08 — Studying for the PTE: Why Fluent Doctors Score 71 in Speaking

Part 8 of the AMC to Australia series. Previously: Start Studying for the AMC, Your EPIC Result Is In, Booking the AMC MCQ Exam, How the Top 10% Prepare, No Test Centre in Your Country?, Exam Day: The AMC MCQ Traps, and Booking the PTE

---

There is a particular kind of email that arrives two days after a PTE sitting.

Overall 74. Reading 78. Listening 76. Writing 72. Speaking 71.

The person who sends it has taken a history in English every working day for six years. They have presented at ward rounds in English, argued with consultants in English, broken bad news in English. They have never once been asked to repeat themselves by a patient. And a scoring engine has just told them they are five points short of the standard Australia's National Boards require.

The instinct is to conclude that the machine is broken, or biased, or that the test is a scam. That instinct is understandable and completely useless, because the machine is not going to change and you still need 76.

The better conclusion — the one that actually gets you to 76 in three or four weeks rather than three or four sittings — is this:

PTE Speaking is not a test of whether you speak English well. It is a test of whether you produce a specific acoustic pattern that an automated scoring engine has been trained to recognise as fluent.

Those two things overlap heavily. They are not the same thing. And the gap between them is filled almost entirely with habits that make you a better doctor and a worse PTE candidate.

This article is about closing that gap. It assumes you have read Part 7 and know the numbers: overall 63, Listening 58, Reading 59, Writing 60, Speaking 76 for tests sat on or after 23 April 2026, with per-skill floors of L53 / R54 / W60 / S66 if you are combining two sittings. Everything below is drawn from Pearson's published Speaking item scoring criteria, the PTE Academic test format pages, and the Ahpra English language skills standard. Read them yourself. This is a guided tour, not a substitute.

---

What the machine is actually measuring

Start with the arithmetic, because it explains everything else.

Pearson publishes the traits scored for each Speaking item type. For the four tasks that carry most of your Speaking score, they are:

Now look at the scales. Pronunciation is marked 0–5. Oral fluency is marked 0–5. Content varies by task: Repeat sentence content is 0–3, Describe image and Re-tell lecture content are 0–5, and Read aloud content depends on the length of the prompt.

Read that table again with a calculator in your head.

On Repeat sentence, ten of the thirteen available marks have nothing to do with whether you understood the sentence. They are awarded for how the sound came out. On Read aloud — where the words are printed on the screen in front of you and comprehension is not in question at all — pronunciation and fluency are ten marks against a content score you can max out by simply reading what is written.

This is the entire explanation for the 71.

You are being marked, in large part, on delivery mechanics. Rhythm. Phrasing. Continuity. Word stress. Whether your consonants land. Whether you stopped mid-sentence to think. A doctor with excellent English who pauses to consider the best word is producing worse acoustics than a nervous undergraduate who reads the whole line in one unbroken breath, and the engine scores acoustics.

The oral fluency descriptors, in Pearson's own words

This is the most useful paragraph in the entire scoring guide, and almost nobody preparing for PTE has read it:

5 — Native-like: "Speech shows smooth rhythm and phrasing. There are no hesitations, repetitions, false starts or non-native phonological simplifications."

4 — Advanced: "Speech has an acceptable rhythm with appropriate phrasing and word emphasis. There is no more than one hesitation, one repetition or a false start."

3 — Good: "Speech is at an acceptable speed but may be uneven. There may be more than one hesitation, but most words are spoken in continuous phrases. There are few repetitions or false starts. There are no long pauses and speech does not sound staccato."

Look at what separates a 5 from a 4. One hesitation. One. In a thirty-second response.

Not one hesitation per task type. Not one per minute. One, and you are capped at Advanced.

And look at what a "false start" is in ordinary human terms: beginning a sentence, deciding halfway through that there is a better way to say it, and starting again. That is the single most characteristic verbal habit of a careful, precise, thoughtful clinician. It is how you speak when you are choosing your words responsibly. It is a fluency penalty every single time.

The pronunciation descriptors matter differently

Pronunciation is where candidates panic unnecessarily, because they read "Native-like" and assume they need an Australian accent. They do not.

Read band 4, "Advanced": "Vowels and consonants are pronounced clearly and unambiguously. A few minor consonant, vowel or stress distortions do not affect intelligibility. All words are easily understandable... Stress is placed correctly on all common words, and sentence level stress is reasonable."

Nothing there is about accent. It is about clarity, completeness and stress placement. An Indian, Nigerian, Filipino, Sri Lankan or Egyptian accent scores 5 comfortably if the consonants are all produced and the stress lands on the right syllable.

