LATEST UPDATE - PTE TEST

Pearson is applying brand-new PTE scoring changes for the Describe Image and Retell Lecture tasks starting November 4, 2024.

The update is the introduction of double scoring, meaning an automated score is now followed by a human check.

pearson is making some upgrades to the scoring process for two PTE Speaking tasks: “Describe Image” and “Retell Lecture.” 

The questions themselves haven’t changed on the content, difficulty, or scoring criteria. The real update is the introduction of double scoring, meaning an automated score is now followed by a human check. 

Additionally, human experts have always been involved in scoring the PTE tests. They routinely review responses to ensure scores hit the mark.

The speaking tasks “Describe Image” and “Retell Lecture” are always based on three criteria: Content, Oral Fluency, and Pronunciation—each scored from 0 to 5.

• Content: This assesses how you can cover all the big-picture details, from key actions to possible outcomes. A full, well-organised recount of main ideas will serve you best, whereas random, disjointed points will get you low scores.

• Oral Fluency: This is about keeping your rhythm and phrasing natural, flowing as if in a casual chat. The ideal response sounds confident, and smooth, with no hesitations or awkward pauses.

• Pronunciation: Clarity is key. Speech should be immediately understood by anyone familiar with the English language. You should make vowel sounds and stresses sounding as close to native as possible.

All three of these have been machine-scored so far, but from 4 November 2024, a human expert will also review your content for an extra layer of assurance.

Blending AI with human oversight is probably the best way to ensure everyone taking the test gets a fair, accurate result. Human experts have been around all along – not just for training the AI but for monitoring live responses too. 

Now, Pearson is expanding their role so that when it comes to the content of these two speaking tasks, a real person will always review it.

So, what will they be looking for? Mainly that what you say sounds like your own words and not something lifted from a template. 

While their scoring systems are already set to detect template use, the human examiner will now cross-check this information for these tasks. Therefore, only genuine, fresh responses will earn a valid score.