Task 1~20 minutes · at least 150 words
The bar chart below shows the average number of hours worked per year by employees in six countries in 1990, 2005 and 2020.Summarise the information by selecting and reporting the main features, and make comparisons where relevant.Write at least 150 words.
Average annual hours worked per employee, 1990, 2005 and 2020 05001,0001,5002,0002,5003,000 hours per year 2,6802,3501,910South Korea2,1902,2802,120Mexico1,8301,8001,790United States2,0301,7801,600Japan1,7701,6801,370United Kingdom1,5801,4301,330Germany 199020052020
Your answer (Task 1)
Band 9 model answer (192 words, counted)

The bar chart compares the average annual working hours of employees in six countries at three points between 1990 and 2020.

Overall, working time fell in every country over the thirty years, although the extent of the decline varied considerably. South Korea saw by far the steepest reduction, whereas hours in Mexico and the United States changed very little, and German employees consistently worked the least.

In 1990, South Koreans worked an average of 2,680 hours a year, almost 500 more than any other nationality. This figure then dropped sharply, to 2,350 in 2005 and 1,910 in 2020, a fall of almost 800 hours. Japan followed a similar but less dramatic path, declining from 2,030 to 1,600 hours. Mexico, by contrast, rose slightly to 2,280 in 2005 before easing to 2,120, which made it the country with the longest working year by 2020.

Among the remaining countries, the United States remained almost unchanged at around 1,800 hours, while the United Kingdom slipped to 1,680 by 2005 before falling steeply to 1,370. Germany recorded the lowest figure at every point, with hours shrinking from 1,580 to 1,330, just 40 below Britain by 2020.

✅ Self-assessment — mark yourself honestly
  • Task Achievement: Did your overview cover the general fall, the country with the biggest change, the stable countries and the lowest country — without specific numbers? Did you avoid giving reasons that are not in the chart?
  • Coherence: Did you group countries by their pattern (big fallers, the stable pair, the late faller, the lowest) instead of describing all eighteen bars one by one?
  • Lexical Resource: Did you vary your language for size of change (steepest reduction, changed very little, eased, shrinking) and avoid repeating "hours worked"?
  • Grammar: Accurate past tenses, comparatives and superlatives (by far the steepest, the longest working year, the lowest figure), with some complex sentences?
Task 2~40 minutes · at least 250 words
Write about the following topic:Increasingly, decisions that affect people’s lives, such as who is offered a bank loan or a job interview, are being made by computer programs rather than by human beings.
Is this a positive or negative development?Give reasons for your answer and include any relevant examples from your own knowledge or experience. Write at least 250 words.
Your answer (Task 2)
Band 9 model answer (304 words, counted)

Banks and employers now routinely use software to screen applications and predict how people will behave. Although this technology offers genuine efficiencies, I believe that, in its present largely unregulated form, it is a negative development rather than a positive one.

The case in favour rests on speed, consistency and scale. A program can assess thousands of applications in the time a human officer takes to read a handful, which lowers costs and shortens waiting times for applicants. Supporters also point out that human decision-makers are far from neutral: they may be influenced by tiredness, mood or personal prejudice, whereas a carefully designed system applies the same criteria to everyone.

However, these advantages depend on the system being carefully designed, and that is where the problems begin. Such programs learn from historical data, so if past decisions were biased against women or ethnic minorities, the software is likely to reproduce that bias while giving it an appearance of scientific objectivity. In one widely reported case, a recruitment tool marked down CVs that contained the word “women’s” because it had been trained on the records of a male-dominated workforce. Furthermore, many of these programs are so complex, or so commercially secret, that rejected applicants cannot discover why they were turned down, let alone challenge the outcome. Responsibility becomes blurred: the bank blames the software, and the developer blames the data.

For these reasons, the shift towards automated judgement seems to me harmful unless strict conditions are met. Organisations should be legally required to explain the main reasons for any decision, to test their systems regularly for unfair results, and to allow a person to review a case on request.

In conclusion, technology can usefully assist human judgement, but when it replaces that judgement without transparency or accountability, it represents a step backwards rather than progress.

✅ Self-assessment — mark yourself honestly
  • Task Response: Did you state clearly whether the development is positive or negative (or under what conditions) and keep that verdict consistent from the introduction to the conclusion?
  • Coherence: Did you present the opposing side fairly in one paragraph, then explain why your verdict outweighs it, using clear signposts (The case in favour…, However…, For these reasons…)?
  • Lexical Resource: Precise vocabulary (screen applications, bias, objectivity, transparency, accountability) with the prompt paraphrased rather than copied?
  • Grammar: A range of complex structures (conditionals, so … that, passive forms, relative clauses) with few errors? Did you reach 250+ words?