What drops you into band 2, "Intermediate," is specific and fixable: "Some consonants are regularly omitted, and consonant sequences may be simplified. Stress may be placed incorrectly on some words or be unclear."

Consonants regularly omitted. Consonant clusters simplified. Word stress in the wrong place. Those are three concrete, drillable, measurable things — not a vague deficiency in your English.

---

The clinical habits that cost you the five points

Here is the uncomfortable and slightly funny part. Almost every habit that makes you good at talking to patients makes you worse at PTE Speaking.

The empathetic pause. You have been trained to leave space. After a difficult sentence, after a question, before a diagnosis. The engine hears a long pause and scores it against continuity.

Hedging and self-correction. "The patient — well, more accurately, the patient's daughter — reported..." This is intellectually honest and clinically correct. It is a textbook false start.

Trailing off at the end of a phrase. Clinicians soften the ends of sentences constantly, because a softened ending is less confrontational. The microphone hears volume falling away and consonants disappearing at exactly the point where the engine is checking whether the final consonant was produced.

Thinking aloud. "So... the main issue here is..." Filled and unfilled pauses are hesitations.

Rate variation for emphasis. Slowing down to stress something important is excellent communication. The descriptors reward "a constant and natural rate of speech" with smooth rhythm. Uneven speed reads as uneven.

Precision of vocabulary. Reaching for the exact right word takes a fraction of a second longer than reaching for an adequate one. That fraction of a second is a hesitation. On a task scored 0–5 on content, the adequate word scores the same as the perfect one.

None of this means those habits are bad. It means that for two hours in a partitioned booth, you are performing a different task than the one you perform at work, and you need to know which one you are doing.

That reframe alone is worth several points, because it removes the shame from the problem. Your English is not the issue. Your delivery profile is, and delivery is a motor skill. Motor skills respond to three weeks of deliberate practice in a way that "improving your English" absolutely does not.

---

Step one: diagnose, do not drill

The most common way to waste a month is to read a forum post saying "Speaking is hard, practise Read Aloud daily," and then practise Read Aloud daily without knowing which of the three traits is actually costing you.

Content, fluency and pronunciation fail for completely different reasons and respond to completely different fixes. Drilling the wrong one is not neutral — it burns the weeks you had.

Sit an official Scored Practice Test before you build a plan. Pearson sells these directly: a two-hour practice test using the same format and the same scoring engine as the real thing, returning a full Score Report. At the time of writing, a single Scored Practice Test is USD 35.99, with bundles — Essential (2 tests + Question Bank) at USD 59.99, Premium (3 tests + Question Bank + Official Guide) at USD 79.99, and Premium Plus (3 tests + Expert Self-Study Course + Question Bank + Official Guide) at USD 109.99. Check the preparation page for current pricing in your market.

Set against a real sitting at roughly AUD 400–500, and against the two-sitting floors that can disqualify a bad attempt from ever being combined, one scored practice test is the cheapest insurance in the entire pathway. Part 7 made this point about booking. It is worth making twice.

You also get a Skills Profile alongside your Score Report — a breakdown of performance across individual language skills with recommendations. Use it. It is the closest thing to a diagnostic the test offers.

Pearson has also launched PTE Academic Official AI Practice, which provides practice items scored on the official traits with instant feedback, plus 15 full-length mock tests and 20 sectional tests. And Smart Prep is free inside your myPTE account — study plan, guided practice tests, online courses and test tips at no cost. If your budget is tight, Smart Prep plus one Scored Practice Test is a perfectly respectable preparation stack.

What to do with the result:

Most doctors reading this land in the 70–75 band. That band has the best return on effort of anything in the entire AMC pathway: a handful of specific mechanical changes, three weeks, one re-sit.

---

The three-week plan that closes a one-skill gap

This is written for the 70–75 candidate — fluent, competent, five points short. Scale it up if you are lower.

Week 1 — Continuity

Goal: eliminate hesitations, repetitions and false starts. Nothing else. Not pronunciation, not content, not vocabulary. Continuity only.

The core drill: record and count. Every day, record five Read Aloud items and five Repeat Sentence items on your phone. Then play them back with a pen and mark a tally for every one of these:

• a pause longer than about a second
• an um, uh, or so
• a word repeated
• a sentence abandoned and restarted
• a final consonant that disappeared

Do not judge the recording. Count it. The number is the whole point, because it converts a vague feeling of "that sounded bad" into a metric that goes down.

Day one, a typical clinician records eight to fifteen marks across ten items. By day seven it should be under four. That fall — from "more than one hesitation" territory into "no more than one" territory — is the difference between band 3 and band 4 on every fluency-scored task in the test.

