A 2026 reanalysis of a well-known “scarcity mindset” experiment has challenged one of the theory’s central explanations: that having too little of something causes people to focus more narrowly on the immediate task.
The reanalysis found that the game’s structure could explain why players given fewer shots appeared to aim more accurately and borrow more often. It reproduced the original effects, but argues that those results do not demonstrate attentional “tunneling.”
The findings do not settle every question about scarcity. In particular, they do not test the separate claim that financial worries reduce overall mental capacity.
What the original game showed
In the original “Angry Blueberries” experiment, 68 online participants shot blueberries at targets to earn points that increased their chances in a raffle. Each level contained seven targets, with three bonus points available for clearing them all.
Players assigned to the “rich” condition received 150 shots—15 per level. Those in the “poor” condition received 30 shots—three per level. In some versions of the game, players could borrow additional shots after using their allocation while targets remained. Each borrowed shot cost two future shots, equivalent to 100% interest.
Across the whole game, “poor” players averaged 2.31 points per shot, compared with 1.67 for “rich” players. They also took longer to aim. The original interpretation was that limited resources encouraged more focused attention, producing better shots despite the pressure.
But the full-game averages combined early shots, when many targets remained, with later shots, when screens were emptier. The first shot averaged 2.81 points, the third 1.80, and a “rich” player’s fifteenth shot 0.69.
When the analysis was limited to the first three shots of each level, the difference was much smaller: 2.63 points per shot for “poor” players and 2.53 for “rich” players. That difference was not statistically significant, with p = 0.72. The reanalysis says neither group differed significantly on any individual shot in this early sequence.
Why the researchers questioned the attention explanation
“Poor” players could take only three shots per level when borrowing was unavailable. “Rich” players used about 5.7 shots per level, including later attempts made after many targets had already been removed. This difference in the number and timing of shots could lower the “rich” group’s average without showing that “poor” players were more accurate.
The original study was too small to establish that the groups were equally accurate during the first three shots. However, the original authors’ 2019 replication, involving about 1,000 players, found no meaningful difference in that early-shot comparison.
The reanalysis also examined the claim that greater focus led “poor” players to make better borrowing decisions. Borrowing was available on 45% of levels for “poor” players, compared with 2% for “rich” players. The difference largely reflected the rules: players had to use their allocated shots while targets remained before the borrowing option appeared.
When borrowing was available, “poor” players borrowed in 69% of cases, compared with 100% for “rich” players. That difference fell just short of statistical significance, with p = 0.052, and only 40% of “rich” players ever reached the borrowing choice.
A reported link between longer aiming and borrowing among “poor” players also weakened when the analysis included everyone who could borrow. The original subset produced a correlation of r = 0.43, but the correlation across all eligible players was essentially zero, at r = 0.03, and did not replicate in the 2019 study.
A new test using insurance offers
Fiona tho Pesch, Antonia Langenhoff and Mahesh Srinivasan conducted a new preregistered experiment with 174 online participants: 91 assigned to the “poor” condition and 83 to the “rich” condition.
Instead of borrowing, participants received insurance offers during the first 10 levels. They could spend current shots to insure a separate field of bonus points against a 50% chance of drought. Eight offers were good value, while two were “scam” offers that cost more than participants could pay back.
When both groups received the same offers, “poor” players bought an average of 4.74 of the eight good-value policies, compared with 4.92 for “rich” players. The difference was not significant, with p = 0.68, and an equivalence test found the groups statistically equivalent.
“Poor” players again aimed longer and scored more points per shot across the whole game, while matching “rich” players on the first three shots. The researchers describe the insurance results as, if anything, showing more care among “poor” players—not the poorer decisions predicted by attentional tunneling.
What the findings do—and do not—show
The reanalysis also found the same overall pattern in the original authors’ 974-person replication and reported similar results in supplementary reanalyses of two other studies from the 2012 paper. A separate 2018 project had already failed to replicate one of the studies in a large retest of social-science experiments from Nature and Science.
Scarcity theory includes two major ideas. One is a “bandwidth tax,” in which scarcity is thought to consume mental capacity. The other is attentional tunneling, in which attention is drawn toward the scarce resource or immediate problem.
The new work addresses the second idea, not the first. It therefore does not show that money worries have no effect on thinking. The best-known test of the bandwidth-tax claim, a 2013 Science study, reported lower test scores after financial prompts among lower-income but not higher-income shoppers, and lower scores among farmers before harvest than after.
The experiments in this reanalysis were online laboratory games using points. They did not recreate poverty or real financial hardship. Participants were recruited through Amazon’s MTurk, and the original samples were not representative of the U.S. population.
The authors also note that the study tests one paradigm and one proposed mechanism. The new experiment analyzed 174 participants rather than the 200 originally preregistered because of budget limits, and its planned multilevel analysis was changed to the simpler approach used by the original authors. The game software did not record who quit, so systematic dropout cannot be ruled out.
For readers, the broader lesson is cautious rather than conclusive: an apparent psychological effect can depend on how an experiment allocates opportunities, measures performance and averages results. The reanalysis disputes the attentional explanation for this game, while leaving the wider question of how scarcity affects thinking open to further testing.
AI tools were used to assist with the preparation of this article.
Leave a Reply