These are inference questions wearing a costume, so start with what an inference actually is on this test. It is not a creative leap or a clever real-world conclusion. It's a small, safe step that is required to be true based on what the passage states. The fastest way to check one is the Because test: you should be able to say, "I can infer this answer because the text explicitly states [specific phrase]." If you can't finish that sentence with concrete proof, the inference is invalid, no matter how reasonable it sounds.
The traps that catch good readers
Almost every wrong inference is a leap that's "too far," and it comes in three flavors. The extreme language trap takes a qualified statement ("some studies suggest a possible link") and hands back an absolute one ("the text proves there is a definitive link"). The real-world plausibility trap is the dangerous one — a conclusion that's perfectly logical in real life but requires an assumption the text never makes. (A company invests in automation and profits rise; the trap concludes it cut its workforce. Plausible, sure, but the passage says nothing about the workforce.) And the reverse logic trap flips the sentence: the text says "all the selected paintings were by Hudson River School artists," and the trap claims "all Hudson River School paintings were selected."
The Last Sentence First method
For completion questions specifically, you should read less, not more. Ignore most of the passage at first and work the blank like a logic puzzle. Five steps:
- Read only the final sentence — the one with the blank. That's your whole mission for the moment.
- Find the command word just before the blank. It tells you exactly what to do. "However / in contrast" means contradict the previous idea. "Therefore / thus / this suggests that" means draw a conclusion from it. "For instance / specifically" means give an example of it.
- Get the setup. Read the one sentence right before the blank and pull out the specific idea your command word is pointing at. Ignore the rest for now.
- Predict the answer in blunt, almost dumbed-down terms. "The setup says the plan was praised; the command word is 'however'; so the blank has to state a flaw or a failure." Simple predictions are hard to fool — a fancy wrong answer won't match them.
- Scan and match. Three choices will be noise. One will be a near-perfect fit for your prediction. Done.
The first sentences of these passages are usually packed with vivid, "interesting" detail designed to color your judgment before you reach the blank. A passage might open by calling a theory a "watershed moment," "elegantly explained," "brilliant" — so by the time you hit a "however," your brain is primed for a positive answer and underweights the turn. Last Sentence First skips the priming entirely.
Watch it cut through a trap
A passage praises the theory of punctuated equilibrium for three glowing sentences, then says: "Some recent genetic studies, however, have failed to identify a specific molecular mechanism that could drive such rapid change. This has led some scientists to propose that ______."
The command phrase is "this has led to propose that" — draw a conclusion. The setup is the genetics problem: no mechanism found. So the prediction is plain: "the theory has a weakness, or is incomplete because of this genetic issue." Now the choices sort themselves. "Punctuated equilibrium remains the most important and well-supported contribution" matches the glowing intro, not the setup — that's the trap for the top-to-bottom reader. "The theory has been largely discredited" is the extreme-language trap, because "failed to identify a mechanism" is not the same as "discredited." The one that says the theory "may not be fully explained by known genetic processes" is the nuanced match. The fog clears the moment you stop reading for the story and start reading for the function.
You're the forensic analyst, not the detective
Your job here isn't to be the brilliant detective who cracks the case with a flash of insight. It's to be the forensic analyst who points at the fingerprint and states the one conclusion it can support. That sounds less glamorous, and it scores better. The Because test is what proves your conclusion is a disciplined deduction and not a wild guess.
Inference is one of the five dimensions Forge reads in your work. As you move through a diagnostic, it can tell whether your misses come from leaping too far or from missing the command word entirely, which are two different fixes for what looks, on the answer key, like the same wrong answer.