Interactive Communication: The One You Cannot Practise Alone

Every other Cambridge speaking criterion can be rehearsed in a quiet room. Grammar, vocabulary, discourse and pronunciation all improve alone.

Interactive Communication does not, and across the 35 candidate profiles Cambridge publishes, it is the top mark more often than pronunciation is.

SpeakShark is free for the one criterion that needs somebody to talk back, three AI conversation sessions a day with no card.

In this guide: the count · the arc · what Cambridge tells teachers · the repeat rule · the phrase Cambridge names · practice · limits · method · FAQ

Key takeaways

• Across 30 non-flat published profiles, interaction is top or joint top 23 times, pronunciation 19.
• Interaction is the single highest mark seven times, and all seven are at B1 or B2.
• At C2 it is the top mark for one of four and the bottom mark for two.
• Cambridge's guides say examiners cannot explain or rephrase, only repeat.
• Cambridge's own teacher checklist calls out candidates who just agree with their partner.

Interaction tops more profiles than pronunciation

We have now read every examiner commentary Cambridge publishes for B1 Preliminary, B2 First, C1 Advanced and C2 Proficiency. That is 35 candidates. Five have completely flat profiles, identical marks on every criterion, so they cannot tell you which criterion stands out and we exclude them from the counts.

Twenty three against nineteen. On the published evidence, the criterion most often sitting at the top of a Cambridge speaking profile is not pronunciation. It is interaction.

That is worth stating carefully. It is a count of how often each criterion lands at the top of a published profile. It is not a statement about weighting, because Cambridge does not publish weights, and it is not a claim about a random sample, because these documents are teaching materials.

It does undercut a story we ourselves helped tell. Our first look at nine candidates put pronunciation at the centre. With 35 profiles rather than nine, interaction is the more frequent strength.

It carries you at B1 and B2, then stops

Read the "alone at the top" column down the levels and the arc is clean.

Interaction is the single highest mark for seven candidates, and every one of them sits at B1 or B2. Four at B1 Preliminary, three at B2 First. At C1 Advanced it reaches the top five times but never alone. At C2 Proficiency it is top for one of the four non-flat candidates.

Then it reverses. At C2, interaction is at the bottom of the profile for two of four. At B1 it is at the bottom for one candidate out of eight non-flat.

So interaction behaves like a lever that works early and flattens out. At B1 a candidate who asks questions stands out, because plenty of candidates at that level simply answer and stop. By C2 everybody in the room can hold a conversation, so interaction stops separating people and other things do.

We saw the mechanism in the commentary rather than only in the numbers. At B1, Andrea takes a 5 for interaction while the examiner notes she is "more reactive than proactive", and Kenza loses half a band because she "does not invite her partner to respond to what she says", both covered in Cambridge speaking criteria by level. At B2, the three candidates whose pronunciation is not their top mark are exactly the three whose interaction is, which we set out in 14 real B2 First speaking scores.

What Cambridge tells teachers to look for

The teacher guides hand teachers an observation checklist for this criterion, phrased as questions about a named candidate. Cambridge summarises the criterion as initiating, responding, developing, then asks:

• Do they start conversations in a suitable way?
• Do they reply to their partner and examiner appropriately?
• Do they link what they say to what their partners say?
• Do they keep the conversation going?
• Do they need any help to do this?
• Do they develop the conversation, adding new ideas, or just agree with their partner?

That last question is the sharpest thing in the document. Agreeing with your partner is named as the weak version of participating. It is not silence, it is not rudeness, it is the polite default that most nervous candidates fall into, and Cambridge has written it into the checklist as the thing teachers should notice.

The C2 guide goes further and glosses its own band wording. "Widens the scope of the interaction" is explained as developing and broadening the discussion, and proactively involving the other candidate with a suggestion or question about further developing the topic.

Both phrasings put the obligation on you to bring your partner in. Not to be pleasant to them, to give them something to answer.

Examiners cannot rephrase, only repeat

Buried in the frequently asked questions of the teacher guides is a rule that changes what you should do in the room, and almost no candidate knows it.

Asked how examiners can rephrase a question a candidate has not understood, Cambridge answers that examiners cannot explain or rephrase, but they can repeat questions and instructions. It then advises teachers to teach phrases like "can you say that again" or "can you repeat that, please".

Read that as a candidate. If you say "sorry, what do you mean?", you are asking for something the examiner is not permitted to give. The best case is that you get the identical sentence back, having spent time and signalled confusion. If you ask for a repeat, you get the identical sentence back on purpose, which is exactly what you wanted, and you have used a natural conversational move while doing it.

The same guides note that a repeat is available for instructions too, not just questions. So if you did not catch what the task is in a collaborative part, asking is the correct move rather than guessing and drifting off task.

The phrase Cambridge names in its own guide

Explaining what "linking contributions to those of other speakers" means, the C2 guide points at a specific technique: referring to things people said earlier in the conversation, using phrases like "as you said earlier".

That is the exam board naming a phrase. It is worth having in your mouth before test day, along with the moves it stands for.

• Refer back. "As you said earlier, the cost matters more than the time."
• Ask, do not just answer. "What would you choose, though?"
• Add a reason to their point rather than agreeing with it. "I agree, and the reason I would go further is..."
• Bring them in on a new idea. "Do you think that would work for the second one as well?"

