What good decisions predict

Decision making is a skill of its own

The strongest evidence that decision making is more than intelligence under another name comes from the Adult Decision-Making Competence (A-DMC) battery. Wändi Bruine de Bruin, Andrew Parker and Baruch Fischhoff built seven short tasks covering different parts of deciding: resisting how a choice is framed, applying decision rules, judging your own knowledge, sticking to consistent risk judgements, ignoring sunk costs and so on.

Two findings matter for you. People performed fairly consistently across the seven tasks, so the aggregate score was reliable. And people with higher A-DMC scores reported fewer negative life events that stem from poor choices, measured by a Decision Outcomes Inventory that ranges from the serious (declaring bankruptcy) to the minor (missing a flight, getting blisters from sunburn). That link held after controlling for cognitive ability, demographics and decision style. The authors concluded that decision-making competence appears to be a distinct construct relevant to real-world decisions.

A follow-up in 2015 went outcome by outcome and found that most of the listed negative outcomes were significantly more common among people with lower A-DMC scores, even after allowing for age and socio-economic status.

Judgment and Decision Making (O*NET)

The US Department of Labor's occupational database defines this workplace skill as considering the relative costs and benefits of potential actions to choose the most appropriate one. It is rated for every occupation, which tells you how routinely jobs are described in these terms.

What each game family has predicted

Each of our decision games descends from a research task with its own evidence trail. The trails are uneven, and it helps to know which is which.

Card tasks: a clinical success story

The Iowa Gambling Task was built to catch a problem that intelligence tests missed. Patients with damage to the ventromedial prefrontal cortex kept choosing decks that paid big now and lost more later, while their IQ scores looked normal. Bechara and colleagues later tested people dependent on alcohol or stimulants: 61 per cent of them performed within the range of the brain-injured patients, against 32.5 per cent of healthy controls. Performance was not explained by IQ, memory or standard executive tests, and it was related to things like years of use, relapses and the ability to hold down a job.

That is strong evidence the task picks up something real in clinical groups. It is weaker evidence about healthy graduates. A review of the task's construct validity by Buelow and Suhr supported its use in clinical assessment but flagged three problems: no concise definition of which part of decision making it measures, too little data on its reliability, and sensitivity to mood and personality. And Steingroever and colleagues, analysing eight data sets, found that healthy participants often prefer the decks with infrequent losses, behave idiosyncratically, and do not show the neat explore-then-exploit pattern the task assumes.

Balloon tasks: linked to real risk-taking

The Balloon Analogue Risk Task was introduced by Carl Lejuez and colleagues in 2002. In their first study of 86 adults, riskier pumping correlated with sensation seeking and impulsivity, and with self-reported addictive, health and safety risk behaviours. The task explained variance in those behaviours beyond what demographics and questionnaires explained, which is the result that made it popular. Its stability is respectable: in a study that retested people on separate days about two weeks apart, adjusted pumps correlated at r = 0.77.

Trust games: they predict being trustworthy

The most striking field result for social decision games comes from microfinance. Dean Karlan measured Peruvian borrowers in a trust game and followed their loan repayments. People the game identified as trustworthy (those who returned more) were less likely to default on their loans. People identified as trusting (those who sent more) were not. A study by Glaeser and colleagues found a similar asymmetry: standard survey questions about trust predicted trustworthy behaviour in the experiment much better than trusting behaviour.

The lesson is subtle. How much you hand over to a stranger says more about your beliefs and risk appetite than about your character. How you treat someone who trusted you says more about you.

The honest limits

One session is a noisy sample

One of the largest comparisons of risk measures asked 1,507 adults to complete 39 of them, including the balloon task, card tasks and lottery choices, and retested 109 of them six months later. Two findings should shape how you read any single game result.

First, behavioural tasks barely agreed with each other or with questionnaires about risk. Second, they were the least stable measures over time.

• General risk factor R0.85
• Self-report propensity0.68
• Frequency of risky activities0.65
• Behavioural tasks0.46

A reliability of 0.46 means a single run of a behavioural task is a rough reading. It is also why careful vendors compare patterns across several games rather than one number, and why our own scores pool many sessions, weight recent ones more, and pull small samples towards the middle until more evidence arrives.

Many of these tasks have no single right answer

Commercial game-based assessments are candid about this. Describing the pymetrics balloon game in 2017, Business Insider reported that a cautious approach is neither better nor worse than an adventurous one, and quoted the company's co-founder and chief executive, Frida Polli.

There is no right or wrong in the spectrum, and we're not just making that up to make you feel good.

In a 2021 audit written jointly by Northeastern University researchers and pymetrics staff, the games are described as derived from peer-reviewed studies and "not meant to be won or lost". Instead, a model compares a candidate's profile to high-performing employees already in the role. What employers look for, then, is fit to a profile, plus the things any decision task reveals: whether your behaviour is consistent across similar situations and adaptive when the evidence changes.

The link to job performance is mostly untested in public

The same audit states plainly that it did not investigate whether the games measure human capabilities or whether those capabilities map to job performance. We found no peer-reviewed study showing that balloon, card or trust-game scores predict job performance in graduate hiring. That does not mean the link is absent; much vendor validation is private. It does mean you should treat claims in either direction with care.

