What the data actually shows
The foundational work is Barry Staw's research on escalation of commitment, beginning in the 1970s, which found that people who were personally responsible for a prior decision allocated more resources to it after it started failing — not less. The worse the early results, the more some people invested, specifically to justify the original choice. That is the counterintuitive core of the finding: failure can increase commitment rather than prompt withdrawal.
A large part of the mechanism is the sunk-cost fallacy. Arkes and Blumer's 1985 work showed that people factor past, unrecoverable costs into present decisions even though, rationally, only future costs and benefits should matter. Their classic demonstrations — finishing a meal you've paid for but no longer want, sitting through a film you're not enjoying because the ticket is bought — generalise directly to jobs, relationships, projects, and investments.
Underneath sits loss aversion, from Kahneman and Tversky's prospect theory: losses loom roughly twice as large as equivalent gains. Walking away crystallises the spent investment as a definite loss, which feels worse than the uncertain prospect of losing more by staying. So the asymmetry of how we weigh losses quietly tilts us toward continuing.
Why this feels different from how it actually is
From the inside, doubling down rarely feels like a bias — it feels like persistence, loyalty, or not being a quitter. Identity does real work here: many people hold a self-image of someone who follows through, and abandoning a failing course threatens that picture more than the practical loss does. The research suggests this self-justification is a major driver of why the person who made the original choice escalates more than an outsider would.
It also feels different because the past investment is vivid and the future cost is abstract. You can clearly recall the years, the money, the effort already sunk; the additional cost of continuing is hazier and easy to discount. So the books look more favourable to staying than they really are.
And admitting error is socially and emotionally expensive. Reversing course can feel like a public verdict on your judgment, so people sometimes keep going partly to avoid the moment of saying 'I was wrong' — to themselves or to others. The decision stops being about the project and becomes about defending the earlier decision.
The thing keeping you in is rarely a clear-eyed bet that things will improve; it's the pull of what you've already put in, plus the discomfort of being wrong.
What the research says to do about it
The most consistent suggestion is to set kill criteria in advance — concrete conditions, decided before you're emotionally invested, that will tell you when to stop. Because the bias operates by recruiting the original decision-maker's ego, deciding the exit rule ahead of time, while you can still think clearly, sidesteps much of the in-the-moment self-justification.
Reframing the decision in forward-only terms also helps: ask what you would choose today if you were starting fresh, with no history and no money already spent. This question deliberately strips out the sunk cost, which research suggests is the input that shouldn't be there in the first place. If you wouldn't start this now, that is information about whether to continue.
Bringing in an outside view is another supported move. Because responsibility for the original choice is what fuels escalation, someone with no stake in having been right — a trusted friend, a neutral adviser — tends to weigh the future costs and benefits more cleanly than you can. Asking 'what would you do here?' borrows their lack of ego involvement.
What the research says does not help
Simply trying harder or telling yourself to be more disciplined does little, because the trap isn't a willpower problem — it's a distortion in how the decision is being scored. More effort applied to a course you'd never freely choose today usually just deepens the commitment.
Waiting for certainty before quitting tends to backfire. People often tell themselves they'll stop once it's clearly hopeless, but escalation thrives precisely in ambiguity, and the spent investment keeps the goalposts moving. Holding out for an unmistakable signal can mean the signal never feels decisive enough.
Dwelling on how much you've already put in — re-totalling the years or the money to motivate yourself — is among the least helpful things, because it foregrounds exactly the sunk cost the research says should be ignored. It feels like accountability, but it strengthens the very bias driving the bad decision.
Walking away crystallises the spent investment as a definite loss, which feels worse than the uncertain prospect of losing more by staying.
What this looks like in real life
The 'would I start this today?' test
Someone stays in a project because of the years and money already poured in. The clarifying question is what they'd choose today with no history and nothing already spent — which deliberately strips out the sunk cost the research says shouldn't be in the calculation. If they wouldn't start it now, that's information about whether to continue.
Not a quitter
From the inside, continuing feels like persistence and loyalty rather than a bias. Abandoning a failing course threatens the self-image of someone who follows through more than the practical loss does — which is why the person who made the original choice escalates more than an outsider would. Borrowing an outside view from someone with no stake in having been right weighs the future costs more cleanly.
Real numbers in context
There isn't a clean population statistic for how often people double down — escalation of commitment is studied mainly through controlled experiments and case studies rather than national surveys, so this is a robustly observed mechanism rather than a measured prevalence. What the studies establish is the direction and reliability of the effect, not a percentage of the public it applies to.
The one number worth holding onto comes from loss aversion: across many studies, losses are felt roughly twice as strongly as equivalent gains. That asymmetry is the engine that makes 'I'll lose what I've spent' outweigh a level-headed look at future costs — and it's why walking away feels disproportionately painful even when it's the better move.