Why I still make students keep a practice journal, in an age of automatic tracking

A student came in last Thursday with a perfect week. His app said 212 minutes across five days, five tidy green bars. Then he played me the piece, and bar nine fell apart in exactly the same place, in exactly the same way, as it had the Thursday before. Two hundred and twelve minutes, and the hardest four beats in the piece had not moved.

Automatic trackers log minutes; they can't tell whether a student spent twenty of them staring at a wall. Research on metacognition and self-regulated learning — a 2025 meta-analysis found a medium, significant effect, g = 0.563 — shows that reflecting on what actually happened during practice, not simply how long it lasted, is what predicts improvement. A journal forces that reflection. A timer never asks you to think.

So I still hand out notebooks. Not out of sentiment, and not because I distrust the software — I build some of it. Because the two instruments measure different things, and only one of them measures the thing that changes players.

What "deliberate practice" actually claims, and where it is under fire

The obvious move here would be to reach for the 10,000-hours literature and let it do the work. I want to resist that, because the ground is less solid than the popular version suggests, and building an argument on it would be dishonest.

The 1993 framework from Ericsson, Krampe and Tesch-Römer defined deliberate practice as effortful, goal-directed activity carried out with immediate feedback and repeated refinement. Notice the second half of that definition — it is the half nobody quotes. Their account of expertise includes performers gradually building the capacity to monitor and evaluate their own playing, becoming progressively less dependent on a teacher standing there telling them what went wrong. That is a claim about self-assessment, not about hours. They did not use the word "journal" and I am not going to put it in their mouths. But the mechanism they describe is exactly the one a journal exercises.

Then, in 2019, Macnamara and Maitra went back to the original violinist data and found something awkward. Accumulated deliberate practice explained only about 26% of the variance in performance level among elite violinists — well below the original claim. More pointedly, the most accomplished violinists in the sample had logged less practice than the merely good ones.

That finding is contested. Ericsson's side disputed the reanalysis, and the argument is genuinely live rather than settled; anyone telling you deliberate practice has been "debunked" is overselling as badly as the people who told you it was a law of nature. But you do not need to pick a winner to draw the useful conclusion: hours accumulated is a much weaker signal than the folklore assumes.

Something similar turns up in a 2024 survey of advanced Brazilian and Portuguese musicians, which identified three separate dimensions in how they organise practice. The more experienced professionals scored substantially higher on personal metacognitive resources and lower on reliance on external resources than the less experienced players — while typically practising fewer total hours than students. Experience appears to buy efficiency, not volume.

Which is a problem for tracking software, because volume is precisely what tracking software is good at counting.

The session that isn't like the others

Every teacher I know has watched a student make more progress in one five-minute stretch than in the preceding fortnight. It is the single most familiar phenomenon in the studio and the hardest to explain to a parent looking at a chart.

The academic source usually pointed at here is McPherson and Renwick's 2001 longitudinal study of self-regulation in children's practice, which followed young instrumentalists over years and looked closely at what they actually did when nobody was watching. The version most people have encountered is the retelling by science writer Daniel Coyle, with its striking figures about one child learning many times faster in one session than another. I would treat those specific narrative details as popular-science framing rather than quotable data from the paper — but the underlying observation, that learning rate tracks the quality of engagement with mistakes rather than elapsed time, is the sort of thing the self-regulation literature keeps producing.

You can see the same pattern in Hallam and colleagues' longitudinal work on how young people's practice strategies develop. As expertise increased, systematic strategies rose — slow practice, sectional repetition, deliberate error-recognition — along with the use of recording devices and metronomes. Ineffective strategies fell away: playing straight through, and the great universal time-waster, going back to the beginning every time you fumble something in the middle.

Read that list again with a timer in mind. A student playing straight through a piece four times and a student isolating one bar and rebuilding it at 60 bpm produce identical logs. Sixteen minutes each. The metronome has been the subject of two hundred years of arguing about tempo and it still cannot tell you whether you deserved the click. Nothing in the automatic data distinguishes the two sessions, because the difference between them is not a quantity. It is a decision, and decisions leave no trace unless somebody writes them down.

