What these questions actually ask
Most Reading and Writing questions ask what a passage says. These ask something else entirely: what a piece of text is doing. That's the whole shift, and once it clicks, the category gets a lot less confusing. A structure question points at an entire short passage and asks how it's built. Does it open with a common belief and then push back on it? Does it lay out a problem first and a fix second? A function question zooms in tighter. It underlines one sentence and asks what job that sentence is doing right there, not what it means sitting on its own.
Picture a sentence that just says, "Not everyone agreed." What it says is thin. Two people disagree, that's it. What it does depends entirely on where it sits. Right after a bold claim, it's pushing back. Right after a list of supporters, it's introducing an exception. Same four words, a different job each time, depending on what comes before it. That's the whole test, really, packed into one short line.
A bank of fluorescent lights flickers at a rate most people cannot consciously see. Vision researcher Anouk Delacroix tested whether the flicker still affects reading. Volunteers read passages under flickering and steady lights of identical brightness. Reading speed was the same in both conditions. Reported headache after an hour was not.
Which choice best describes the function of the last sentence of the text?
Structure questions: track the passage like a sequence of moves
Treat the passage like a sequence of moves, not a block of meaning. Read it sentence by sentence. After each one, jot a short label: claim, evidence, objection, example, conclusion. By the end, that string of labels is the structure. The right choice is just the one that matches it.
Try a short one. For decades, city planners treated rooftop gardens as a nice-to-have, good for photos but not much else. Then researchers measuring urban heat found something planners hadn't expected: roofs covered in plants ran noticeably cooler than bare ones, and the buildings under them used less energy on cooling. Planners now budget rooftop gardens as part of a building's energy system, not as landscaping. Label each sentence and the shape pops out fast: a dismissive assumption, a finding that undercuts it, a changed practice. A question asking how this is structured wants exactly that sequence, presents a common view and then shows evidence against it before describing the resulting change. A choice describing some other order, like two methods being compared, or a claim that never gets challenged, is wrong no matter how well it's worded.
Function questions turn on what sits right before and after
Function questions work differently. They point at one underlined sentence inside a longer passage and ask what that sentence is doing there, not what it means in isolation. Context decides the answer, so read the sentence right before the underlined one and the sentence right after it before glancing at a single choice. Skipping that step is probably the most common reason students miss these.
Try this passage instead. Mapmakers debated for years whether color on a map should mark elevation or political borders first. Some evidence suggests early colored maps used color mostly for decoration. By the eighteenth century, though, cartographers began shading consistently to show mountains and valleys. This shift mattered because travelers could judge how rough the terrain was from a glance instead of reading a key. Modern topographic maps still use that layered coloring, long after borders stopped being the main reason for color on a map. Underline the sentence that starts with "This shift mattered" and ask what job it's doing. It isn't restating the sentence before it. It's explaining why the change described there was significant. "Explains the significance of a change described in the previous sentence" is the sort of phrase a correct choice uses. A choice that says "describes how travelers read maps" sounds related but names the wrong action.
The traps that catch students who know the passage cold
The most common wrong answer just summarizes the sentence. It can restate the content with total accuracy and still lose, because the question never asked what the sentence says. It asked what the sentence does. Big difference. Any choice that reads like a plain recap is a signal to slow down and check it against the actual job, not the content.
A sneakier trap pairs a correct-sounding verb with the wrong object. A choice might say the sentence "illustrates a claim made earlier," and the verb fits fine. But if no claim actually sits earlier in the passage, or the sentence illustrates something else, that choice fails on the second half. Check both halves on their own: the action word, then what it's acting on. A third trap is judging the underlined sentence alone, cut off from its neighbors. A sentence can sound like an example by itself and turn out to be a rebuttal once you've read what comes right before it. Context isn't optional here. It's the whole method.
How to practice this without drowning in theory
Pull three or four short passages from anything real you're reading and label every sentence the way described above: claim, evidence, example, objection, conclusion. No question attached. Just the labeling. Once that feels fast and almost automatic, answering these questions on the real test gets a lot less tense, since the reading part is already done before the choices even show up.
If a practice test flags this as a weak spot, StudyPebble's practice by skill drills it on its own with an explanation after every question, instead of mixing it into a pile of unrelated reading questions. Missed a batch of these on a real Bluebook test? A screenshot upload at /mistakes/upload sorts misses by skill automatically, so a cluster of function-question mistakes shows up together instead of scattered across a results page.