None of these are advanced English. Mohammed, the B1 candidate who beats his more accurate partner on interaction, does it with "What about this bag, what do you think about this bag?"

Cambridge also answers a teacher asking how to prepare students who feel intimidated by an examiner: do role plays. That is the exam board's own recommendation, and it is why role play scenarios and a regular speaking partner do more for this criterion than any workbook.

How to actually work on it

The parts where interaction is scored are the ones with another candidate in them.

• B1 Preliminary: Part 3 and Part 4.
• B2 First: Part 3 and Part 4.
• C1 Advanced: Part 3 and Part 4.
• C2 Proficiency: the collaborative task.

Two practical rules follow from the data. If you are working at B1 or B2, interaction is the cheapest band you can buy, because a single question asked of your partner is often the difference. If you are at C1 or C2, do not expect it to carry you, and check whether grammar and lexical range is the column actually holding you down.

Before planning anything, test your speaking level free, because the advice genuinely differs by level, and CEFR levels for speaking explains what the bands describe. If you want to know who is marking what, the two examiner roles is the short version: the silent one marks this.

What we could not verify

These are curated teaching materials. Thirty five candidates chosen by Cambridge to illustrate marking. Every proportion here describes that published set and nothing wider.

Five profiles are flat and carry no signal. We excluded them from every top and bottom count and have said so at each point rather than letting them inflate a figure.

We are not claiming interaction is weighted more heavily. Cambridge does not publish weights for its speaking criteria. Our counts are about where marks land in a profile, not about how the final grade is computed.

We did not read the C1 Advanced teacher guide. Cambridge publishes one and we read the B1 Preliminary, B2 First and C2 Proficiency guides. Quotations here come from those three.

We are not reproducing the assessment scales. We quote the guides' own explanations and checklists, not the band descriptor tables.

The repeat rule is quoted from the guides, not from an exam regulation. It appears in the frequently asked questions of more than one guide, which is why we treat it as a general rule, but the authority on exam conduct is Cambridge itself.

How we researched this guide

The score counts come from every examiner commentary document Cambridge publishes across the four levels, downloaded and read rather than summarised. Getting to 35 took two corrections along the way, and both came from reading a preparation page rather than trusting our own notes: B1 turned out to have four documents and nine candidates rather than two and four, and B2 turned out to have six documents and fourteen candidates rather than three and seven.

The counting was done in code with the rules fixed before we looked at the output. A profile counts as flat when every criterion carries the same mark, and flat profiles are excluded from any top or bottom count, because a candidate level across the board is simultaneously at both. That rule removes five candidates and is why the denominator is 30 rather than 35.

The guidance and the quotations come from three teacher guides, each around twelve megabytes with the prose embedded among font data, so the text had to be filtered to readable runs before it could be read at all. The frequently asked questions section is where the repeat rule sits, several pages after the assessment material, which is a long way from where a candidate would think to look.

We checked the repeat rule appeared in more than one guide before presenting it as a general rule rather than a level specific one. It did.

Practise speaking, from SpeakShark

SpeakShark is an AI English speaking practice app, and this is the criterion it exists for. You cannot rehearse initiating, responding and developing against a textbook, because all three need a reply you did not write.

You talk, the AI answers what you actually said, and you get speaking feedback while the conversation is still running. The free tier gives basic feedback; the detailed pronunciation and grammar breakdown is on Premium.

Being straight about the limits: a free session runs five minutes with four turns, and it is one on one. A Cambridge speaking test puts you with another candidate whose goals differ from yours, and we do not reproduce that. For the negotiation part of the higher levels, practise with people.

Start free with three sessions a day and no card. Paid sessions run ten minutes with unlimited turns. Limits are on the pricing page, and how it works walks through a session. Sign up here if you want to practise asking questions back rather than answering them.

We are a speaking improvement tool. We are not an exam preparation provider, and we are not affiliated with Cambridge English, the IELTS partners, Pearson or any exam board. For test format, booking and official practice material, go to the exam body directly. Use SpeakShark to make your spoken English stronger, and use official material to learn the test.

• Site: speakshark.com
• Support: speaksharksupport@gmail.com
• Free tier: 3 AI conversation sessions per day, no card

Sources

• Cambridge English, B2 First and B2 First for Schools, Assessing speaking: a guide for teachers: the initiating, responding, developing summary, the six question observation checklist including the just agree with their partner line, the examiners cannot explain or rephrase answer and the role play recommendation
• Cambridge English, C2 Proficiency, Assessing speaking: a guide for teachers: the gloss on widening the scope of the interaction, the proactively involves the other candidate wording and the as you said earlier phrase, plus the same repeat rule
• Cambridge English, B1 Preliminary and B1 Preliminary for Schools, Assessing speaking: a guide for teachers: the vary feedback guidance and the assessor and interlocutor roles
• Cambridge English, B1 Preliminary preparation and B2 First preparation: the pages publishing the examiner commentary documents behind every score counted here
• Council of Europe, CEFR Table 1, common reference levels global scale: the level scale these qualifications report against

FAQ