Comprehension Item That Rewards Inference Beyond the Literal Text
You are marking the unseen passage section, worth 10 marks, of an SSC Arabic paper. The passage describes a farmer who finds a leaking water channel at dawn and repairs it before his neighbours wake. Question 3, worth 2 marks, asks «ماذا فعل الفلاح عندما رأى الماء يتسرّب؟» and is answered literally by most candidates. Question 4, worth 3 marks, asks «ماذا نتعلّم من سلوك الفلاح؟». Candidate D answered Question 4 with «نتعلّم أنه كان خجولاً» ('we learn he was shy'), apparently inferring shyness from the pre-dawn timing. Candidate E answered «نتعلّم أن نساعد الآخرين دون انتظار الشكر», which is not stated in the passage but follows from repairing the channel before anyone wakes. The passage never uses the words 'شكر' or 'خجل'. The rubric says inference answers must be 'supported by the text'. Fifty-eight scripts show a similar spread, and 19 candidates left Question 4 blank.
1. Score D and E on Question 4 out of 3 and justify the difference.
Read the worked answer
D should score 0 or, at most, 1 if the school's rubric allows a mark for a relevant attempt. Shyness is not merely unstated; it is contradicted by the detail that he acted at dawn to help neighbours, which points to unostentatious service rather than avoidance of people. An inference that the text resists cannot be 'supported by the text'. E should score 2: the answer captures the implicit motive, and the pre-dawn timing is a genuine textual clue for helping without display. It falls short of 3 because 'without waiting for thanks' adds an element the passage never mentions; full credit should require the candidate to anchor the inference in a specific detail, for example «قبل أن يستيقظ جيرانه».
2. Explain how a marker should apply the phrase 'supported by the text' consistently across 58 scripts.
Read the worked answer
Convert the phrase into a two-step test. First, locate the textual detail the answer depends on: dawn timing, repairing before neighbours wake, no request for reward. Second, ask whether the stated lesson is a reasonable generalisation from that detail or an invention. So 'help others quietly' passes; 'he was shy' fails because no detail supports avoidance of people; 'he liked water' fails because the detail is about the channel, not preference. A useful marking shorthand is: detail plus lesson = full credit; lesson with a plausible but unstated detail = partial; lesson contradicted by the text = zero. Apply the same shorthand to all scripts and note the detail candidates cite, so the standard is visible at re-check.
3. Propose a revised wording or rubric so that this item discriminates fairly in the board paper.
Read the worked answer
Make the inference task explicit and give part of the credit for evidence. One option is to reword Question 4 as «ماذا نتعلّم من سلوك الفلاح؟ اذكر عبارة من الفقرة تدل على إجابتك» (state the lesson and cite a phrase), with 2 marks for the lesson and 1 for the citation. Another is to keep the question but print band descriptors on the key: 3 = valid lesson plus textual clue; 2 = valid lesson without a clue; 1 = weak but relevant attempt; 0 = contradicted or blank. The reworded version is better because it teaches the evidence habit and reduces disagreement between markers, and because 19 blank answers suggest candidates did not understand what kind of answer was wanted.