The shape of the trap
Every true-but-irrelevant answer has the same anatomy. It is a real fact, pulled straight from the passage or the data, welded onto a claim or a goal it does not serve. Nothing about it is false. You can point to the line that proves it. That is the point: the test wants you to confirm that it is true, feel the click of recognition, and stop reading.
This is a different animal from the other wrong-answer families. An out-of-text answer is unsupported: it says something the passage never said. A too-strong answer overreaches with an always or a proves. True-but-irrelevant does neither. It is fully supported and perfectly measured. It just answers the wrong question. That is what makes it the hardest of the wrong-answer types for careful students specifically, because careful students are the ones checking whether the answer is true, and true is not the test.
Why accuracy fools you
Your brain runs a shortcut under time pressure: is this true? yes, then it must be right. Most of school rewards recall, so that shortcut usually works. The SAT breaks it on purpose. The question is never "is this choice true," it is "does this choice do the specific job the stem named." Those are two different tests, and the trap lives in the gap between them.
The behavioral root is anchoring. You read the four choices before committing to what a correct answer must do, and a true statement lights up as familiar the instant you see it. Familiar feels like right, so you match on truth instead of on task, and the trap, engineered to be both true and off-target, wins. The deeper miss underneath it is the wrong-task leak: you solved cleanly, you just solved for the wrong target, and that quietly bleeds wasted points on questions you actually knew how to do.
The move: re-read the stem before the choices
The whole fix is one habit, and it costs about five seconds. Before you read a single answer choice, re-read the stem and write down what the answer must do in two or three words. Not what the passage says. What the correct choice has to accomplish.
This is Predict-Then-Peek aimed at the task instead of the answer. On Command of Evidence, the stem hands you a specific claim and asks which choice most directly supports it. So name the claim first. "Supports: moths navigate by moonlight." Now that phrase is your filter, and every choice gets tested against it, not against "is this a real fact." A true statement about the moth's wing color fails the filter instantly, no matter how correct it is.
People worry the re-read costs time they do not have. Weigh it honestly. The five seconds you spend naming the task is cheaper than the far longer stretch you lose talking yourself into a true-but-off-claim choice, then re-reading, then second-guessing. Skipping the stem is not faster: it is the reason the time-sink shows up two questions later, disguised as a careless miss. Naming the task up front is the same discipline as the Receipt Rule: you do not accept a choice until you can say exactly which claim it supports.
The wrong-task variants
The trap wears a few costumes. Learn the three you will actually meet.
Off-claim (Command of Evidence, text). The passage makes several true statements. The stem asks which one supports one specific claim. The trap is another true statement from the same passage that supports a different claim. Both are real; only one does the assigned job.
Irrelevant data (Command of Evidence, quantitative). On a quantitative evidence item, the trap cites a real number from the table, read correctly, but for the wrong row or entity. The number is accurate; it just describes the wrong thing. That separates it from a plain data-misread, where the student grabs the wrong number outright.
Serves a different goal or wrong audience (Rhetorical Synthesis). On rhetorical synthesis items, the stem gives you a goal ("emphasize a contrast"). The trap accomplishes a real rhetorical purpose, cleanly built from the notes, but not the stated one. It stresses a similarity when you were told to stress a difference, or it is pitched at the wrong reader. Right craft, wrong assignment.
Original examples
Command of Evidence, off-claim. A passage argues that a desert moth navigates by polarized moonlight. The stem asks which finding best supports that navigation claim.
- A trap choice: "The moth's wings reflect ultraviolet light visible to predators." True in the passage. It supports a claim about predation, not navigation. Off-claim.
- The credited choice would report that moths released under a filtered sky that scrambled polarization lost their direction. That one does the navigation job.
Command of Evidence, irrelevant data. A table lists nitrogen levels for four lakes. The stem asks for evidence that Lake Maren's nitrogen rose after 2015.
- A trap choice: "Lake Osel measured 3.1 milligrams per liter in 2015." That number is really in the table, read correctly. It is the wrong lake. Irrelevant data.
Rhetorical Synthesis, wrong goal. The notes cover two bridge-building methods. The goal sentence: emphasize a difference between them.
- A trap choice combines the notes accurately to stress what the two methods share. Real synthesis, real facts, wrong goal.
Notice the pattern across all three. In every case you can defend the trap as "true." In every case it flunks the two-word task you would have written down.
The self-test you can run today
Pull your last ten Command of Evidence and Rhetorical Synthesis misses. For each, write the stem's task in two or three words: the claim, the number's target, or the goal. Now look at the choice you picked and ask one question: was it false, or was it simply true-and-off-task? Tally the off-task ones. If they are the bulk of your misses, your problem is not knowledge or reading speed. You are grading choices on "is it true" instead of "does it do the job," and the fix is the five-second stem re-read, every time.
Practice it now: run the free Command of Evidence drill. It makes you name the claim before it shows you the choices, so the true-but-irrelevant traps have nothing to anchor on. No signup, works as a guest.
The filter, in one line
Stop asking "is this choice true." Ask "does this choice do the exact job the stem named." Truth is the bait. The task is the answer. Once you re-read the stem, name the job, and test every choice against it, the accurate-but-off-target choice stops looking like a right answer and starts looking like what it is: a correct answer to a question you were never asked. Pair this with first-instinct override discipline on the review screen and you close the last door the trap uses.