Evidence

The memory research in this area is genuinely strong. It is also routinely cited for things it does not support, and the gap between the two is most of what this section is about.

The spacing effect is real and not uniform. Distributed study beats massed study for durable retention, replicated thousands of times. One meta-analysis found benefits ranging from about g = 0.29 to g = 2.16 with significant heterogeneity; a later analysis in mathematics found smaller effects, around g = 0.24 within a course. The range is the honest headline, not the mean.

A related workplace concept is employee attendance tracking software, which offers a useful contrast with the personal knowledge-work problem discussed here.

The testing effect holds broadly and has a documented exception. Attempting recall beats restudying, failed attempts still help, and there is a replicated forward effect on subsequent learning. In mathematics, testing versus restudy came out at g = 0.18 with a confidence interval crossing zero — which means no effect was reliably detected, not that a small one was found.

The forgetting curve measures something other than its usual axis label. Ebbinghaus measured savings — reduction in relearning time — not percentage recalled, on nonsense syllables chosen specifically to have no meaning, with a sample size of one. It replicated closely in 2015, and the replication found the curve is not smooth.

And the most-cited note-taking study did not replicate. The 2014 longhand-beats-laptop finding failed to reproduce in two direct replications. What survived is the distinction it popularised: encoding, from the act of writing, and external storage, from having the record.

For an independent external reference related to this topic, see PubMed Central.

Underneath all of it sits an uncomfortable fact: almost nothing specific to personal note archives has been studied at all. The research is about laboratory materials and deliberate learning, and the bridge to four thousand notes on your own disk is built out of reasoning rather than evidence — including here.

• Does Note-Taking Help Memory?

  The most-cited study in this field did not replicate. What survived it is a distinction more useful than the finding ever was.

• Effect Sizes in Learning Research

  A number that looks like precision and usually is not. What g means, what counts as large, and why the range matters more than the average.

• What Ebbinghaus Actually Showed

  The most reproduced graph in learning writing measures something other than what its axis label says, on the least meaningful material possible.

• Reading a Claim About Learning

  Six checks that take two minutes and catch most of what is wrong with productivity writing that cites cognitive science.

• Second-Brain Claims

  Five specific promises, checked one at a time. Two are supported, one is plausible and unstudied, and two are not claims about memory at all.

• The Spacing Effect, Carefully

  One of the most replicated findings in psychology, and one of the most misapplied. What it shows, and what it does not.

• The Testing Effect

  Retrieval beats rereading — one of the strongest findings in learning research, and one with a documented exception worth knowing.

• What We Still Don't Know

  Six open questions that matter more than the answered ones, and an account of why the evidence base has the shape it has.