TOEFL Speaking Mistakes by Criterion

TOEFL Speaking mistakes by criterion are easier to repair than a single low band, because the next recording can aim at one bar. Listen and Repeat uses Repeat Accuracy, Intelligibility, and Fluency. Take an Interview adds Language Use and Organization. A mistake that belongs to one bar gets worse when you drill a different bar. This guide sorts the common misses and tells you which task to record next.

In this guide

1. Which criteria apply
2. Repeat Accuracy
3. Intelligibility and Fluency
4. Language Use and Organization
5. Picking tomorrow’s recording
6. Try this yourself
7. Frequently asked questions

Which criteria apply

A Listen and Repeat item does not have an opinion to organize. Judging it for Organization will send you into interview practice while your repeats still drop room numbers. An interview item does not ask you to reproduce a sentence verbatim. Judging it for Repeat Accuracy will make you memorize phrases that do not answer the question.

How Speaking is scored in 2026 is the definition page. The section overview is the timing page: seven sentences, then four 45-second answers, no preparation screen. Use those pages when you need the format. Use this page when a report is already in front of you and you need the pile.

The illustrative marks in the Listen and Repeat sample show two opposite patterns: clear but cut off, and fast but muddy. Those are Fluency-and-accuracy versus Intelligibility. They are not five problems at once.

Repeat Accuracy

Repeat Accuracy fails when the sentence’s content or order changes. Paraphrase is the polite version of this miss. “The upstairs printer is broken” does not preserve “second floor” and “offline.” Swapped room numbers preserve the topic and change the campus. A dropped “don’t” preserves the rhythm and reverses the instruction.

The repair is not a bigger vocabulary. It is a rule: numbers, names, rooms, and negatives are copied, not improved. Record another Listen and Repeat set with that rule and ignore your words-per-minute. If accuracy rises and fluency falls for one set, that trade is acceptable. You can put fluency back after the content survives.

Do not drill this bar on interview answers. An interview that copies the question word for word can still be off-topic. Different task, different miss.

Intelligibility and Fluency

Intelligibility fails when a listener cannot tell the words apart: endings disappear, content words blur, or stress hides the noun. Fluency fails when the first word is late, a pause swallows the middle, or fillers replace the verb. A fast muddy answer can score well on fluency and poorly on intelligibility. A slow clear answer can do the reverse. Pronunciation and intelligibility and fillers and pauses split the drills. Do not run both in one evening.

The repair for intelligibility is one ending or one stress pattern, spoken clearly, even if the window feels less full. The repair for fluency is the beep: first word immediately, then a steady line, not a faster line. Restarting a Listen and Repeat sentence inside an 8-second window usually costs the ending. One take is the constraint. A second private take is a different task and should not be compared with the report.

Language Use and Organization

These two apply to Take an Interview. Language Use is range and control: verb forms that match the time you are describing, words that mean the thing you mean, sentences a listener can parse once. Organization is the shape: a direct answer, a reason, an example, and a stop before the 45 seconds become a new topic.

A common Language Use miss is a string of correct words in an order that hides the point. More idioms will not fix it. A shorter sentence will. A common Organization miss is a true story that starts too late. The question was “which do you prefer.” The first sentence was background. The preference arrived at second 30 and the example was cut off. The 45-second strategy is the shape. This page only tells you not to mix that shape into a repeat-accuracy drill.

Organization does not mean a memorized template that ignores the question. A four-part answer to a question that was not asked is fluent and off-task. Answer first. Then use the shape.

Picking tomorrow’s recording

Read the lowest bar on the last report. Record only the task that uses that bar.

Lowest bar	Next recording	Ignore for that day
Repeat Accuracy	Seven new sentences	Speed, idioms
Intelligibility	Seven sentences, slower endings	New interview topics
Fluency	Seven sentences, start on the beep	Accent projects
Language Use	Four interview answers, shorter sentences	Listen and Repeat
Organization	Four answers, preference in sentence 1	Vocabulary lists

If two bars are tied, pick the one that changes the information: accuracy before fluency on repeats, organization before language polish on interviews. A listener can forgive a plain verb. A listener cannot recover a swapped room or a missing preference.

Speaking practice and Listen and Repeat are the repeat URL. Take an Interview is the interview URL. The checker page, Speaking checker, explains the report you are sorting. One credit when the task starts. The 3-credit full section is for the day you want both tasks, not for diagnosing a single bar.

What the timing line is allowed to decide

A Speaking report can show words per minute, pauses, fillers, and time to first word. Those numbers belong to Fluency unless a missing ending is why the word count looks low. Do not let a high words-per-minute number talk you out of an intelligibility miss. The fast take can be the muddy take. Do not let a low words-per-minute number talk you out of a Repeat Accuracy miss if the words that were said were the wrong rooms. Speed did not cause the swap. Memory for the content words did.

Use time to first word as a fluency decision only when accuracy and intelligibility are already acceptable. If the first word is late and the sentence is also wrong, fix the sentence. A prompt start on the wrong announcement is a confident error. Interview timing works the same way. A preference that arrives at second 30 is an organization problem even if the language after that second is clean. Move the preference, then judge the verbs. The 45-second shape is the picture. The criterion list is how you refuse to work on all of it tomorrow.

If the report has no timing line because the attempt failed to record, do not invent a fluency story. Check the microphone and record once more. A missing attempt is not evidence about your English. Paying for another credit before the microphone works is the waste described in free versus paid practice. The criteria cannot score silence.

Try this yourself

Open your latest Speaking report, or use Take B from the sample if you do not have one: clear speech, sentence 6 cut off. Name the lowest bar in four words. Record only that task once. If you notice a second problem while you speak, write it down and do not chase it until the next day.

Start Listen and Repeat when the bar is accuracy, intelligibility, or fluency. Use the interview task when the bar is language or organization.

Frequently asked questions

Do all five criteria apply to every Speaking item?

No. Listen and Repeat uses Repeat Accuracy, Intelligibility, and Fluency. Interview items add Language Use and Organization. Scoring a repeat for organization mixes the tasks.

Should I fix fluency by speaking faster?

Only if the words stay distinct and the content survives. Faster muddy speech often lowers intelligibility. Start on time before you add speed.

Can one recording repair two criteria?

It can show two problems. It rarely fixes two. The second problem becomes tomorrow’s recording. Today’s recording has one rule.

Are these official ETS criteria names?

They are the criteria this practice report uses, aligned with what the tasks measure. An official score is produced by ETS, not by this list. Practice marks are not official results.

Next steps

Name the lowest bar from the last report and record that task once. The 14-day Speaking plan will hold the rest of the week if you keep one bar per day.