Do fillers like "um" lose marks in the O-Level English oral?

What do the band descriptors say about fillers?

Nothing by name. We searched every band of the three K300 oral scales (Response, Delivery and Spoken Interaction) for fillers, hesitation, pauses or repetition, and none of those words appears. What does appear is fluency, and in the band descriptors only in one place (the assessment objective on the same page also says "fluently"): the Delivery criterion of the Planned Response, which SEAB titles "Delivery with Awareness of Purpose, Audience and Context" and assesses on three things: pronunciation, fluency and intonation.

What does the fluency wording change from band to band?

Read only the fluency words and they form a ladder: "fluent", then "generally fluent", then "quite fluent", then "patches of fluent delivery", and at Band 1 fluency is not mentioned at all. Our reading, not SEAB's explanation, is that the ladder describes how much of the delivery flows. Fillers are one thing that can interrupt that flow; so are restarted sentences and losing your place. SEAB does not say how many of any of them move a response down a band, and we will not invent a number.

Each Delivery band also names pronunciation and intonation. We quote them and stop: this guide gives no advice on how anything should sound, because a written claim about sound is one we cannot source. That is a rule across OralPrep, not a judgement that those strands matter less.

Do fillers count in the Spoken Interaction?

Not under any fluency heading, because that scale has none. Spoken Interaction, worth 15 marks, is judged on "Communication of perspective(s)" and "Engagement in a discussion"; its descriptors mention pronunciation inside the vocabulary bullet, and never fluency. What the scale does reward is a "sustained discussion" with "well-developed perspective(s)". A turn so broken by fillers that its point is hard to follow is less developed and less sustained, so in our view fillers still matter there, through the content rather than as a separate deduction. The asymmetry between the two parts is laid out on how Response, Delivery and Spoken Interaction are marked.

Is a hedge like "I think" or "it depends" a filler?

No. In our reading, treating a hedge as a filler would throw away something useful. A filler fills a gap: "um", "er", "like", "you know", "so yeah". It carries no meaning, and removing it changes nothing. A hedge qualifies a claim: "probably", "in most cases", "it depends on…", "for older people, at least". Removing it changes what you have said.

The descriptors never mention qualifying or hedging, but in our view a qualification is often how a view shows it is "well-considered", the top Response band's own word. Our top-band worked answer to the recalled wigs prompt opens with a hedge: "Yes, I think they are — although 'good use of their time' probably means something rather different to them than it would to me." That "although" is the plan for the whole response.

Where hedges do harm, in our view, is when they stack up and never land: "I think maybe it is kind of good, I guess." That sentence never commits to a perspective. A hedge followed by a position ("it depends on who is using it, and for most teenagers I would say yes") is a perspective; three hedges and no position is not.

Why do fillers happen, and what helps?

SEAB does not say, and we hold no study of it. Our working answer, which you can test on your own recording, is that fillers bunch where the speaker does not yet know what comes next: at the very start of an answer and in the gap between one point and the next. If that is true of your recording, our guidance is to fix it on the content side.

• Know your first sentence. A prepared opening move (answer, then qualify) means the first seconds are not spent searching. See openings and endings for the Planned Response.
• Know the link to your next point. If your recording shows fillers between points, a planned connection ("But the more important thing is…") gives that gap words with meaning.
• Write notes as prompts, not a script. A glance at a two-word prompt gets you to the next idea; hunting for your place in a paragraph produces exactly the gap fillers fill.
• In the discussion, start with something true. "That is not something I have considered, but my guess is…" fills the gap with words that mean something. Recovering when you blank has more.

What we do not suggest is trying to suppress fillers by force while speaking. Our guidance keeps your attention on the argument, which is what the Response criterion and the whole Spoken Interaction scale judge.

How can you check your own fillers?

Count them, and note where they fall. Record one answer, write out what you said word for word (fillers included), and mark each one. You do not need to judge your own voice to do this; it is counting. Then look at where they cluster: at the start, between points, or at the end. That tells you which part of the plan to strengthen. A weekly routine built around this is in practising the O-Level English oral at home.

One related habit is worth catching at the same time: the pet linking phrase. "And also… and also…" is not a filler, but it is one structure used over and over, and both the Response and Spoken Interaction scales grade a "range of vocabulary and structures". The Spoken Interaction guide covers it under what loses marks in the discussion.