Causal Claim about Wi-Fi Slowdown
A student writes this paragraph in a 250-word essay: 'The campus Wi-Fi slowed dramatically during exam week. On Monday of that week, the IT office installed a new access point in the library. Therefore, the new access point caused the slowdown. To fix the problem, the office should remove the access point.' The student also cites a social-media poll in which 18 of 25 respondents said Wi-Fi was worse that week; the poll was posted in a library study group. The essay adviser notes that exam week also brings a large increase in users, that the new access point was installed in a different building from most reported problems, and that the poll respondents were self-selected. The adviser asks the student to separate what the evidence supports from what it does not.
1. Identify the main causal flaw and one sampling flaw in the argument. Explain how each weakens the conclusion.
Read the worked answer
The main causal flaw is post hoc ergo propter hoc combined with confounding. The new access point was installed on Monday, and the slowdown occurred during exam week, but exam week also brought a large increase in users and traffic. The increased demand is a plausible common cause of the slowdown, so the timing alone does not show that the access point caused it. The sampling flaw is self-selection and narrow reach: 18 of 25 respondents is a small, non-random group recruited from a library study group, so it does not represent all campus Wi-Fi users. People who respond to such a poll are likely those who had a bad experience, which inflates the apparent severity.
2. Describe two pieces of evidence the student could collect to test the causal claim more fairly.
Read the worked answer
First, collect before-and-after throughput and latency measurements in the affected buildings and in the library, at the same times of day, with user counts recorded. If the slowdown appears in buildings far from the new access point and scales with user numbers, the access point is a poor explanation. Second, run a controlled test: temporarily disable the new access point during a high-demand period and compare performance with it enabled, ideally repeating on several days. A random sample survey of users, rather than a self-selected poll, would also test whether the problem is widespread.
3. Rewrite the conclusion as a cautious claim that the evidence supports, and state what would still be needed before recommending removal.
Read the worked answer
A cautious conclusion: 'The evidence is consistent with several explanations, including increased exam-week demand; the new access point may have contributed, but the current evidence does not establish that it caused the slowdown.' Before recommending removal, the office should run the controlled on/off test and measure whether removal actually improves performance without shifting the problem elsewhere. Removal would be justified only if disabling the access point repeatedly improves throughput or latency under comparable load.