What the data actually shows
A big part of the engine here is cognitive dissonance, the idea developed by Festinger that holding conflicting thoughts is uncomfortable and motivates us to resolve the conflict. After making a decision, one cheap way to reduce that discomfort is post-decision rationalization: we mentally upgrade the chosen option and downgrade the rejected one, so the choice feels more clearly correct than it was when it was genuinely a close call.
Memory research gives this a name: choice-supportive bias, studied by Mather and Johnson among others. The finding is that people tend to misremember the features of a decision in a self-flattering direction — attributing more good qualities to what they chose and more bad qualities to what they rejected than the original information supported. The reconstruction is biased toward justifying the choice already made.
All of this is possible because memory is reconstructive rather than a recording. Work associated with Loftus shows that remembering is an active rebuilding of the past, which makes it editable — vulnerable to suggestion, expectation, and our present needs. When the present need is to feel good about a past decision, that editability quietly works in the choice's favor. These are well-supported patterns, though their size varies across people and situations, so it's a tendency rather than a fixed quantity.
Why this feels different from how it actually is
It doesn't feel like rewriting because the editing happens beneath awareness and produces a memory that feels just as solid as any other. You don't experience a revision; you experience a recollection — one that conveniently lines up with the choice you made. The seam doesn't show.
It also feels different because the rewrite usually makes the story more coherent, and coherent stories feel true. 'I chose well, for good reasons' is tidier than 'it was a genuine toss-up and I'll never know.' The mind tends to prefer the clean narrative, so the version that justifies the choice tends to win out and then feels like simple fact.
And there's no contradicting record to check it against. Once the original ambivalence has faded, the only account left is the edited one, so there's nothing to flag the change. The absence of a 'before' copy is exactly what lets the 'after' version pass as the truth.
You don't experience a revision; you experience a recollection — one that conveniently lines up with the choice you made. The seam doesn't show.
What the research says to do about it
The most useful move the research points to is humility about your own memory of past decisions — treating a strong feeling of 'I clearly chose right' as a possible product of the bias rather than proof. Because choice-supportive memory works quietly in your favor, the confidence you feel about a past choice is weak evidence that it was actually the better one.
Where it matters, recording your reasoning at the time helps, since memory is reconstructive and editable. Decisions you wrote down — what you expected, what the trade-offs were — give you a 'before' copy the bias can't quietly revise, which is closer to honest self-knowledge than a later recollection.
When evaluating others, the same caution applies in reverse. People defending a past choice may be reporting a sincerely rewritten memory, not spinning you, so reading confident hindsight as proof of a good decision tends to overstate the case. None of this is a way to switch the bias off — it's mostly there to keep you appropriately skeptical of the tidy stories memory hands you.
What the research says does not help
Assuming your memory is an accurate playback of how a decision actually felt doesn't help, and is roughly the opposite of what the research suggests. Memory is reconstructive, and after a choice it tends to reconstruct in a self-justifying direction, so 'I remember it clearly, so it must be right' is exactly the judgment the bias would produce.
Trying to simply stop the rewriting through willpower also tends not to work. Choice-supportive bias operates largely outside awareness, so you can't reliably catch it as it happens; aiming to feel no pull toward justifying your choices generally fails. The realistic target is skepticism about the conclusion, not control over the process.
Swinging to the other extreme — assuming every favorable memory of a past choice is a delusion — isn't supported either. The bias is a tilt, not a guarantee that every recollection is wrong, and some of your past choices genuinely were good. Treating all self-favorable memory as fabricated overcorrects past what the evidence shows.
The confidence you feel about a past choice is weak evidence that it was actually the better one.
What this looks like in real life
The 'obviously right' choice that was a toss-up
You picked one job over another after agonizing — it was genuinely close at the time. A year later you remember it as clear-cut: your job had all the upside, the one you turned down had all the red flags. That confident memory feels like solid recollection, but it may be choice-supportive bias tidying a close call into a clean story so the decision looks wiser than it was.
The written note that catches the rewrite
Before deciding, you jot down what you expect and the trade-offs on both sides. Months later, when memory has quietly rearranged the story in your favor, that note is a 'before' copy the bias couldn't revise. Comparing it to your current recollection is often the only way to see how much the past has been edited to justify the choice.
Real numbers in context
It's worth being precise about what this evidence claims. These findings are about direction and mechanism, not magnitude: cognitive dissonance (Festinger) explains why we rationalize after deciding, choice-supportive bias (Mather and Johnson) describes the self-flattering slant in what we remember, and reconstructive-memory research (Loftus) explains why memory is editable enough for this to happen. None of them put a clean number on how much any given memory has been rewritten, and this shouldn't be read as one.
The honest summary is a reliable tendency rather than a fixed quantity. Most people, most of the time, remember their chosen paths a little more kindly and their rejected ones a little more harshly than the original facts warranted — and the size of that slant varies by person and situation. It eases regret, which is a genuine benefit, while quietly costing accuracy about our own history, which is the part worth keeping in mind.