Actions tend to hurt more in the short term, but over a lifetime people often dwell more on the chances they never took — a classic psychological finding that has survived several modern replication attempts, even if the effect isn't universal.

A classic line of psychology research describes a reversal in regret. Soon after a bad outcome, people tend to be more troubled by something they did. When they look back across a much longer period, failures to act become more prominent.

The pattern has survived modern replication attempts, but in a qualified form. A large 2022 preregistered replication found support in three of four tested paradigms, generally with smaller effects. A 2023 field study of 2,600 people reproduced the shift across time horizons, yet did not find that inaction clearly dominated people’s lifetime regrets.

This is a pattern across several studies, not a universal rule about how every person will remember every choice. It is also not a licence to treat action as automatically wiser than restraint.

The original paper reconciled two conflicting findings

Thomas Gilovich and Victoria Husted Medvec published “The temporal pattern to the experience of regret” in the Journal of Personality and Social Psychology in 1994. The paper addressed what looked like a contradiction.

In short-term scenarios, people often judged an unfortunate action as more regrettable than an otherwise similar failure to act. This became known as the action effect. Yet when people were asked about their largest regrets in life, they often named omissions: an education not pursued, a relationship not repaired or an opportunity left alone.

Gilovich and Medvec argued that time helped explain the difference. A regrettable action has a concrete outcome, making it easy to replay soon afterwards. People can also work on explaining, repairing or learning from a completed act. A failure to act leaves no single outcome to process. Its imagined alternatives can remain open and become more favourable in hindsight.

Their 1995 review of the evidence proposed several mechanisms rather than one simple cause. Some reduce the sting of action over time; others preserve the appeal of the path not taken or make omissions easier to retrieve later.

Five studies did not amount to a lifetime diary

The 1994 paper reported five studies. It combined telephone surveys, written questionnaires, face-to-face interviews and hypothetical scenarios. That variety was useful because the same broad pattern appeared through more than one method.

It also sets an important boundary. The researchers did not follow the same people and the same regrets over several decades. Some studies compared imagined short-term and long-term reactions. Others asked participants to retrieve regrets from different periods of their lives.

In the original Study 5, 32 adults considered their biggest regret of action and inaction from the previous week and across their lives. For recent regrets, 53 per cent said an action troubled them more and 47 per cent chose an inaction. For lifetime regrets, 84 per cent chose an inaction and 16 per cent an action.

That 84-to-16 split is memorable, but it came from 32 people. It should never have been treated as a population constant.

The 2022 replication supported three of four paradigms

Siu Kit Yeung and Gilad Feldman revisited the work in a 2022 preregistered replication in Collabra: Psychology. Their online sample comprised 988 US participants recruited through Amazon Mechanical Turk. They replicated Studies 1, 3, 4 and 5 from the original paper in one data collection, with randomised display order.

Studies 1, 3 and 4 supported the temporal pattern. Depending on the design, participants associated relatively more regret with action in the short term and more with inaction in the long term. The effect sizes were weaker than those reported in 1994.

Study 5 did not replicate. In the task based on personal recent and lifetime regrets, the action-inaction split barely changed across the two time horizons. The authors concluded that the series gave overall support to the effect, while the inconsistency around real-life personal regrets needed further work.

That is a more informative result than a clean “replicated” or “failed” label. Three related demonstrations returned. The most autobiographical one did not.

A 2,600-person field study found a partial replication

Thomas Gilovich, Shai Davidai and colleagues then returned specifically to Study 5. Their results-blind, in-principle accepted study was published in Royal Society Open Science in 2023. It used regret responses written by visitors to the Griffin Museum of Science and Industry in Chicago.

The sample contained 2,600 usable responses, more than 80 times the original Study 5 sample. Among people prompted to consider their biggest regret over the past few days, 58 per cent chose an action and 42 per cent chose an inaction. Among those asked about their life to that point, the split was 51 per cent action and 49 per cent inaction.

The interaction between time horizon and regret type was statistically significant. In plain terms, action mattered more in recent regret than it did in lifetime regret. But the lifetime group did not favour inaction. Its 49 per cent was statistically indistinguishable from half.

The researchers therefore called the result a partial replication. It recovered the temporal shift, not the original paper’s large long-term inaction majority.

Memory may be doing part of the work

Action and inaction are not the only things changing between last week and a lifetime. The structure of memory changes too.

Ian Davison and Aidan Feeney examined regret as autobiographical memory in two studies published in Cognitive Psychology in 2008. In samples of adults in their forties and sixties, recent regrets were more likely to concern specific events. More distant regrets were more likely to be general, and general regrets were more likely to concern inaction.

That offers a different reading of the temporal pattern. A recent action can be recalled as one bounded episode: I sent the message, took the job or made the purchase. A long-term omission can become a broad theme: I never learned, never moved or never asked. General life themes may be easier to preserve across decades than the details of a particular decision.

Other work has also separated frequency from intensity. A 1999 series by Jack Feldman, John Miyamoto and Elizabeth Loftus found more long-term inaction regrets when participants listed regrets, but did not find that inaction regrets were necessarily more intense. “Which kind comes to mind more often?” and “Which hurts more?” are different questions.

This is not a rule to take every chance

I think the easiest mistake is to turn the research into advice to say yes more often. The studies do not show that action is inherently better, that caution is cowardice or that every declined opportunity becomes a lifelong wound.

Some actions produce irreversible damage. Some opportunities are poor. An inaction can reflect sound judgement, limited resources, conflicting responsibilities or information that was available at the time but fades from memory later. Regret is a retrospective feeling, not proof that the alternative would have worked out.

The word “dwell” also needs care. These experiments generally asked which regret was stronger, more troubling or easier to retrieve. They did not continuously measure how many hours people spent thinking about an omission over an entire life.

What has held up is a relative tendency: recent regret leans more towards action than distant regret does. Whether missed chances actually become the majority, and whether they hurt more intensely, changes with the sample, task and definition of regret.