Finance & investing
Investors who switch financial advisors tend to recall their previous portfolio performance as worse than it actually was, inflating the perceived benefit of the new advisor and justifying switching costs.
Change Bias
aka Implicit Theory of Change Bias · Exaggerated Change Effect
Misremembering your past self as more different from your present self than they actually were.
WHAT IT IS
The glitch, explained plainly.
Imagine you took painting lessons. Afterward, you look back at your old paintings and think, 'Wow, those were terrible!' — but actually, they were okay. Your brain made them seem worse in your memory so it feels like you got way better from the lessons, even if you didn't improve that much.
Change Bias occurs when individuals who expect or believe that change has occurred in themselves reconstruct their past selves as more different from their present selves than is actually warranted by objective measures. When people hold an implicit theory that they should have changed — for instance, after therapy, a training program, or simply with the passage of time — they unconsciously revise their memories of past attributes, emotions, or behaviors to manufacture evidence of that expected change. This creates an illusory narrative of personal growth or transformation that may not reflect actual improvement. The bias is particularly potent when people are motivated to see themselves as having progressed, improved, or matured.
SOUND FAMILIAR?
Where it shows up.
- 01 After finishing a diet program, remembering having been much heavier before than was actually the case, making the results seem more dramatic.
- 02 Looking back at a first year at a job and recalling being far more clueless and incompetent than colleagues at the time would have described.
IN DIFFERENT DOMAINS
Where it shows up at work.
The same glitch looks different depending on the terrain. Finance, medicine, a relationship, a team — same mechanism, different costume.
Medicine & diagnosis
Patients in clinical trials or therapy programs misremember their pre-treatment symptom severity as worse than originally reported, inflating perceived treatment efficacy and complicating the assessment of genuine therapeutic outcomes.
HOW TO SPOT IT
Ask yourself…
- Am I remembering my past self as worse or more deficient than I would have described myself at the time?
- Do I have objective records (journals, scores, measurements) from before this change that I can compare to my current recollection?
HOW TO DEFEND AGAINST IT
The playbook.
- Keep contemporaneous records: journal entries, skill assessments, mood ratings, or measurements taken before any expected change so you have an objective baseline to compare against later recall.
- Use prospective rather than retrospective measurement: assess yourself at regular intervals using the same instrument rather than asking yourself to remember what you were like before.
FAMOUS CASES
In history.
- Conway and Ross (1984) demonstrated change bias in a study-skills program: participants who completed an ineffective program recalled their prior skills as worse than they had actually reported, while control group members recalled accurately. Six months later, program participants also overestimated their academic performance.
- Research on psychotherapy outcomes has shown that patients systematically recall pre-therapy distress as more severe than originally measured, inflating perceived treatment success even when objective improvement is minimal.
WHERE IT COMES FROM
Academic origin
Michael Ross (1989) formalized the theory of implicit theories of stability and change in memory construction. The empirical demonstration was provided by Michael Conway and Michael Ross (1984) in their study 'Getting what you want by revising what you had,' published in the Journal of Personality and Social Psychology.
Evolutionary origin
Perceiving personal growth and progress likely served adaptive functions by maintaining motivation to continue effortful self-improvement behaviors, such as skill acquisition and social learning. Ancestors who felt they were improving — even when objective change was minimal — would have been more persistent in practicing survival-relevant skills.
IN AI SYSTEMS
How the machines inherit it.
AI systems trained on self-reported improvement data (e.g., customer satisfaction surveys, therapy outcome surveys, educational assessment surveys) may inherit change bias from training data, as human respondents systematically overstate pre-intervention deficits. This can lead models to overestimate the effectiveness of interventions or programs when relying on retrospective self-reports rather than prospective measurements.
EXPLORE MORE
Related glitches.
LAUNCH PRICE
You read about it. Now drill it.
This page taught you the name. The deck turns the name into reflex. 1,100+ swipeable scenarios, 1,100+ defenses, 650+ detection prompts — spaced-repetition Swipe Deck, unlimited Spot-the-Bias Quiz, Defense Playbook, Pre-Flight, My Blindspots, Cheat Sheets, Field Guide e-book. $39.53$59.