The one rule that fixes most of it: never stop for accuracy. If you misread a word in Read Aloud, keep going. If you cannot recall the second half of a Repeat Sentence, say something plausible in rhythm and finish the phrase. A single content error costs you one point in a length-dependent content score. Stopping to fix it costs you a hesitation and a false start, capping fluency at 3 across a five-point scale.

Say it out loud once, then never forget it: a wrong word said smoothly outscores a right word said twice.

The second drill: chunking. Take any paragraph of text. Mark it with slashes at natural phrase boundaries — roughly where you would breathe. Read it looking at the slashes, not the words. You are training your eye to run ahead of your mouth by a phrase, which is precisely what produces "most words spoken in continuous phrases" in the band descriptors.

Fifteen minutes a day. That is the whole of week one.

Week 2 — Sound

Goal: consonants produced, clusters intact, stress correct.

Go back to your week-one recordings and listen for something different. Not the pauses — the word endings.

The three specific things the descriptors punish:

1. Omitted final consonants. English marks tense and number at the ends of words: asked, fifths, lasts, helped, risks. Many first languages do not permit those clusters, and speakers simplify them automatically without hearing themselves do it. Every simplification is a pronunciation deduction and can also read as a content error, because "ask" is not "asked".

Drill: read a paragraph deliberately over-producing every final consonant. It will feel absurd and sound stilted. Do it for five minutes a day for a week and then read normally — the endings will survive.

2. Simplified consonant clusters. Strengths. Sixths. Twelfths. Clothes. Find the clusters your language does not have and drill them in isolation, then in words, then in sentences.

3. Misplaced word stress. This one is invisible to the speaker and glaring to the engine. Medical vocabulary is full of traps that IMGs consistently stress wrongly: PHAR-macy but phar-ma-CEU-tical; la-BOR-atory in Australian usage, not LAB-ratory; dia-BE-tes, an-aes-THET-ic, res-pi-RA-tory, cer-VI-cal. Build a list of thirty words you use daily, check each against a dictionary's audio, and drill the ten you had wrong.

Microphone technique belongs in week 2, because it is a sound problem, not a language problem. Two rules: a consistent distance of roughly a hand's width from your mouth, and no movement during a response. Candidates lean in when they concentrate and back off when they relax; the engine hears the volume swing as inconsistency. Fix your position before you start speaking and stay there.

And begin practising with noise. Part 7 flagged this and it belongs in your drills, not just your expectations: you will be in a room with ten to fifteen other candidates, all speaking, and if the first time you hear that is on test day it will cost you a hesitation on every one of the first three items. Practise with a recording of background chatter playing quietly. Not loud enough to distract — just loud enough that silence stops being the condition you need.

Week 3 — Full-length simulation

Goal: hold the mechanics for eighty minutes.

Part 1 of PTE Academic runs 76–84 minutes and contains nine question types. It is more than half the test, it comes first, and it is where your entire Speaking score is decided. Reading (23–30 minutes, five question types) and Listening (31–39 minutes, eight question types) follow.

The failure mode in week 3 is not technique. It is decay. Your continuity is excellent for twenty minutes and then your rate drifts, your pauses lengthen, and by Re-tell Lecture you are back to your normal speaking pattern. The engine samples the whole section.

So: do at least two full-length timed runs in week 3, in the morning, at roughly the hour your real test is booked for. Use a Scored Practice Test for one of them if budget allows — the second one, not the first, so the number you finish on is the one you carry into the exam.

Two days before the test, stop. No new material, no new technique. A light Read Aloud warm-up on the morning and nothing else. You are not going to acquire a skill in forty-eight hours and you can absolutely lose freshness.

---

Task by task: what each one actually rewards

Read aloud. The only task where the words are given to you, and therefore the purest test of delivery. Use the preparation time to scan for the words you might stumble on, and mark your phrase boundaries mentally. Then read at a steady, slightly slower-than-conversational pace, all the way to the end, without stopping for anything. Do not "perform." Even is better than expressive.

Repeat sentence. Content is 0–3: all the words in the correct sequence scores 3, at least half scores 2, less than half scores 1. Crucially, Pearson states that in scoring content, "hesitations, filled or unfilled pauses, leading or trailing material are ignored." But they are not ignored by the fluency trait. So the strategy is: repeat as much as you accurately can, in one continuous run, and if you have lost the middle, keep the rhythm and the sentence shape rather than stopping to search. Half a sentence said fluently scores 2 + 5 + 5. A full sentence said in three broken pieces scores 3 + 2 + 3.

