> ## Content Index > Fetch the complete content index at: https://www.indykarveli.com/llms.txt > Use this file to discover other available public pages before exploring further. # What the Marshmallow Test Really Proved - URL: https://www.indykarveli.com/what-the-marshmallow-test-really-proved/ - Published: 2026-06-18T14:00:00.000Z - Updated: 2026-06-18T14:00:00.000Z - Author: Indy Karveli - Tags: Mind, Articles, #Import 2026-10-03 18:23 The marshmallow test is one of the most famous experiments in psychology. In it, young children were offered a choice: one marshmallow now, or two if they could wait. The children who waited, researchers later reported, did better in life. For decades, this was held up as proof that willpower and self-control in childhood predict success. But when scientists tried to repeat the experiment, it didn't hold up the way everyone believed. ## The original study In the 1960s and 1970s, researchers at Stanford ran the now-legendary experiment with young children at a campus nursery school. Kids were left alone with a marshmallow and told they could eat it now or wait a while and get two. Years later, follow-up research reported that the children who'd waited had better life outcomes as teenagers, better test scores, better behavior. The conclusion that captured the public imagination: self-control in childhood predicts success, and it's a fixed trait you either have or don't. ## What the replication found In 2018, researchers Tyler Watts, Greg Duncan, and Haonan Quan tried to reproduce the finding with a much larger and more diverse group, around 900 children, compared to the small original sample. Their results were far more modest. They did find some link between waiting and later achievement. But once they accounted for the children's *backgrounds*, their families, home environments, and socioeconomic circumstances, the link became much weaker. In other words, a large part of what had looked like pure "willpower" was actually explained by the child's environment and circumstances. ## Why the original was misleading The original children came largely from families connected to Stanford University, a narrow, privileged, unrepresentative group. Part of what looked like self-control may actually have been the security of a stable, well-off background. A child from a secure home has more reason to trust that waiting will be rewarded; a child from an unpredictable environment may rationally grab the sure thing now. What appeared to be a character trait was partly a response to circumstances. ## What it means The marshmallow test's collapse under scrutiny doesn't mean self-control is meaningless. A separate, larger study that followed about a thousand people from birth found that broader childhood self-control *did* predict adult health and wealth. So the underlying idea. That self-control matters, survived. What didn't survive was the famous *test* and the tidy, oversimplified story built around it: that a single marshmallow choice reveals a fixed trait that determines your future. ## The bigger lesson The marshmallow test is a perfect example of why it's worth checking famous claims rather than accepting them. A compelling story, repeated for decades, turned out to prove far less than everyone believed once it was examined carefully. This is exactly the kind of scrutiny worth applying to *all* confident claims about money and success, including the appealing ones. ## The honest limit To be fair to the original researchers, science advances precisely by testing and refining earlier findings, the replication isn't proof the original scientists were dishonest, just that the initial finding was narrower than the popular story made it. And self-control does appear to matter for life outcomes, per better-designed research. The specific lesson here isn't "self-control is a myth." It's that the famous marshmallow *test* proved much less than claimed, that environment explained a lot of what looked like willpower, and that even beloved scientific stories deserve scrutiny. --- ## Sources - Watts, Duncan & Quan (2018), *Psychological Science* (n=918) — a replication finding the marshmallow-test link much weaker once background is accounted for; original: Mischel et al. (1960s-70s). --- *The Selfmade Newsletter: one idea every Friday on what self-made people do differently. Free, on Substack.* [*Join here*](https://www.indykarveli.com/newsletter/)*.* *This article is for education only and isn't financial advice. Returns are never guaranteed.*