Renée's Test-Optional Op-Ed and the 62 Dining-Hall Surveys

What she may claim: that among 62 first-year students approached outside one dining hall, 71 percent agreed with a statement about the tests not reflecting their ability. That is a description of her sample, and nothing more.

What she may not claim: that "most first-years feel this way," and certainly not that the tests are invalid. A belief about a test is not evidence about a test's predictive validity; those are different claims that require different data, such as predictive studies comparing test scores and first-year GPA.

Two problems with the sample. First, non-probability selection: diners at one location and time are not a random draw, and people with strong opinions are more likely to stop and answer. No margin of error can repair a non-random sample.

Second, even under ideal random sampling, 62 responses out of a first-year class of thousands yields a margin of error near 12.5 percentage points (roughly 1/√62). So 71 percent could plausibly be anywhere from about 58 to 84 percent - a range that still would not settle whether the policy should change.

Third, the question is leading: "do not reflect my ability" invites agreement. A neutral version would ask students to rate how well scores represent their preparation.

The assumption: that the 18 percent application increase was caused by the test-optional policy. The comparison supplies a sequence - policy, then increase - and treats sequence as causation.

Rival explanations that fit the same facts: national application inflation (many universities saw double-digit growth in that period); a new marketing campaign or fee waiver; a new major or online program; a change in the Common Application; a rankings shift; or a tuition freeze. Any of these could produce 18 percent without the policy doing anything.

Evidence that would test it:

• A difference-in-differences comparison - the peer university's increase minus the average increase at comparable institutions that stayed test-required over the same years. If peer schools also rose 18 percent, the policy explains nothing.
• The composition of the increase: did it concentrate among applicants without scores, or was it uniform? A policy effect should show up disproportionately in the group the policy was meant to reach.
• Enrollment and graduation outcomes for the first cohorts, since applications are an input, not a result.

Renée can keep the comparison if she reports it as correlation and says so plainly. Written as correlation, it still supports a modest claim: this is a policy other institutions have adopted without evident collapse in applications.

Original: "Standardized tests are useless and should be abolished nationwide."

Three problems: "useless" is a global absolute that no evidence could establish and no reader could vote on; "nationwide" exceeds the senate's jurisdiction and Renée's evidence; and "should" asserts a conclusion without naming the criterion that makes it follow.

Revision: "Because a peer institution's move to test-optional coincided with an 18 percent rise in applications, and because students in our own survey report that scores do not capture their preparation, this university should make SAT and ACT submission optional for first-year applicants starting in fall 2027 - and report on the first two admitted cohorts before the change is made permanent."

Function of each element:

• The because-clauses attach the claim to the evidence she actually has, and their wording ("coincided," "report") keeps her from overstating causation.
• "This university" brings the claim inside the body that is voting.
• "Optional" is a specific, votable policy rather than an abolition.
• The pilot dates and reporting clause build in a test of her own claim, which is what makes the thesis arguable rather than asserted.

That last clause is also the practical move with the admissions director: he can oppose abolition and still vote for a time-limited pilot with published outcomes.