Zeigarnik Effect

011 min

Where the Zeigarnik effect shows up

An illustrative case, using a role rather than a person. A backend engineer is halfway through a code change when an incident page arrives. They spend two hours on the incident. That evening, the half-finished change keeps coming back to their mind: the failing test, the function they meant to rename. The incident was more serious, but it is the unfinished change that returns.

Change one input and run it again. Before leaving for the incident, the engineer writes one line in the ticket: "Next: rename the parser function, then run the integration tests after lunch." The research on plan making, described below, predicts fewer returning thoughts in this second case. That is a prediction from a laboratory study. Nobody measured it in this scenario.

The first run mixes three separate claims. The first claim is that an interrupted task returns to mind. The second is that the person wants to go back and finish it. The third is that the person would later recall the details of the interrupted task better than the details of a finished one. The first two have support. The third is the Zeigarnik effect in its original form, and it is the one that has not held up. The sections below separate them.

022 min

Where the Zeigarnik effect comes from

Bluma Zeigarnik was one of Kurt Lewin's first students. She published the study in 1927, in the series on action and emotion psychology that Lewin edited. According to an account that is often repeated, Lewin noticed that a waiter remembered unpaid orders in detail and could not recall the orders after they were paid. This entry treats that story as an anecdote, not as evidence.

The experiment was larger than its summary suggests. According to a later overview, Zeigarnik tested 164 people one at a time, including students, teachers and children, plus a group session with 47 adults and 45 school children aged 13 to 14. Each person received 18 to 20 small tasks, such as moulding an animal from modelling clay, drawing a page full of crosses, or solving a puzzle. Most took three to five minutes. Half of the tasks were interrupted, in an order the person could not predict, and which tasks were interrupted was varied across people. Right after the last task, the person was asked which tasks they remembered.

Zeigarnik's own report was strongly positive. In the group experiment, interrupted tasks were recalled 90 per cent better by adults and 110 per cent better by children. The meta-analysis described later in this entry describes the study's design this way.

A year later, Maria Ovsiankina, a colleague of Zeigarnik, used a similar method to test a different outcome: whether people go back to an interrupted task when given the chance. That second effect is the one that has been easier to confirm.

032 min

Why the Zeigarnik effect was proposed

Lewin's theory came before the experiment. He proposed in 1926 that forming an intention creates a tension, which he called a quasi-need. A quasi-need is a pressure to act that comes from an intention, not from a biological need such as hunger. According to the 2025 meta-analysis, this is how the theory describes it.

When a task is finished, the tension is released. When a task is interrupted, the tension stays. The theory then makes two predictions: the interrupted task should be easier to recall, and the person should be drawn back toward it. Those two predictions became the Zeigarnik effect and the Ovsiankina effect.

The tension itself was never measured in these studies. It is a theoretical term, and the later evidence supports only some of what follows from it.

A more recent line of work describes the same idea as goal activation. In a 2011 paper, Masicampo and Baumeister reported that an unfinished goal led to intrusive thoughts during an unrelated reading task. It also made goal-related words easier to access, and in one study it lowered performance on an unrelated anagram task. These are three observable effects of one unfinished goal, and the paper's authors present them as evidence that an unfulfilled goal stays active in the mind.

Why would the mind work this way? One plausible reason is that keeping an unfinished goal active makes it more likely that the person returns to it. That reason is a hypothesis, and the studies discussed here did not test it.

A related question is which mental resource the active goal uses. Working memory is the limited capacity to hold and use information for a few seconds. Masicampo and Baumeister write about freeing "cognitive resources", but whether goal activation uses working memory in particular was not measured in the studies this entry reviewed.

042 min

Where the Zeigarnik effect is contested

The memory advantage is the part of the Zeigarnik effect that has failed to replicate. Several later studies, among them an in-depth replication by Van Bergen in 1968, found no significant difference in recall between finished and interrupted tasks.

The strongest evidence comes from the 2025 meta-analysis by Romain Ghibellini and Beat Meier, published in Humanities and Social Sciences Communications. It combined studies that used the same interrupted-task method as Zeigarnik and Ovsiankina.

For recall, the authors report that interrupted tasks made up an average of 49.16% of all tasks recalled. If interruption helped memory, that figure would be well above 50%. They report a weighted ratio of recalled interrupted tasks to recalled finished tasks of 0.99, and a small average effect size of 0.15. They conclude that Zeigarnik's own figures were inflated.

The authors also offer an explanation for why the original result may have appeared. In the 1920s, a professor running an experiment held more authority than a researcher does today, and participants may have worked harder to do well. The authors suggest that this task involvement may be needed to produce the effect, and that it is rarer now. They present this as a possibility, not as a tested cause.

A fair objection to the meta-analysis is that it only covers the interrupted-task method, so it says little about other ways of measuring unfinished goals. Masicampo and Baumeister, for example, did not use recall. A second objection is that the pooled studies differ in task type and in how participants were treated.