Writing forces a different kind of attention

Here is where the evidence gets genuinely good, and notably better than the evidence for raw hours.

The 2025 systematic review in the Journal of Intelligence pulled together 31 studies of metacognition and self-regulated-learning instruction in music from 2015 to 2025, seven of them pooled for meta-analysis. The result was a statistically significant, medium-sized positive effect on learners' metacognitive and self-regulatory ability, Hedges' g = 0.563, with effects strongest at secondary and tertiary level. I want to be careful about what that number licenses: those interventions bundle reflection with teacher scaffolding, structured support and technology feedback. It is not a measurement of notebooks. It is a measurement of the family of approaches that notebooks belong to.

What makes it convincing is a 2024 study comparing beginner violin students with professional brass-quintet musicians. The professionals showed greater reflective capacity: they could evaluate whether a practice strategy had actually worked. The beginners could describe their own behaviour perfectly well — they knew what they had done — but did not always know how to judge whether the strategies they used had been effective. Beginners also found organising and time-managing their sessions significantly harder.

That gap is the whole argument, in one finding. Beginners are not short of data about themselves. They are short of the habit of evaluating it. And a tracker, by design, supplies more data and demands no evaluation whatsoever.

There is also the 100-week case study of a professional concert cellist — a long, close look at one player who wrote regularly about her actions, thoughts and feelings after practising. Over the two years, the researchers observed increased intrinsic motivation, decreasing reliance on external regulation, and steadily more sophisticated practice strategies. One case, so treat it as illustration rather than proof; but it is illustration at a level of detail that group studies never reach.

And my favourite, because of how physical it is: a review by Jabusch describes a study by Lisboa and colleagues in which an eighteen-year-old piano student with no memorisation experience wrote her playing-related thoughts directly onto her score over six and a half weeks. Those written thoughts became stable performance cues. Nine and a half weeks later, with no practice in between, she performed the piece from memory. The writing had become part of the four kinds of musical memory she could actually retrieve under pressure.

There is one more reason to distrust unexamined self-knowledge. When researchers compared teachers' own estimates of time spent on music activities against video of the same lessons, the self-reported figures came out higher than what the recording showed. That 1997 study established the direction only, not a magnitude — I have seen people invent a percentage for this and there isn't one. But the direction alone is worth sitting with: we overestimate, and we do it without meaning to.

What automatic logging is genuinely good for

I am not making an anti-technology argument, which would be a strange position for someone who spends his evenings building a practice tracker.

Crescender logs practice time automatically, and that is real value. Parents get an honest picture without policing anyone. Students exam-bound in term three can see whether the schedule they promised themselves survived contact with Year 11. Teachers can spot the fortnight where everything quietly stopped, before it becomes a conversation about whether to continue lessons. Duration is a legitimate measurement and automation removes the friction of capturing it.

It just answers a different question. So I still ask for two or three handwritten lines about what happened in the session — because no timer ever built was designed to make anyone think, and mine is no exception.

The counter-argument deserves better than a shrug

Journals get abandoned. Anyone who says otherwise has not taught long enough. The notebook handed out in February turns up under a bed in November with nine entries in it, and that is the ordinary outcome rather than the exception.

The strongest evidence I can point to here is a 2024 scoping review in the Journal of Medical Internet Research on why adults abandon lifestyle-behaviour and mental-health apps. It found a curvilinear pattern — sharp drop-off soon after download, then a slowing tail — with roughly 70% of users discontinuing within the first 100 days. Retention varied enormously by category: alcohol-reduction apps saw 95–97% abandonment, while physical-activity apps held on to between 54% and 75% of users. Crucially, "time to enter data" was named specifically as a driver of quitting.

I should be straight about what that is and isn't. It is evidence about general habit-tracking and health apps, not a study of music practice journals — nobody, as far as I can find, has run that study. Extrapolating is reasonable, because the mechanism being described is the same one — self-report decays when the reporting costs more than it returns — but it is extrapolation and I am not going to dress it up as a finding about notebooks.

What it changes is the design, not the conclusion. If data-entry burden is the killer, the answer is a smaller entry. I ask for three things: what you worked on, in bar numbers rather than piece titles; what went wrong, specifically; and what you will try first next time. Thirty seconds. No mood ratings, no star scores, no colouring in — those feel like reflection and are actually just more logging.

