How to do it
- Write one line the day it happens. How bad, out of ten, in a month.
- Set a reminder for a month out. Then read the line and score today.
- Guess how long, not just how bad. The how-long part is where people got it most wrong.
- Keep the old lines. Next time you have your own record to check against.
How the forecast goes wrong
- Your forecastguesses too long
After break-ups, their candidate losing and job knock-backs, people expected the bad feeling to last longer than it did.
- The recoverygets left out
The authors' explanation is that when you predict, you forget you'll make sense of it and move on. They call it immune neglect.
- How badless clear
A later paper found the overestimate was mainly about how long, less about how bad at first. The original authors disputed it.
- Youcheck the forecast
Writing a prediction and reading it a month later is my idea. Nobody has tested it.
Why it works
When something goes wrong, you make a forecast: this will sit on me for months. In six studies, covering break-ups, missing out on tenure, their candidate losing an election, bad feedback and being turned down by an employer, people overestimated how long the bad feeling would last. Evidence: decent
The how-long part holds up better than the how-bad part. A later paper found the overestimate was mainly about duration, less about how bad it feels at first, and that it shrank when the questions were asked more carefully. Evidence: decent The original authors disputed that, and it isn't settled. Evidence: decent So the useful version is narrow: it'll probably pass sooner than you think, not necessarily hurt less than you think.
The long readEvery step of the chain, each study with its numbers, and what's still unproven. 2 min read.
What happens in your brain
This chain is behaviour, not brain areas. The first two stops are one paper, the third another, the last is mine.
Your forecast guesses too long. Across six studies, people predicting how they'd feel after a setback expected the bad feeling to last longer than it did. Evidence: decent
The recovery gets left out. The authors' explanation: when you predict, you leave out the ways you'll make sense of it and move on. They call this immune neglect, meaning you forget your mind has its own ways of absorbing bad news. Evidence: decent
How bad is less clear. A later paper found the overestimate was mainly about how long, less about how bad at first. Evidence: decent The original authors disputed it. Unresolved. Evidence: decent
You check the forecast. A line written today, read in a month, puts your own forecast next to what happened. That this helps is my guess. Evidence: worth trying
The proof
- Gilbert and colleagues, 1998: six studies, including break-ups, being denied a permanent academic post, and seeing their candidate lose an election. People overestimated how long the distress would last. Evidence: decent
- Levine and colleagues, 2012: the overestimate was mainly about duration, less about initial intensity, and it shrank with better questions. Evidence: decent Wilson and Gilbert disputed this. Evidence: decent
How to do it
- Write one line the day it happens: how bad, out of ten, you think it'll feel in a month. Evidence: worth trying
- Set a reminder for a month out. Read the line, then score today. Evidence: worth trying
- Guess how long, not just how bad. Duration is where the overestimate was clearest. Evidence: decent
- Keep the old lines. Next time you have your own record, not just a study. Evidence: worth trying
Watch out
The studies measured how long bad feelings lasted, not whether writing a forecast helps. The action here is mine and untested. Evidence: worth trying People overrate how long a setback hurts; whether they also overrate how bad it hits at first is disputed. Evidence: decent "You always bounce back fast" is the oversold version.
This is for setbacks: a knock-back, a lost deal, a break-up. It isn't for grief. Losing someone is.
Flat for a few weeks, with sleep or appetite off and no interest in things you used to like: that's not a forecasting error, it's a doctor visit.