Worried that what you studied last week is already gone? The forgetting curve describes how memory for learned material changes as time passes, and a modern replication supports the classic picture. Plan to relearn material, not just hope it sticks.
This page is part of the wider guide to study techniques that work.
| Item | Value |
|---|---|
| Year of the original forgetting curve | 1880 |
| Year the replication was published | 2015 |
| Number of subjects in the replication | One subject |
| Time that subject spent learning and relearning lists | 70 hours |
| Delays before relearning | 20 min, 1 hour, 9 hours, 1 day, 2 days, 31 days |
| Possible upward jump in the curve | Starting at the 24 hour data point |
These figures come from the EconPapers record of the study and the FLoRA Replication Atlas entry. Figures can change as records are updated, so check the sources before you rely on them.
What the forgetting curve describes
The forgetting curve is a plot of how much of something you can still use after a delay. The classic version dates from 1880, and the replication abstract on EconPapers credits it to Ebbinghaus.
That original work used the method of savings. In plain terms, a learner studies lists, waits, and then relearns them, so the data come from relearning rather than from simply recalling.
The idea matters to anyone who studies, because it turns a vague feeling of “I knew this” into something you can measure and plan around.
What the 2015 replication did
A paper by Murre and Dros, published in PLOS ONE in 2015, set out to repeat Ebbinghaus’s work. According to the FLoRA Replication Atlas, the paper is titled “Replication and Analysis of Ebbinghaus’ Forgetting Curve”.
As the abstract describes it, one subject spent 70 hours learning lists and relearning them after 20 min, 1 hour, 9 hours, 1 day, 2 days, or 31 days. The authors report that the results are similar to Ebbinghaus’ original data.

The authors also looked at serial position, meaning how forgetting differs across a list. They tested which mathematical equations fit the curve and its replications.
A possible jump at the one-day mark
Many people picture the curve as a smooth slide downward. The authors of the replication conclude otherwise: in the abstract, they say it is not completely smooth but most probably shows a jump upwards starting at the 24 hour data point.
The wording “most probably” is a hedge, and it is worth keeping. It means the shape near that point is a finding to be weighed, not a rule to memorize.
How to use the curve in your own study plan
The replication measured forgetting and did not test remedies, so treat the steps below as planning ideas drawn from the method, not proven results.
- Pick a defined set of material, such as a vocabulary list or a set of definitions.
- Study it, then leave a gap before you return and relearn it.
- Note how much came back easily and how much you had to rebuild.
- Vary the gaps, from short to long, so you see how your own memory behaves.
If you want a way to restate material during a review, the feynman technique is one option. Structured notes also give you something to return to, and the cornell note taking method is a common format for that.
What the evidence does not show
The replication rests on one subject, so it shows that a classic result can be reproduced, not how every learner will behave. Your own results will depend on the material and on you, and no method here guarantees a grade.
The FLoRA Replication Atlas says it has no replication or reproduction of this study on record yet. It adds that an absent record is not evidence that the finding failed to replicate.
Frequently asked questions
Has Ebbinghaus’s forgetting curve been confirmed?
Show answer
The Murre and Dros abstract calls its replication successful and says the data resemble the original. That is one study of one learner, so read it as support rather than as the final word.
How long were the gaps the replication tested?
Show answer
They ran from a short break of 20 min up to 31 days, according to the study abstract. Four other delays fell between those ends.
Is the curve a perfectly smooth line?
Show answer
Not according to the authors. They say it probably rises again around the 24 hour data point, as stated in the same abstract.
Does this research say which study method works best?
Show answer
No. It documents how forgetting unfolds over time and does not compare study techniques, so any method you choose should be tested against your own results.



