The Exam That Rewards the Wrong Skill

The test does not measure understanding of the passage; it measures recall and writing fluency.

• Most questions are recall. 'What colour was the boat?' and 'Name two things the narrator carried' can be answered by locating a detail, not by interpreting the passage. Recall is a lower skill than understanding.
• Length is being rewarded. Students who write long, fluent answers score well even when they misread the passage. That means the marking scheme is giving credit for prose rather than accuracy. The signal is inverted: fluent misreaders outscore careful readers.
• Short accurate answers are penalised. A careful reader who writes 'the boat was blue' may score less than a fluent student who writes a paragraph containing a misreading. The test therefore discourages the very behaviour it should reward.
• The stated aim is not operationalised. 'Understanding of the passage' has not been translated into questions that require inference, summary, or explanation of why something happened.

The colleague's suggestion makes this worse: more questions of the same kind simply increase the weight on recall and fluency.

A workable revision keeps the passage and the sitting date, changes the question types, and adjusts the preparation advice.

1. Replace recall items with inference and explanation items. For example: 'Why did the narrator carry those two things?' or 'What does the writer suggest about the boat's owner?' These require the reader to connect details to meaning.
2. Add a short summary task. Ask for a three-sentence summary of the passage. This tests understanding economically and is quick to mark with a clear rubric.
3. Write a rubric that separates content from expression. Award most marks for accurate reading and a smaller, explicit share for clarity. A fluent answer that misreads should lose content marks.
4. Keep the total at 40 marks by allocating, say, 10 for recall, 20 for inference, and 10 for summary, so the weighting matches the aim.
5. Tell the class the change. Since they were told to reread, add one lesson on inference and summary so the change is fair. Rereading is still useful, but it is no longer sufficient.

This fits the one-week limit because the passage and logistics are unchanged; only the questions and rubric are rewritten.

The teacher should look for evidence that the test now separates careful readers from fluent writers.

• Compare the two groups. Identify students whose answers were accurate but short, and those whose answers were long but misread. On the old test the second group scored higher; on the revised test the first group should score at least as well. If the pattern is unchanged, the revision failed.
• Check item-level performance. If the inference items are answered correctly only by students who also did well on recall, the inference items may be too easy or too dependent on recall. If they are answered randomly, they may be ambiguous.
• Look at the summary task. Strong summaries should capture the main idea in few words. Summaries that copy phrases from the passage signal recall rather than understanding.
• Ask for a second marker. A colleague marking a sample of scripts against the rubric checks whether content and expression are being scored separately in practice.
• Review the rubric's discrimination. If almost everyone scores in the middle, the marks are not distinguishing readers, and the weighting should be adjusted next time.

The test worked if accuracy, not length, predicts the score.