Set Shadowing Homework That Produces Useful Recordings, Not Complet...

Use Take 1 → Name → Repair → Take 2 → Transfer

1. Take 1: record one honest baseline after the student understands the line.
2. Name: choose one observable target such as thought groups, prominence, one connected-speech feature or one intonation move.
3. Repair: practise privately. The teacher does not need every rehearsal file.
4. Take 2: record the same source again with the same target.
5. Transfer: record one new sentence or short response that uses the same feature with different wording.

The teacher then listens with one question: Did the named feature change, and did any of that change survive outside the copied line?

Replace repetition counts with evidence

“Shadow this five times” is easy to assign and easy to tick off. It is much harder to know whether repetition three, four and five had a purpose.

You can still suggest a reasonable practice window if students need structure, but do not make the number of repetitions the main evidence of learning. A better submission tells you what the learner tried to change.

For most homework, three short audio artifacts are enough:

• Take 1: the starting attempt.
• Take 2: the same source after targeted practice.
• Transfer: a new line using the same speaking feature.

That is enough to create a before/after comparison without turning your inbox into an audio archive.

Take 1: capture the baseline without turning it into a punishment

The first take should happen after the student understands the sentence, but before targeted repair. It is a baseline, not a “gotcha” recording.

Tell students that Take 1 is there so they can hear a change later. Do not grade it for polish.

If students are allowed unlimited “baseline” retakes, the first recording quietly becomes another polished final take. One reasonable first attempt gives you a clearer comparison.

Name one target the student can actually hear

British Council TeachingEnglish guidance directs attention to features such as pronunciation, rhythm, intonation and connected speech. For homework, make that attention more specific.

Good target statements sound like this:

• “I want clearer thought groups.”
• “I want the important contrast words to stand out.”
• “I want to hear and reproduce the weak form in this phrase.”
• “I want my final fall to sound more complete.”

Weak target statements sound like “improve pronunciation” or “sound more native.” They are too broad to compare fairly in two short recordings.

The target statement makes the homework inspectable. The teacher already knows what to listen for before pressing play.

Repair privately; submit selectively

Students can rehearse as much as they need within the assignment’s practical limits. They do not need to upload the whole rehearsal history.

A 2025 systematic review of shadowing research reports promising evidence in areas including fluency and some prosodic dimensions, while also emphasizing substantial variation in methods and outcomes. That is a good reason to avoid fake precision such as “seven repetitions guarantees improvement.”

Instead, give the repair phase a job:

1. Listen specifically for the named target.
2. Shadow the smallest useful chunk.
3. Compare with the model.
4. Try again only while the next repetition is fixing something identifiable.
5. Record Take 2 when the student can hear a meaningful change or has reached the limit of what they can repair alone.

Homework is allowed to contain struggle. The submission should make that struggle interpretable.

Take 2: same source, same target, cleaner evidence

For the comparison to mean anything, keep the source and target stable between Take 1 and Take 2. Do not switch sentences and then ask whether the pronunciation “improved.”

That gives the teacher a precise feedback sentence: “Your grouping is clearer in Take 2; next time keep the same grouping when the sentence gets longer.” That is more useful than “Good job — 5/5 completed.”

Transfer: change the words and see what survives

A better second take proves that the student can improve one copied line. The transfer recording asks a harder question: can the feature survive when the wording changes?

Keep the transfer short: one new sentence or brief response is usually enough to make the target feature audible without creating a second full homework assignment.

If the target disappears completely in Transfer, that is not a failed homework submission. It is excellent diagnostic information: the feature is still dependent on the practiced line.

Review the homework with three questions

The assignment should make teacher feedback faster, not create three times as much grading.

1. What was the named target?
2. What changed from Take 1 to Take 2?
3. Did the feature survive in Transfer?

Then give one next target. Avoid turning a short shadowing assignment into a full pronunciation report unless that is genuinely the purpose of your course.

A compact feedback pattern

Change: “Your contrast words are clearer in Take 2.”

Transfer: “The contrast is still audible in the new sentence, although the second word loses some prominence.”

Next: “Next time, keep the second contrast word strong even when the phrase is longer.”

Do not grade the microphone

Recordings contain technical noise as well as language. A quiet room, expensive microphone and fast device can make one file easier to judge than another. That is not the same thing as stronger English.

There is also a more specific shadowing problem: Audacity documents recording latency, where an overdubbed recording can be laid down slightly out of sync with the playback. Exact offset varies by setup.

So avoid high-stakes judgments based on millisecond-perfect overlap, waveform alignment or tiny timing offsets across different devices. Listen for the language feature you assigned.

• Accept a clean voice-only recording if mixing the model creates technical problems.
• Do not grade room noise as pronunciation.
• Do not require public posting when a private submission works.
• Provide a reasonable alternative when a learner cannot record privately or reliably at home.

Would this submission actually tell you anything?

A copy-ready shadowing homework brief

You can adapt this for your class:

1. Listen until you understand the short source passage.
2. Record Take 1 once before targeted practice.
3. Write your one target: for example, “clearer thought groups” or “stronger contrast on the key words.”
4. Practise privately. You do not need to submit rehearsal files or report a repetition count unless your teacher specifically needs that information.
5. Record Take 2 using the same source and focusing on the same target.
6. Record one Transfer sentence with different wording but the same speaking feature.
7. Submit the three recordings plus one sentence: “The clearest change I hear is ___.”

That final sentence forces a small act of self-listening before the teacher supplies the interpretation.

Micro-challenge for teachers

Open your current shadowing homework instructions. Delete any requirement that exists only to prove time-on-task — for example, a fixed repetition count with no learning reason — and replace it with one observable piece of evidence.

Four practical homework questions

How many recordings should students submit?

Usually you only need the recordings that answer a question: a baseline, a repaired version and a brief transfer. More files are useful only when each adds evidence you plan to review.

Should Take 1 be graded?

It can receive credit for honest completion, but treating an unpractised baseline as a polished performance can discourage the very diagnostic honesty the assignment needs. Use it primarily as a comparison point unless your course has another explicit assessment purpose.

What if the recording is noisy or slightly out of sync?

Judge whether you can hear the assigned language feature. Small synchronization offsets can come from recording latency, and background noise can come from the learner’s environment. Ask for a rerecording only when the evidence is genuinely unusable, not because the file is not studio-quality.

Do students need to report how many times they practised?

Not automatically. A repetition count can be useful for research or a particular classroom routine, but it is not the same thing as evidence that the target changed. If you collect it, be clear about why.

Collect evidence, not audio clutter

A completion tick tells you the student submitted something. A useful shadowing assignment tells a small learning story.

Take 1. Name one target. Repair it. Take 2. Change the words and transfer it.

Two comparable takes and one short new line can tell you far more than a folder full of unexplained “final” recordings — and they give the student a much clearer reason to listen back.