Describe image. Content is 0–5, and the top band requires describing "all elements of the image and their relationships, possible development and conclusion or implications." This is the one task where content genuinely competes with delivery for your attention. The workable compromise is a light structure you have internalised — what it is, what the main trend or comparison is, one specific data point, one implication — spoken continuously. Note that this is a structure, not a memorised script. See the warning on templates below.

Re-tell lecture. Scored on content, pronunciation and oral fluency, with content requiring you to deal with the key points and their implications. This is a note-taking task with a speaking component attached. You get a notepad and pencil — use them. Write nouns and numbers, not sentences. Then speak from four or five noted anchors in continuous phrases rather than trying to reconstruct the lecture in order.

Answer short question. Scored on vocabulary only, correct or incorrect: appropriate word choice scores 1, inappropriate scores 0. There is no fluency or pronunciation trait here. Answer in one or two words, immediately, and move on. Do not build a sentence.

Summarize group discussion and respond to a situation were added to Speaking & Writing after 7 August 2025. Pearson's published item-scoring guide predates them, but they sit within the same trait framework and reward the same thing: continuous, clearly articulated speech that addresses the prompt. Prepare them exactly as you prepare Describe Image — light internal structure, no memorised text, continuity above elegance.

---

Three things not to do

Do not use memorised templates. They circulate widely, they are seductive, and they are a poor bet. Automated scoring systems are specifically trained to detect and penalise formulaic, memorised and off-topic content, and a template that thousands of candidates have submitted is by definition detectable. Beyond the risk, a template competes with the actual prompt for your speaking time and pushes content scores down. A light internal structure that you rephrase every time is safe and does the same job. A word-for-word script is not.

Do not speed up. Candidates who learn that pauses are penalised often respond by talking faster. The descriptors reward "an acceptable speed" with "smooth rhythm and phrasing" — not maximum words per minute. Speeding up increases the rate of consonant omission and cluster simplification, so you trade a fluency problem for a pronunciation problem and often end up lower.

Do not over-articulate. The opposite error. Deliberately separating every syllable produces exactly the "staccato" and "syllabic timing" that the descriptors name as markers of band 1. Aim for smooth and complete, not slow and separated.

---

The skills that are not Speaking

Speaking is where the gap usually is, but it is not the only number on the report, and a two-sitting strategy makes the others matter more than they look.

Writing 60 is the requirement, and 60 is also the two-sitting floor. There is no margin at all. A Writing 57 does not just fail the Writing requirement — it makes that entire sitting ineligible to be combined with another one. If your practice Writing is sitting at 62, that is not comfortable, that is one bad essay away from an unusable sitting. Spend a few hours on Summarize Written Text (one sentence, 5–75 words, all the key points, correct punctuation) and on essay structure. It is the highest-risk low-attention item in the test.

Reading 59 and Listening 58 are the requirements, with floors of 54 and 53. These are the most forgiving numbers on the table and most IMGs clear them without targeted work. Note, though, that Repeat sentence, Re-tell lecture and Summarize group discussion all feed your Listening score as well as your Speaking score, and Answer short question feeds Listening too. Work that improves your Speaking delivery on those tasks improves two communicative skills at once. It is the only genuine two-for-one in the test.

---

While you are doing this

A word about sequencing, because a PTE re-sit has a way of eating a month that was supposed to belong to the AMC MCQ.

Fifteen to twenty minutes a day of Speaking drills, plus two full simulations in week three, is the entire plan above. That is not a month of study. It is a month in which something else should still be your main work.

If your MCQ is still ahead of you, keep it running. The mistake candidates make emotionally — and Part 7 made this point about a first-attempt miss — is treating one setback as a reason to pause everything. Three weeks of "I'll get back to the MCQ once English is sorted" turns into six, and six turns into a cold restart on a question bank you had built real momentum in. AMC Prep Pro holds your bookmark-and-review queue and your category-level analytics while your attention is elsewhere, so a daily set of twenty questions is enough to keep the plan alive and warm rather than starting again from zero in October.

Fifteen minutes of Read Aloud in the morning. Twenty MCQs in the evening. That is a perfectly sustainable four weeks, and it ends with both things moving.

