The Wiley Blackwell Handbook of Judgment and Decision Making (the authors' copy of the chapter).
