Item 3 on remote work: upholding an appeal and spotting a defective key
A teacher in Valladolid is marking the June mock of the Portuguese PAU. The reading passage is an original ninety-word article on remote work in Portugal. Item 3 asks: «O que se pode inferir sobre a atitude das chefias?» The keyed option is C, «As chefias continuam a preferir o contacto presencial para avaliar o desempenho». Two students appeal. Student A chose B, «O teletrabalho aumentou a produtividade das empresas portuguesas», arguing that the text states that seven workers in ten feel more productive. Student C chose the keyed option but writes in the margin: «não é inferência, está no texto». The relevant lines read: «Num inquérito de 2023 junto de 1 200 trabalhadores, sete em cada dez afirmaram sentir-se mais produtivos em casa; as chefias, por seu lado, continuam a preferir o contacto presencial para avaliar o desempenho.» The teacher has twenty minutes of department meeting to settle both appeals and to decide whether the item should be rewritten before the bank is reused next year.
1. Student A appeals option B. Should the appeal be upheld? Cite the words that decide it.
Read the worked answer
Reject the appeal. The sentence is: «sete em cada dez afirmaram sentir-se mais produtivos em casa». Three elements do the work. afirmaram marks reported speech: the article reports a belief, it does not assert the fact. sentir-se marks subjective perception: feeling productive is not being more productive, and nothing about output is measured anywhere in the passage. And the grammatical subject is trabalhadores, not empresas: option B silently swaps the group. B therefore needs two unlicensed steps - from perception to fact, and from workers to firms - which is exactly what «inferir» does not allow. A rule worth writing on the board: an option can be plausible in the real world and still be wrong if the text does not license it. Concede to A that productivity is the passage's topic; that is precisely why the trap works.
2. Student C chose the keyed option but says it is «no texto, não é inferência». How do you respond, and what does this reveal about the item?
Read the worked answer
C is right about the option, wrong about the reason, and the reason exposes a defect in the item. The keyed sentence - «as chefias… continuam a preferir o contacto presencial para avaliar o desempenho» - is almost a verbatim restatement of the last clause. A stem beginning «O que se pode inferir» cannot be answered by copying; if the keyed option is explicit, the item measures scanning, not inference. Marking: award C the point, because in this format the option is what is scored, but log the item. Before reuse, take one of two routes. Either change the stem to «Segundo o texto, …», which makes the key honest and leaves B wrong for the same reasons; or keep the stem and build a key that requires a step, e.g. «As chefias desconfiam de indicadores que não possam observar directamente.» The second route is harder to write and is the only one that measures the skill the stem names.
3. Rewrite B so that it becomes a legitimate inference, and add one distractor that is textually true but does not answer the stem.
Read the worked answer
Legitimate inference: «Parte dos inquiridos associa o teletrabalho a um aumento da sua própria produtividade.» This keeps the hedges (parte, associa, própria) and stays inside the evidence. Textually true but irrelevant to a stem about chefias: «O inquérito de 2023 envolveu 1 200 trabalhadores.» It is verifiable in the passage, so weaker candidates choose it, and it tests whether they read the subject of the question at all. A second useful distractor: «Os trabalhadores queixam-se do isolamento» - true, but it answers a question about employees, not managers. Keep both; they cost nothing and separate scanning from reading the stem.