For perspective, even long-established selection tools are less predictive than once thought. A 2022 re-analysis by Sackett and colleagues concluded that range-restriction corrections had inflated many classic validity estimates, cut typical estimates by 0.10 to 0.20, and found structured interviews the top-ranked single method.

Why it still matters for you

Put the evidence together and three practical conclusions follow.

1. The underlying skills matter even where the games are noisy. Updating on evidence, calibrating risk and reading counterparts are what A-DMC-style competence is made of, and that competence tracks real outcomes beyond IQ.
2. Consistency is information. Because single sessions are noisy, what stands out to an assessor is a coherent pattern: you explore and then commit, you bank earlier on fragile balloons and later on sturdy ones, you trust partners who repay and stop trusting those who do not.
3. Understanding the task removes avoidable noise. Misreading the rules, panicking after one loss, or rejecting offers out of pique are not your personality. They are errors, and they are the part of your result you can fix.

Remember

• Decision-making competence is measurably distinct from cognitive ability and predicts fewer bad real-world outcomes.
• Card tasks are strongest in clinical groups; in healthy adults the classic assumptions often fail.
• Balloon-task riskiness relates to self-reported risk behaviour and is reasonably stable across days.
• In the field, trust games predict trustworthiness (loan repayment) better than trust.
• Behavioural risk tasks have modest reliability (about 0.46 over six months), so single sessions are rough readings.
• Vendors say these games have no right answer; what shows is consistency, adaptiveness and fit to a role profile.

Next, where decisions are tested looks at which employers and exams use these tasks, and what they say they measure.

Sources

1. Bruine de Bruin, W., Parker, A. M., & Fischhoff, B. (2007). Individual differences in adult decision-making competence. Journal of Personality and Social Psychology, 92(5), 938–956. doi.org/10.1037/0022-3514.92.5.938
2. Parker, A. M., Bruine de Bruin, W., & Fischhoff, B. (2015). Negative decision outcomes are more common among people with lower decision-making competence: An item-level analysis of the Decision Outcome Inventory (DOI). Frontiers in Psychology, 6, 363. doi.org/10.3389/fpsyg.2015.00363
3. O*NET OnLine. Transferable skills: Judgment and Decision Making (2.B.4.e). US Department of Labor. onetonline.org
4. Bechara, A., Dolan, S., Denburg, N., Hindes, A., Anderson, S. W., & Nathan, P. E. (2001). Decision-making deficits, linked to a dysfunctional ventromedial prefrontal cortex, revealed in alcohol and stimulant abusers. Neuropsychologia, 39(4), 376–389. doi.org/10.1016/s0028-3932(00)00136-6
5. Buelow, M. T., & Suhr, J. A. (2009). Construct validity of the Iowa Gambling Task. Neuropsychology Review, 19(1), 102–114. doi.org/10.1007/s11065-009-9083-4
6. Steingroever, H., Wetzels, R., Horstmann, A., Neumann, J., & Wagenmakers, E.-J. (2013). Performance of healthy participants on the Iowa Gambling Task. Psychological Assessment, 25(1), 180–193. doi.org/10.1037/a0029929
7. Lejuez, C. W., Read, J. P., Kahler, C. W., Richards, J. B., Ramsey, S. E., Stuart, G. L., Strong, D. R., & Brown, R. A. (2002). Evaluation of a behavioral measure of risk taking: The Balloon Analogue Risk Task (BART). Journal of Experimental Psychology: Applied, 8(2), 75–84. doi.org/10.1037/1076-898X.8.2.75
8. White, T. L., Lejuez, C. W., & de Wit, H. (2008). Test–retest characteristics of the Balloon Analogue Risk Task (BART). Experimental and Clinical Psychopharmacology, 16(6), 565–570. doi.org/10.1037/a0014083
9. Karlan, D. S. (2005). Using experimental economics to measure social capital and predict financial decisions. American Economic Review, 95(5), 1688–1699. doi.org/10.1257/000282805775014407
10. Glaeser, E. L., Laibson, D. I., Scheinkman, J. A., & Soutter, C. L. (2000). Measuring trust. Quarterly Journal of Economics, 115(3), 811–846. doi.org/10.1162/003355300554926
11. Frey, R., Pedroni, A., Mata, R., Rieskamp, J., & Hertwig, R. (2017). Risk preference shares the psychometric structure of major psychological traits. Science Advances, 3(10), e1701381. doi.org/10.1126/sciadv.1701381
12. Feloni, R. (2017, June 28). Consumer-goods giant Unilever has been hiring employees using brain games and artificial intelligence. Business Insider. businessinsider.com
13. Wilson, C., Ghosh, A., Jiang, S., Mislove, A., Baker, L., Szary, J., Trindel, K., & Polli, F. (2021). Building and auditing fair algorithms: A case study in candidate screening. Proceedings of FAccT '21, 666–677. doi.org/10.1145/3442188.3445928
14. Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2022). Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection for restriction of range. Journal of Applied Psychology, 107(11), 2040–2068. doi.org/10.1037/apl0000994