The Library's 24-Hour Opening: Did It Cause the Rise in Use?
Last October the Rialto University library extended weekday opening from 08:00–22:00 to 24 hours and added Sunday hours. In the following spring semester, recorded book loans rose 12% and turnstile visits rose 9% compared with the same semester a year earlier. The director's report concludes that 'the extended hours caused the increase in use' and asks the senate to make 24-hour opening permanent at an annual cost of €180,000, calling the cost 'clearly justified by the extra use.' In the same semester the library launched a new discovery system and a hold-and-collect service; undergraduate enrolment also grew 6%, and the neighbouring town closed its public library for refurbishment. A student representative argues that the report proves nothing, while the director replies that the 12% rise is 'too large to be a coincidence.' The senate has one meeting before the budget is set.
1. Identify the central unstated assumption in the director's argument and assess whether the evidence supports the causal claim.
Read the worked answer
The argument runs: hours were extended → loans rose 12% → the extension caused the rise. The unstated assumption is that nothing else changed in the same period that would have raised loans or visits, or equivalently that usage would not have risen by a similar amount without the extension.
The evidence is a simple before/after comparison with no control group. Several documented changes occurred simultaneously: a new discovery system, hold-and-collect, 6% enrolment growth, and the closure of the town library. Each plausibly raises loans or visits. The report therefore establishes correlation and a plausible mechanism, not causation. The claim 'caused' is not supported at the strength asserted, although the extension may still have contributed.
2. Which rival explanation most seriously threatens the causal claim, and what data would distinguish it from the extension's effect?
Read the worked answer
The strongest rival is the new discovery system plus hold-and-collect, because these act directly on the measured outcome: they make catalogue items easier to find and borrow, so they raise loans even with unchanged hours. Enrolment growth and the town library closure mainly affect visits and are weaker for the loan figure.
Discriminating data:
- Loans per enrolled student, to net out the 6% enrolment growth.
- Loans by time of day, especially the 22:00–08:00 window; if the extension caused the rise, much of it should appear in newly available hours.
- Use of the discovery system and hold-and-collect, and loans through those channels.
- A comparison library without extended hours, or another term's data, to provide a counterfactual.
If the 12% rise is concentrated in evening loan requests through the new system and per-student loans are flat, the extension's causal role shrinks sharply.
3. Evaluate the director's reply that the rise is 'too large to be a coincidence.'
Read the worked answer
The reply confuses size with identification. A 12% rise is not evidence against confounding: 6% enrolment growth alone plus a new borrowing channel and a closed rival library could readily produce 12%. The very magnitude the director cites is what several simultaneous changes would predict.
A large effect can also reflect other biases: the comparison semester may have been unusually weak, and recorded visits may rise simply because more people enter the building for non-library reasons.
What would support the claim is not a bigger number but a sharper comparison: usage in newly opened hours, a matched control library, or per-capita usage adjusted for enrolment. Without such evidence, the €180,000 commitment is being justified by an effect size rather than by a causal estimate. The senate should request the by-hour and per-student data before making the opening permanent.