The Zeigarnik Effect: Why Unfinished Tasks Stay in Your Head

Unfinished tasks do keep occupying mental space, but almost certainly not through the mechanism the famous version of this effect describes. The claim that you remember interrupted tasks better than completed ones has a poor replication record and was recently pooled across 37 published studies with a result of essentially zero. What survives is a narrower and more useful finding: unfinished goals generate intrusive thoughts, and the thing that shuts those thoughts off is making a specific plan, not completing the task.

What Bluma Zeigarnik reported in 1927

Bluma Zeigarnik was a graduate student in Kurt Lewin's group in Berlin. The origin story, which is repeated in nearly every popular account, is that Lewin noticed a Viennese waiter who could recall unpaid orders in detail and then forgot them the moment the bill was settled.

Zeigarnik built an experiment around the observation. Participants worked through a sequence of short tasks, roughly half of which were interrupted before completion, and were later asked to recall as many as they could. She reported that interrupted tasks came back close to twice as often as finished ones. The theoretical frame came from Lewin: an intention creates a tension state, a quasi-need, and that tension only discharges on completion.

It is an elegant story, and it has been repeated so consistently for ninety-odd years that it now functions as a fact. The evidence underneath it is weaker than the repetition suggests, which is a pattern this cluster keeps running into, most visibly in the distance between what goal-setting theory is popularly said to prove and what its own researchers documented.

The replication record, plainly

The most thorough attempt to reproduce the original was a 1968 doctoral monograph by A. van Bergen at the University of Amsterdam, published as Task Interruption. It was a serious, in-depth replication, and it did not find the memory advantage.

The picture since then has been inconsistent enough that in 2025 Romain Ghibellini and Beat Meier at the University of Bern pooled the whole literature. Their meta-analysis, published in Humanities and Social Sciences Communications, screened 59 publications using the interrupted-task paradigm. The headline numbers are worth stating exactly:

• Across 37 publications excluding Zeigarnik's own, the weighted ratio of recalled interrupted tasks to recalled completed tasks came out at 0.99. That is not a small effect. That is no effect.
• Interrupted tasks accounted for an average of 49.16 percent of everything recalled, which is what you would expect if interruption did nothing at all.
• Where an effect size could be computed, the average was dz = 0.15, small enough to be indistinguishable from noise given the study sizes involved.
• The authors conclude that the Zeigarnik effect "lacks universal validity."

The same meta-analysis found that a related and much less famous effect held up well. The Ovsiankina effect, named for Maria Ovsiankina's 1928 work, is the tendency to resume an interrupted task when given the chance. Pooled across 20 publications excluding the original, the resumption rate was 66.79 percent. People do go back to unfinished things. They just do not remember them better.

The version that does hold up

There is a second reading of the Zeigarnik idea that has nothing to do with recall accuracy. It says unfinished goals stay cognitively accessible: they intrude on unrelated thinking, they pull attention, and they degrade performance on tasks they have nothing to do with. That version has direct experimental support, and it comes with an unexpected finding attached.

In 2011, E. J. Masicampo and Roy Baumeister published a set of studies in the Journal of Personality and Social Psychology (volume 101, issue 4, pages 667 to 683) under the title "Consider it done! Plan making can eliminate the cognitive effects of unfulfilled goals." They confirmed the interference: participants primed with an unfulfilled goal showed more goal-related intrusions while reading, faster access to goal-related concepts, and worse performance on unrelated tasks.

Then they added a condition. Participants wrote a specific plan for when and where they would deal with the unfinished goal. The goal remained unfinished. The interference disappeared anyway. The authors also found that the plan had to be made in earnest to work, which rules out the possibility that writing anything down is simply a distraction.

This is the genuinely useful result. The open loop is not closed by finishing. It is closed by specifying. Once a goal has a designated situation and response, the cognitive system apparently treats it as handled and stops raising it.

That is the same structure documented in the research on implementation intentions and if-then planning, where naming a cue and a response outperforms restating the goal. Masicampo and Baumeister's finding is arguably the cleanest demonstration of why: the plan does not just make the behavior more likely, it also stops the unexecuted goal from taxing you in the meantime.

Two caveats that belong on this finding

First, provenance. Baumeister's most widely cited claim, ego depletion, held that self-control draws on a limited resource that empties with use. That claim failed a 23-laboratory preregistered replication published in Perspectives on Psychological Science in 2016, which found essentially no effect across 2,141 participants. The plan-making result does not depend on ego depletion being true, and it is not the same claim. But it came out of the same laboratory tradition and the same era of methods, and it deserves the same scrutiny.

Second, the 2011 studies were modest laboratory experiments of the kind that era produced in volume. As far as I can find, they have not been subjected to a large preregistered replication. Treat the finding as promising and mechanistically sensible rather than settled.

What to do with an open loop, ranked by evidence

1. Write a plan with a situation attached. "If it is Thursday after dinner, then I will call the pharmacy" is the form Masicampo and Baumeister tested. A to-do list entry that says "call pharmacy" is not this, because it names no moment.
2. Anchor the plan to an event, not a clock time. Cue reliability is the constraint on every if-then plan. An event you already perform without deciding to is worth more than an hour you intend to respect.
3. Accept that you will resume rather than remember. The Ovsiankina resumption effect is the part of this literature with real support. Design for returning to things, not for recalling them at the right moment.
4. Do not treat the intrusion as motivation. A goal nagging at you is a cost, not a signal that the system is working. It is exactly what the plan is supposed to remove.
5. Stop invoking the memory version. If your productivity system rests on the claim that unfinished tasks are better remembered, the pooled evidence says the foundation is not there.

The broader point is one this cluster keeps arriving at from different directions. The distance between intending something and doing it is not usually closed by wanting it more, a pattern quantified in the research on the intention-behavior gap. It is closed by specifying when and where, which costs about thirty seconds and is boring enough that almost nobody bothers.