Scoring Recorded Responses Using Official Performance Criteria

Hello again. Your baseline recordings and timestamped log are ready. Last time, you identified what happened when speech broke down. Today you will take one response and make a more demanding judgment: how well did it meet the task, how understandable was the speech, and how smoothly was it delivered?

Allow about 40 minutes. Use a Describe Image recording from your baseline if you still have its image. It gives you one clear prompt against which to judge content and, like other scored speaking tasks, uses pronunciation and oral-fluency criteria. If you did not keep the image, use a recording for which you can retrieve the exact prompt, or make a new Describe Image recording while saving its image. Keep any new recording separate from your original baseline. You cannot score content reliably from memory of the prompt.

Three scores, three different questions

Pearson assesses Describe Image on Content, Pronunciation, and Oral Fluency. Start with its current task page to see the distinction, particularly the difference between covering the image and sounding smooth.

For self-assessment, keep the questions separate:

• Content: Did you describe what this particular image shows, accurately and with meaningful relationships between its elements?
• Pronunciation: Could a regular English speaker understand your words readily, including their sounds and stress?
• Oral fluency: Did your speech move in natural phrases at a reasonably steady rate, without disruptive hesitations, repetitions, or false starts?

This separation matters for your template concern. A memorized opening might sound fluent while consuming time without describing the image. That is chiefly a content concern. Conversely, a response may identify the right trend but stumble repeatedly while expressing it: good information does not erase the fluency evidence. A regional accent is not, by itself, a pronunciation error; intelligibility is the central question.

Anchor each rating to Pearson’s descriptors

Pearson’s Score Guide supplies a 0–5 content scale for Describe Image. Its levels distinguish, for example, covering most key elements and an implication from mentioning only basic or disconnected elements. Read the whole scale before choosing a number; do not begin with a score and search for wording to justify it.

The same guide gives detailed 0–5 pronunciation and oral-fluency scales. Unlike Describe Image content, these are shared criteria for speaking tasks scored on those traits.

One useful distinction is between pronouncing a word unclearly and pausing before it. If “distribution” is hard to recognize because its sounds or stress are distorted, note pronunciation evidence. If you say it clearly after a long search for the word, note fluency evidence. Both can occur together, but explain each rating with its own evidence.

Rate your recording in three passes

Set aside about 18–20 minutes. Put the saved image, recording, and last lesson’s timestamped log where you can see them. Work from what was actually recorded, not what you intended to say.

First pass — content, with the image visible. Before replaying, jot down the image’s main subject and its key elements: for a chart, these might include the categories, prominent values or changes, and the relationship that makes the chart informative. This is an assessor’s reference, not a script for a better answer. Listen once and mark what your response covered, misstated, or missed. If you offered an implication, check that the image supports it rather than rewarding an impressive-sounding guess. Compare your evidence with all six Describe Image content levels and write a tentative 0–5 rating.

Second pass — pronunciation, listening rather than reading a transcript. Replay the answer without looking at any words you wrote down from it. Mark two or three timestamps where words were immediately understandable or where a listener might reasonably struggle. Ask whether the issue is a speech sound, word stress, or sentence stress—not merely that you have an accent. Then select the closest 0–5 pronunciation descriptor. If the microphone clipped or muffled a passage, label that passage technical uncertainty; it is not sound evidence of how you pronounced it.

Third pass — oral fluency. Listen for the overall rhythm and where phrases begin and end. Consult the P, R, and FS observations from your previous log, but check them against the audio again. Did a pause occur between complete thoughts or interrupt an unfinished phrase? Were most words grouped continuously, or did the response become word-by-word? Compare the overall pattern with neighbouring levels on Pearson’s scale and assign a tentative 0–5 rating. Note-reading problems and incomplete ideas from your log may explain what happened, but they are not extra fluency traits to score separately.

Record the result compactly:

Here is how the distinctions work in an invented case, not a rating of your recording. Suppose a chart shows public-transport use rising, driving falling, and cycling staying unchanged. A speaker accurately describes the first two changes and their relationship, but omits cycling. If the response also contains one brief “um” yet otherwise flows in clear phrases, its content rating must reflect the missing element while its fluency rating can reflect the largely smooth delivery. Nothing in that written account establishes a pronunciation rating: you would have to hear the audio. Do not infer one trait’s score from another.

Turn three ratings into one useful diagnosis

Your three numbers are provisional self-ratings, not an official PTE result or a prediction of a 90. Pearson’s operational scoring uses more than your judgment of one recording, and even careful listeners may disagree near a boundary. If you are torn between adjacent levels, write both—such as “3–4”—and state what additional evidence would decide it.

Now write one sentence that names the strongest constraint on this particular response. For example: “I covered the main comparison, but omitted a key category; my delivery stayed mostly continuous, and pronunciation was readily understandable.” That is more actionable than “Speaking: bad” or “I need to speak faster.” It also prevents a familiar trap: fixing a pause will not restore missing image content, while adding more facts will not automatically smooth a broken sentence.

If you have time, replay the response once at normal speed without stopping. Check whether your written ratings still match the impression of the whole answer. Keep the recording, prompt, ratings, and evidence together for later comparison.

Takeaway

You have now assessed one recorded response on three independent official traits: task-specific content, understandable pronunciation, and smoothly phrased oral fluency. The valuable result is not just three tentative numbers; it is the evidence beside them and an honest note about uncertainty.

In the next module, you will begin working on fast sentence construction from keywords. Today’s rating gives that work a purpose: you can see whether difficulty forming a sentence quickly is showing up mainly as missing content, interrupted fluency, or both.