CELPIP Speaking Task 4: How to Make Confident Predictions About a Scene

CELPIP Speaking Task 4: How to Make Confident Predictions About a Scene

CELPIP Speaking Task 4: How to Make Confident Predictions About a Scene

Many test takers feel relaxed after Speaking Task 3, where you describe a scene, and then get caught off guard by Task 4. The picture is the same, but the question changes: now you must say what is likely to happen next. If you freeze, repeat your description, or switch between tenses without noticing, your answer can lose clarity quickly.

The good news is that Making Predictions is one of the most learnable tasks on the CELPIP Speaking test. It rewards a simple structure, a handful of reliable future-tense patterns, and the habit of reasoning from visible evidence. This guide shows you exactly how to do that, with example phrases and a practice routine you can start today.

What Task 4 Actually Asks You to Do

In Task 4, you look at a picture and predict what will happen next. The scene is typically the same one you described in the previous task, so you already have a few moments to notice details. A short preparation period is followed by a short speaking period. Check the current official CELPIP materials for the exact timings, and practise with those limits so nothing about the clock surprises you on test day.

Notice what the task is not. It is not another description of the scene. It is not a story about your own life. It is a reasoned forecast: you look at clues in the image, decide what is most likely to happen, and explain why.

When your response is assessed, the things that matter are the same ones that matter across the Speaking test: how well your ideas are developed, how clearly and naturally you deliver them, the range and accuracy of your vocabulary, and the control of your grammar. For this task, future forms and linking language carry a lot of weight, because the whole answer is about what comes next.

A Simple Four-Part Structure for Predictions

A clear structure frees your mind to focus on language. Here is a framework that fits comfortably into the speaking time without sounding memorised.

  1. Quick anchor (one sentence). Name the situation so the listener knows what you are predicting about. Example: "In this scene, a group of coworkers are carrying boxes into a new office, so I think a busy afternoon is ahead of them."
  2. Main prediction with a reason. Say what will probably happen and point to the visual evidence. Example: "The man at the front looks worried about the narrow doorway, so they will probably struggle to fit the large desk through it."
  3. Second prediction or consequence. Show that one event leads to another. Example: "As a result, they will likely have to take the desk apart, which could delay the rest of the move."
  4. Wrap-up outcome. End with a confident final sentence. Example: "In the end, I expect they will finish setting up, but everyone will be tired by evening."

Spend only a few seconds of your preparation time on the anchor. Use most of it to choose your two predictions and one piece of visual evidence for each. Having a reason attached to every prediction is what separates a developed answer from a list of guesses.

Language That Makes Predictions Sound Natural

Strong predictions rely on a variety of future structures. Repeating "will" in every sentence sounds flat, while mixing patterns shows range. Build a small toolbox and practise it until it is automatic.

Certain and probable futures

  • "She will probably realize she forgot her umbrella."
  • "They are going to miss the bus if they don't hurry."
  • "It is likely that the line will get longer."
  • "I would expect the children to start running around soon."

Softer, less certain futures

  • "He might decide to ask for help."
  • "There is a good chance that the café will run out of seats."
  • "It could turn into a long delay."

Linking ideas together

  • Because of that / As a result — to show a consequence
  • Meanwhile — to move to another person or action in the scene
  • Eventually / In the end — to close the prediction
  • Judging by / Based on — to introduce visual evidence

A useful sentence frame is: "Judging by [detail], I think [person] will [action], which means [consequence]." It carries a clue, a prediction, and a result in one smooth sentence. Do not use it every time, but having it available keeps you moving when you are short of words.

Also be careful with tense control. A common slip is drifting back into the present tense: "Then he opens the door and the dog runs out." In a prediction, keep the future perspective: "Then he will open the door, and the dog will probably run out." If you want to practise using vivid, precise words for images, the guide on CELPIP vocabulary for image tasks pairs well with this one.

Finding Predictions Fast: The Clue, Cause, Consequence Method

The hardest part for many learners is not grammar; it is thinking of something to say within a short preparation window. Try this three-step thinking routine.

  1. Clue. Pick the most interesting detail in the image: an expression, an object, a weather condition, a crowd, a time of day.
  2. Cause. Ask, "What is this person or situation heading toward?" A child holding a melting ice cream is heading toward a mess. A crowded platform is heading toward a delayed train.
  3. Consequence. Ask, "And then what?" This second question gives you your second prediction almost for free.

A worked example

Imagine a scene of a family at a park picnic table. Dark clouds are gathering, a father is cutting a cake, and a young girl is looking up at the sky. Here is a model response built with the four-part structure:

"In this scene, a family is having a picnic in a park, and it looks like they are about to enjoy a birthday cake. However, judging by the dark clouds above them, I think it will start raining very soon. The little girl is already looking at the sky, so she will probably be the first to notice, and she will likely warn her parents. As a result, the family will have to pack up quickly, and the father might need to cover the cake with his jacket. In the end, I expect they will run to their car, a little wet but laughing, and finish the celebration at home."

