Summary for Educators: Forgetting Takes No Holidays

Recent neuroscience research reexamines the classic forgetting curve first mapped by Hermann Ebbinghaus, showing that even deeply encoded memories decline over time—and the rate of that forgetting does not differ much between deep vs. shallow encoding.

In a set of experiments, scholars from the University of Edinburgh had participants learn words in two different contexts: a shallow encoding task (e.g. judging whether a word rhymes) and a deep encoding task (e.g. thinking about how the word fits a sentence or meaning). On immediate recall tests, the deep encoding group outperformed the shallow group—demonstrating the value of semantic processing. That performance advantage persisted at longer delays: 30 minutes, 2 hours, and 24 hours.

Crucially, though, the rate at which participants forgot information (i.e. the decline in recall over time) was nearly parallel for both groups. In other words, deeper encoding doesn’t slow down forgetting—it shifts the starting point higher. This supports the conclusion that forgetting may act as a default, somewhat inevitable process of the brain (Rivera-Lares et al., 2023). Edinburgh Research+1

Thus, even if students encode information deeply (through analysis, connection, elaboration), that doesn’t protect them from losing detail over time. The positive news is that deeper knowledge maintains a persistent advantage: such learners start further ahead and remain ahead, even as decay occurs.

From an educational perspective, this means that repetition, retrieval practice, spaced review, and scaffolding are still essential—because no matter how well material was initially learned, the brain naturally sheds traces over time. It also argues for designing instruction that doesn’t assume that deep learning is “done,” but plans for reinforcement.

Some practical takeaways for educators:

• Build spaced retrieval into curriculum—plan revisits, cumulative review, interleaving.

• Use elaborative encoding strategies (concept maps, student explanations, analogies) to increase how deeply material is encoded (raising the starting point).

• Recognize forgetting is natural: don’t treat it as failure, but as a signal to refresh.

• Monitor retention over intervals (days, weeks) rather than only immediate recall.

• Consider scaffolding supports for students who may struggle to encode deeply or to sustain retrieval practice.

In sum, this study reaffirms the power of deep learning but also warns educators that even our best‐learned content will fade without ongoing engagement. Knowing that forgetting takes no holiday helps us design teaching that anticipates memory decay rather than denying it.

Original Article

Peng, N., Logie, R.H. & Della Sala, S. Effect of levels-of-processing on rates of forgetting. Mem Cogn 53, 692–709 (2025). https://doi.org/10.3758/s13421-024-01599-4

------------------------------

Prepared with the assistance of AI software

OpenAI. (2025). ChatGPT (4) [Large language model]. https://chat.openai.com