Vocabulary and Word Use: Collocation, Hedging, and One Noun
A candidate is writing a 150-word academic summary of a cross-sectional, self-report survey of 300 nurses. Her draft contains four decision points. (1) 'The researchers ___ a questionnaire about sleep habits.' She hesitates between made, conducted and did. (2) 'The findings ___ that night-shift nurses sleep less than six hours.' Options: prove, show, suggest. The study used self-reports only, with no clinical measurement and no follow-up. (3) 'Caffeine intake rose by 12% (p = 0.03), a ___ increase.' Options: significant, considerable, important. (4) In her topic sentence she writes 'the affect of shift work on sleep'. Her teacher's only written feedback is 'Mind your collocation and watch your hedging.' She has one revision pass and eight minutes.
1. Resolve decision point (1) and explain what kind of knowledge is being tested.
Read the worked answer
'Conducted' - 'The researchers conducted a questionnaire about sleep habits.'
This is a collocation item, not a meaning item: made, did and conducted are near-equivalent in sense, but only one combines naturally with this noun in academic prose. Two refinements worth knowing: 'administered a questionnaire' is the most precise choice when you focus on distributing it to respondents, while 'conducted a survey' is the usual pairing when the instrument is a survey rather than a questionnaire.
Did is not wrong so much as register-marked - it belongs to speech and informal writing. Made a questionnaire is the classic false step; it can only mean the physical act of creating the document, not running it.
What is being tested: collocation is convention rather than logic, so it has to be stored in chunks - conduct research, administer a questionnaire, carry out a survey, draw a conclusion, make a comparison. When two options are both grammatical, the exam is asking which one a published paper would print.
2. Resolve decision point (2), then decide between 'significant', 'considerable' and 'important' in (3).
Read the worked answer
(2) 'Suggest'. The design is a single self-report with no clinical measurement and no follow-up, so it cannot establish proof and it cannot firmly establish causation. 'Suggest' matches the evidential strength exactly. 'Prove' is an overclaim about method, and 'show' is borderline - acceptable for describing a pattern in your own data ('the results show'), but risky when the claim is causal.
Hedging is not weakness; it is calibration. Examiners reward a candidate whose verbs match the strength of the design.
(3) 'Statistically significant' is the precise choice, because p = 0.03 is below the 0.05 threshold. The judgement here is that bare 'significant' is ambiguous in academic writing: it can mean statistically detectable or merely large. Since the test tells you p = 0.03, state it fully: 'a statistically significant but modest 12% increase.' 'Considerable' makes a claim about size, which 12% may not support, and 'important' is a value judgement the study cannot make.
With eight minutes left, fix (2) and (3) first: mis-stated certainty and a mis-signalled statistic are substantive; a clumsy collocation is cosmetic.
3. Resolve decision point (4) and explain how to keep 'affect' and 'effect' apart under time pressure.
Read the worked answer
Correction: 'the effect of shift work on sleep'.
'Affect' is normally the verb - caffeine affects sleep - and 'effect' is normally the noun - the effect of caffeine on sleep. A shorter version of the same sentence would be 'shift work affects sleep', where the verb is required.
Fast test: if the word follows 'the' or an adjective and is followed by 'of', it is a noun slot, so write effect. If it carries tense (affects, affected, affecting), write affect.
The one exception to keep in mind: in psychology, 'affect' is a noun meaning an observed emotional state ('a flat affect'). That reading is irrelevant here, because the sentence is about influence, not emotional display.
Revision order for the eight minutes: correct 'effect', change 'prove' to 'suggest', then add 'statistically' before 'significant'. Three edits, all substantive, and no rewriting required - which is what 'one revision pass' should mean.