Look at what this answer does well. It anchors the situation, uses evidence (the dark clouds, the girl's gaze), links events with "as a result" and "in the end," and mixes will, will probably, will likely, and might. It also includes a human touch that makes the prediction memorable without needing complicated vocabulary.

Common Mistakes and How to Fix Them

  • Describing instead of predicting. If more than a sentence or two of your answer is about what is already in the picture, you are repeating Task 3. Keep description short and use it only as evidence.
  • Making one prediction and running out of time. Aim for two or three connected predictions so that you can fill the speaking time naturally.
  • Going silent while searching for the perfect idea. Any reasonable prediction is acceptable. Choose the first logical one and develop it with reasons rather than hunting for something original.
  • Using the same structure repeatedly. If every sentence starts with "He will...", vary your openings with time expressions, conditionals, and linking phrases.
  • Contradicting yourself. Avoid saying it will rain and then that the family will enjoy a sunny afternoon. Plan your prediction path before you speak.
  • Rushing at the end. Many speakers speed up and mumble in the last few seconds. Finish with a calm, complete concluding sentence instead.

Delivery also matters. Steady pacing, clear stress on key words, and natural pauses between ideas make your prediction easier to follow. For more on sounding smooth rather than rehearsed, see these CELPIP fluency tips.

A Practice Routine You Can Repeat

Skills like this improve through short, repeated, focused sessions. Try this 20-minute routine a few times a week.

  1. Collect scenes (3 minutes). Choose a photo from a news site, a photo library, or your own camera roll. Aim for scenes with several people and some tension or movement.
  2. Prepare under time pressure (1 minute). Set a short timer. Write only keywords: two clues, two predictions, one linking phrase.
  3. Record your answer (1 minute). Speak without stopping, even if you make a mistake. Recovering smoothly is a skill the real test rewards.
  4. Listen back and check four things (5 minutes). Did you predict rather than describe? Did you give reasons? Did you vary your future structures? Did you stay in the future tense?
  5. Re-record (1 minute). Fix one issue only. A second attempt with a single clear goal produces faster improvement than trying to fix everything at once.
  6. Keep a phrase bank (2 minutes). Write down any new, natural expression you used or want to use next time.

Task 4 also fits into a larger preparation picture. If you are working through the tasks in order, you may find it helpful to compare the reasoning skills here with those needed for a very different situation in CELPIP Speaking Task 6: Dealing with a Difficult Situation, where tone and diplomacy matter more than forecasting.

How USpeak Helps You Practise This Task Under Real Conditions

The biggest challenge with self-study is the black box: you record yourself, but you do not know how the exam would actually score you. USpeak is an AI-powered speaking coach designed to close that gap by mirroring the real CELPIP experience.

  • Guided practice for all 8 CELPIP speaking tasks, including Making Predictions, with authentic timing so preparation and speaking windows feel like the real thing.
  • Predictive AI scoring aligned to CELPIP's 0-12 rubric, so you can see where your response currently sits and what to improve.
  • Diagnostics for fluency, pronunciation, grammar and vocabulary, which helps you notice patterns such as tense drift or repetitive structures.
  • AI-generated Level 12 sample responses with instant tips, so you can compare your prediction to a model answer and borrow useful phrasing.
  • Recording history and progress charts, which make improvement visible week by week.
  • Test Packs for intensive prep before test day, when you want to practise full sets of tasks back to back.

Use the tool as a feedback loop rather than a shortcut: practise, review the diagnostics, apply one change, and record again. No app can guarantee a particular score, but consistent, informed practice gives you the best chance to perform at your own highest level.

Conclusion and Next Steps

Speaking Task 4 is less about creativity and more about clear thinking. Look for clues, predict what comes next, explain why, and connect your ideas with confident future language. With the four-part structure, a varied toolbox of future forms, and short daily practice, you can turn this task into a reliable strength.

Here is a simple plan for this week. First, collect five scenes and practise the Clue, Cause, Consequence routine on each. Second, record and review at least three answers, checking tense control and variety. Third, build a personal phrase bank of ten prediction expressions that sound natural in your own voice.

When you are ready to see how your answers hold up under real exam conditions, run a full CELPIP simulation in USpeak and get detailed feedback on every task. Download USpeak on Google Play and take your next practice test today.

Share this article:

Ready to Improve Your Speaking Skills?

Download USpeak and start practicing with AI-powered feedback today!

Download Free App