Scanning Versus Reading: When to Do Which

The question types that reward scanning

Numbers, prices, dates and durations. Anything with a printed form. A Part 2 chart question about which option costs twelve dollars less is a scan: you are looking for two numerals in two columns, and nothing about comprehension helps you do it faster.

Proper nouns. Names of people, places, organisations and programmes are capitalised, which makes them the easiest things on a page to find. In Part 4, once your ledger tells you which speaker a question concerns, finding their quotation is a scan for a capital letter.

Distinctive concrete nouns. A Part 3 statement about a wheelchair ramp or a snow clearing contract usually has one place it can live, because concrete objects are named rather than paraphrased. When a statement contains a noun the passage would have no reason to rephrase, scanning is the right first move.

Confirming an answer you have already chosen. This is scanning's most underrated use. You decided the answer from comprehension; now you want the line that proves it. You know exactly what you are looking for, so the scan is quick and the confirmation is worth several seconds of insurance.

The question types that punish it

Implication. Stems of the form the writer suggests, it can be inferred or the writer's attitude toward the move is have no printed target. The letter never says the writer is reluctant; it says she has started the letter four times. There is no word to scan for, because the answer is a relationship between sentences rather than a sentence.

Attribution. Part 4 questions about who holds which view look scannable, because names are easy to find. The trap is that the position you are looking for genuinely appears in the article, often near a name, and often near the wrong one. A scan finds the position and hands you the nearest person, which is precisely the distractor design.

Paraphrased matching. Part 3 statements are written to avoid the paragraph's vocabulary. Scanning for their words finds nothing in the right paragraph and something in the wrong one. Part 3 is the part where scanning is not merely weak, it is a liability, because it produces a confident wrong answer rather than an empty search.

Tone and register. Questions about how the writer feels, or how formal a reply should be, are judged across a whole text. A scan has nothing to land on.

Telling them apart from the stem

You do not need to know the theory on test day. You need one question, asked before every search: can I say what this will look like on the page?

If you can finish the sentence I am looking for a... with something concrete, scan. A number. A name. A date. The word ramp. If the only way you can finish it is with something abstract, such as the reason she hesitated or the person who cares about safety, do not scan. Read the relevant stretch of text properly instead, and accept that it costs more seconds.

A second, faster proxy works in Part 3 specifically. Look at whether the statement shares vocabulary with the passage. If it does, the shared word is more likely a plant than a signpost, and you should translate the statement before searching. If it clearly does not, translation is obviously required and you have lost nothing by checking.

A third proxy applies to Part 1. If the stem half-quotes the letter, the order rule will place it and a short scan of the right paragraph will find it. If the stem paraphrases the letter, you are in implication territory and the read you did at the start of the part is what has to carry you.

The three things scanning cannot do

It cannot establish that something is absent. This matters most in Part 3, where E means the fact is in none of the four paragraphs. A failed scan is weak evidence of absence, because the fact may be present in words you did not scan for. Choosing E requires a comprehension pass over the candidate paragraphs, not four quick scans.

It cannot distinguish two similar things. Part 2 charts are built from four near-identical options. A scan for evening access finds evening access, and finds it in three columns. The work of deciding which column the question is about is comprehension work, and doing it before the scan is what makes the scan useful.

It cannot detect that you have answered a different question. This is the failure mode behind most wrong answers in the section. You scanned, you found something true, and it was true about the wrong question. The only defence is rereading the stem after you find your candidate answer, which takes about three seconds and is the highest-return three seconds in Reading.

A working combination

The productive way to hold these two skills is not to choose between them but to sequence them. Comprehension establishes where and what. Scanning retrieves and confirms.

In Part 1 that means reading the letter once so the implications are available, then scanning within the paragraph the order rule points at. In Part 2 it means understanding the person's constraints, then scanning the one named column. In Part 3 it means labelling the four paragraphs, translating the statement, then scanning the paragraph the label selected. In Part 4 it means building the ledger, then scanning for the name.

Notice that in all four cases the comprehension step comes first and is short, and the scan is narrow because the comprehension step narrowed it. A scan across a whole text is almost always a sign that a comprehension step was skipped.

Practising this is easier than it sounds. Take any reading question and, before you look at the text, write down in five words what you are about to look for. If you cannot write it, you are about to read. If you can, you are about to scan, and you now know exactly where to stop.

Practise it

Five questions on when to scan, when to read, and what scanning can never do for you.