The Sentence That Proved Too Much
The school magazine's science page carries this draft, written by a first-year student, summarising a study reported in a news digest:
'Researchers fitted 40 urban foxes with tracking collars and recorded when the animals were active over eight weeks in and around one city. In streets with brighter night-time lighting, foxes were recorded as active later after sunset. The authors write that "the pattern is consistent with foxes shifting their activity to avoid people, although refuse availability was not measured."'
The student's draft sentence reads: 'The study proves that bright street lighting forces urban foxes to become nocturnal.'
The editor's checklist states three rules: every claim must match the strength of the evidence; causal language needs an experiment; and any variable the original authors explicitly left unmeasured must be acknowledged in the summary. The editorial meeting is in twenty minutes.
1. Identify the words in the draft that overstate the evidence and explain what the study design actually supports.
Read the worked answer
Two words carry the excess.
'Proves' claims certainty from a single observational study of 40 animals in one city over eight weeks, with no replication and no statement of effect size or uncertainty. Observational data can establish association and can rule out some explanations, but they cannot prove a mechanism.
'Forces' is a causal verb attributing the behaviour to lighting, whereas the researchers hedged with 'consistent with'. The design recorded the timing of activity in streets that already differed in lighting; it did not manipulate lighting, did not randomly assign foxes or streets, and did not measure refuse, a plausible common cause, since bright streets may also contain more food waste. The data support only this: foxes in brighter streets tended to be active later after sunset. Note also that 'become nocturnal' is unsupported: the study recorded timing of activity, not a switch from day to night.
2. Rewrite the sentence in no more than 25 words, keeping the finding and acknowledging the unmeasured variable.
Read the worked answer
A version of 20 words:
'In brighter streets, urban foxes were recorded active later after sunset, though food availability from refuse was not measured.'
What makes it pass the checklist: the verb 'were recorded' describes the observation rather than a mechanism; the comparison 'brighter streets' preserves the actual finding, including its correlational form; and the final clause names the confounder the authors themselves flagged. It stays under 25 words and drops 'proves', 'forces' and 'become nocturnal'. A weaker alternative such as 'lighting may affect fox behaviour' is not wrong, but it is less informative because it removes the specific, checkable detail about timing that the study did establish.
3. The editor wants one follow-up sentence proposing a study that could support a causal claim. Write it, and state what result would weaken the lighting explanation.
Read the worked answer
Proposed sentence: 'A follow-up could change lighting on matched streets, comparing fox activity before and after the change against control streets whose lighting is untouched, while logging refuse.'
Why this works: manipulating lighting removes the common cause, and the before-and-after comparison within a street holds constant the food supply, street layout and fox population, provided control streets are included to absorb seasonal change.
The result that would weaken the lighting explanation: if fox timing shifted by the same amount on the control streets, where lighting was unchanged, but shifted in step with refuse volume, then the lighting association would look like a by-product of food, which the original study left unmeasured. Equally, if the lighting change produced no shift while refuse changed, the causal claim would fail outright.