The version that dies is the one that asks a fourteen-year-old to write a paragraph. The version that survives is three lines in pencil, in the case, next to the rosin.

This isn't nostalgia

It would be easy to read all this as an older teacher preferring paper. It isn't, and the institutions give the game away.

The Australian Music Examinations Board — the body most Australian students encounter when choosing between AMEB, ABRSM and Trinity — sells its own branded Practice Diary as a study resource. It runs to a 26-week lesson-and-practice record, plus a glossary of musical terms, a composer timeline and a key-signatures guide, and it is marketed straight at private studios and classrooms. AMEB does not require anyone to use it, and I am not claiming they do. But an exam board in 2026 is still manufacturing a physical diary, which tells you something about what teachers keep asking for.

At the other end of the training pipeline, Carey, Harrison and Dwyer studied first-year conservatoire students at Queensland Conservatorium, Griffith University who kept reflective journals documenting their one-to-one performance lessons. The journalling process prompted increased reflection, and the researchers found it associated with growing autonomy, a sense of shared responsibility with teachers, and greater clarity about their own learning. These are students who already practise several hours a day and already know how to work. They were still being asked to write it down.

Between an exam board's shop and a conservatoire's first-year curriculum, the written practice record has survived every generation of technology that was supposed to make it redundant. That is not sentimentality. That is a method that keeps outliving its obituaries.

Bar nine, six weeks later

The student with the 212 minutes is back on the chair by the window, the one where the afternoon light comes in sideways and the 19 tram goes past every eight minutes and nobody has ever managed to record anything usable.

His notebook is open on the stand. Last Tuesday's entry, in pencil, reads: bar 9 — shift lands late every single time above 80. Tried it at 60 tomorrow, just the shift, nothing either side. Underneath, Wednesday: worked at 60. Late again at 84. It's the left thumb — I'm dragging it.

He plays bar nine. It is clean, at tempo, twice.

Then he tells me why, unprompted, before I have opened my mouth — and that is the part I actually wanted. Not the four beats. The fact that he can now describe the mechanism of his own mistake, in his own words, in a sentence he wrote himself before he knew it was true. No timer in the world was ever going to hand him that.

Common questions

Do practice journals actually help musicians improve faster than apps that just log time?

The honest answer is that no study has isolated journalling alone with a percentage attached. What exists is strong: a 2025 meta-analysis of metacognition and self-regulated-learning instruction in music found a medium, significant effect (Hedges' g = 0.563). Those interventions bundle reflection with teacher support and structure, so credit the reflective approach, not the notebook itself.

Why do music teachers still use practice journals when automatic tracking apps exist?

Because the two measure different things. An app records that practice happened and for how long; it cannot record what the student attended to, what went wrong, or what they tried next. Research comparing beginners with professionals found beginners could describe their own behaviour but struggled to evaluate whether a strategy had worked. Writing is where that evaluation gets built.

Is writing down what happened in practice better than tracking how long you practised?

They answer different questions, but duration is the weaker predictor. A 2019 reanalysis of elite violinist data found accumulated practice explained only about 26% of performance variance, and the very best players had logged fewer hours than the merely good ones. Time tells you effort was spent. Reflection tells you whether it was spent well.

Do practice journals get abandoned — is a written log worth the friction?

Abandonment is real. A 2024 review of lifestyle and mental-health apps found roughly 70% of users stopped within 100 days, with data-entry burden a named driver. That is general habit-tracking evidence, not a study of practice journals, but the lesson transfers: keep entries to two or three lines, or they will not survive the term.

What should go in a music practice journal entry?

Three things, in about thirty seconds. What you actually worked on — the bar numbers, not the piece title. What went wrong, specifically enough that you could describe it to someone else. And what you will try first next time. Skip mood ratings and star scores; they feel productive and tell you nothing you can act on.

If you are setting one up alongside exam preparation, the same three-line discipline works for helping a child prepare for a music exam, and it transfers cleanly to listening work too — ear training from zero improves fastest when you record what you got wrong rather than how many you attempted.