---

The study checklist

Before you start

• Sit an official Scored Practice Test — same format, same scoring engine, full Score Report
• Read your Skills Profile, not just the overall number
• Identify which trait is short: content, oral fluency, or pronunciation
• Set up Smart Prep in your myPTE account (free)
• Confirm which score table applies to your test date (on or after 23 April 2026: S76)

Week 1 — continuity

• Record 5 Read Aloud + 5 Repeat Sentence daily
• Count hesitations, repetitions, false starts and dropped endings — track the number down
• Rule: never stop to correct. Wrong word said smoothly beats right word said twice
• Chunking drill: mark phrase boundaries, read the slashes not the words

Week 2 — sound

• Over-produce final consonants for five minutes daily
• Drill the consonant clusters your first language simplifies
• Check word stress on thirty words you use daily; drill the ones you had wrong
• Fix microphone distance — a hand's width, no movement mid-response
• Practise with background chatter playing, not in silence

Week 3 — simulation

• Two full-length timed runs, in the morning, at your booked hour
• Watch for decay after minute twenty — that is the real failure mode
• Second Scored Practice Test if budget allows, late in the week
• Two days out: stop. Light warm-up only

Do not

• Memorise templates
• Speed up to avoid pauses
• Over-articulate into staccato speech
• Let the AMC MCQ go cold for a month

---

The point

The 71 is not a verdict on your English. It is a measurement of a delivery pattern, produced by a scoring engine that counts hesitations because hesitations are countable.

Which is, in the end, good news. A verdict on your English would take years to overturn. A delivery pattern takes about three weeks, because you are not learning a language — you are unlearning six specific habits that a careful clinician spends their whole training acquiring: the thinking pause, the mid-sentence correction, the softened ending, the emphatic slowdown, the hunt for the precise word, and the instinct to go back and fix a mistake.

For two hours in a partitioned booth, drop all six. Say the words in front of you, in order, in one continuous run, all the way to the final consonant. Then pick every one of those habits back up on the ward, where they are worth having.

76. Speaking. That is the number, and it is almost certainly the only one standing between you and an Ahpra application you could otherwise lodge.

Download AMC Prep Pro for iOS or Android and keep your MCQ preparation running through your PTE weeks.

---

Next in this series: PTE exam tips and tricks — the test-day decisions, timing calls and small mechanical choices that move a borderline score across the line.

Preparing for the AMC MCQ? AMC Prep Pro gives you 3,000+ AMC-style questions with full explanations and clinical imaging, full-length 150-question timed mocks with a live countdown and question navigator, a bookmark-and-review queue for exam week, and category-level analytics with a live pass prediction. Free daily practice sets on the free plan. Get the app for iOS and Android.

---

Sources

1. Pearson. "PTE Academic Speaking Scoring" (published item-scoring criteria). — Trait tables for Read aloud, Repeat sentence, Describe image, Re-tell lecture and Answer short question; full Pronunciation (0–5) and Oral fluency (0–5) band descriptors, including the Native-like and Advanced wording quoted above. Note that this published guide predates the two Speaking & Writing task types introduced after 7 August 2025.
2. Pearson. "PTE Academic scoring." — https://www.pearsonpte.com/pte-academic/scoring/ (AI scoring with human expert review on some responses; results typically within 48 hours; overall score 10–90 on the Global Scale of English; Score Report and Skills Profile; free unlimited score sharing)
3. Pearson. "PTE Academic & UKVI test format: Speaking & Writing." — https://www.pearsonpte.com/pte-academic/test-format/speaking-writing/ (Part 1: 76–84 minutes, 9 question types; Part 2 Reading: 23–30 minutes, 5 question types; Part 3 Listening: 31–39 minutes, 8 question types)
4. Pearson. "Prepare for PTE Academic & UKVI." — https://www.pearsonpte.com/pte-academic/preparation (Scored Practice Tests — same format and scoring engine, full Score Report, three versions; indicative pricing USD 35.99 single / Essential 59.99 / Premium 79.99 / Premium Plus 109.99; PTE Academic Official AI Practice with 1,000+ items, 15 full-length mocks and 20 sectional tests; free Smart Prep via myPTE)
5. Ahpra. "Accepted English language tests." — https://www.ahpra.gov.au/Registration/Registration-Standards/English-language-skills/Accepted-English-language-tests (PTE Academic minimums from 23 April 2026: overall 63, Listening 58, Reading 59, Writing 60, Speaking 76; two-sitting per-skill floors of L53, R54, W60, S66)
6. Ahpra. "English language skills." — https://www.ahpra.gov.au/Registration/Registration-Standards/English-language-skills.aspx (registration standard, pathway selection tool, evidence guide)
7. Australian Medical Council. "English language proficiency." — https://www.amc.org.au/pathways/english-language-proficiency/
8. AMC Prep Pro. — https://amcexampreparation.com/

Scores, fees, task types and scoring criteria change. Verify every requirement against the current Ahpra registration standard and Pearson's official documentation before you book or pay. AMC Prep Pro is not affiliated with the Australian Medical Council, Ahpra or Pearson.