For a practitioner, the position is this. Do not rely on the claim that unfinished work is remembered better. Treat the tendency to return to unfinished work, and the benefit of planning, as better supported, with the caveats in the next section.

052 min

What still has support: resuming and planning

The same meta-analysis reports a clearer result for the Ovsiankina effect.

In 20 publications on resumption, interrupted tasks were resumed about 67% of the time. This figure leaves out Ovsiankina's original data. Chance would be 50%. The authors say this rate should be read as indicative, not as an exact figure, because the studies used different ways of interrupting and of offering the chance to resume.

The second line of evidence is plan making. Masicampo and Baumeister published it in 2011, in the Journal of Personality and Social Psychology. Their finding is the most useful one for this audience, because it turns the effect into something a person can act on. They asked participants to make a specific plan for an unfulfilled goal, and the intrusions and the other effects went away, even though the goal was still unfinished. Participants who later carried out their plans were also the ones who showed no more intrusions.

A press summary from the author's university describes the same study.

Two caveats apply. The first is that this entry did not find an independent replication of the plan-making result, so it rests on one research group. The second is that the paper's own wording is about specific plans. Writing a goal on a list is not the same thing.

These two cases differ in one variable: what the researchers measured. The first case measured recall of tasks, and the effect was near zero. The second measured intrusive thoughts and the accessibility of goal words, and the effect was clear. The contrast suggests that unfinished goals may stay active without the person remembering the task itself any better.

062 min

How the Zeigarnik effect shows up in product, design, and AI

Accidentally suffered: interruptions in engineering work. The opening scenario applies to any task switch: an incident, a review request, a meeting. The research supports one prediction here. If a person leaves a task without a plan for it, thoughts about it may return later. A short note on the next step is the cheapest response, and it does not depend on the contested memory effect.

Deliberately used: resume prompts in product design. Many products show an unfinished state, such as a checklist that is partly done, a draft that can be continued, or a "pick up where you left off" button. The better-supported ground for these designs is the Ovsiankina effect, the tendency to resume an interrupted task. The memory claim is not a good ground. This entry reviewed no sourced case of a named product, so it makes no claim about how well any particular design works.

The line between a prompt and a dark pattern. A prompt that lets a person resume work they chose to start is a convenience. A repeated reminder that makes leaving feel like a failure is pressure. The research reviewed here does not say where one becomes the other, so this is a product decision, not a finding.

For analysts. When a team cites "the Zeigarnik effect" to justify a feature, ask which claim it is using. A feature meant to bring people back should be judged on the rate of return, which is the Ovsiankina measure, and an A/B test can measure that directly. A feature justified by better recall of the interrupted step has the weaker support and should be tested before anyone relies on it.

071 min

How to use the Zeigarnik effect, and what not to use it for

Awareness does not help much here, and nothing in the sources reviewed shows that it does. What the evidence supports is a change in procedure.

Write the next step before you stop. A next action is the one concrete step to take when work resumes. The 2011 plan-making study suggests that making the plan, and not only intending to do it, is what reduces intrusive thoughts. The author of that study said that the plans that work "specify exactly what you are going to do, including when and where you are going to do it."

Make the plan specific. "Finish the report" restates the goal. "Tomorrow at the start of the day, write the summary table from the usage data" is a plan. The studies tested specific plans, so a vague note should not be expected to do the same.

Do not use interruption as a memory tool. The meta-analysis found no reliable recall advantage, so stopping a study session in the middle of a task is not a tested way to remember it better.

Use the pull to resume deliberately. If a task is easy to start again, a visible unfinished state can help a person return to it. If the work needs a break, a stopping point with a note costs less than a person carrying the task around in their head.

081 min

Common misunderstandings

"People always remember unfinished tasks better." This is the popular version of the effect, and the 2025 meta-analysis does not support it. The pooled ratio of recalled interrupted tasks to recalled finished tasks was 0.99.

"The effect explains why cliffhangers and open-ended games keep people engaged." This is a common claim in popular writing, and none of the sources reviewed for this entry tested it. The better-supported effect is the tendency to resume, and a cliffhanger tests that tendency, not memory.

"A to-do list gets unfinished tasks out of your head." The plan-making study used specific plans, not lists of goals. A list of names of tasks may not do the same job.

"Unfinished work always causes distress." The studies reviewed here measured intrusive thoughts, word accessibility and anagram performance. They did not measure stress, so they do not support a claim that unfinished tasks make people anxious.

Where the pull to resume is useful. A person who returns to a half-finished draft is doing what the theory predicts, and nothing needs to be guarded against. The problem is only when the pull interrupts other work with no plan to resolve it.

091 min

The Zeigarnik effect vs. nearby concepts

The fast test is this: was the task stopped before it was finished, and is the claim about what the person recalls? If so, it is the Zeigarnik effect. If the claim is about returning to the task, it is the Ovsiankina effect.

?8 questions

Questions people ask

§6 sources

Sources