Inference and Implication: Taking Exactly One Step
Inference and Implication questions ask for a conclusion the passage supports but does not state outright. They are worded “suggests,” “implies,” “most likely agrees,” “would probably,” or “can be inferred.” The DAT deliberately separates a warranted inference-one short step from stated evidence-from a plausible guess that requires outside knowledge.
The one-step rule
Find the sentence or sentences anchoring the question, then take the smallest logical step. A statement needing two or three unstated assumptions is almost always wrong. If the passage says a drug reduced biofilm in laboratory cultures but that no clinical trials have been run, the warranted inference is “its effectiveness in patients is not yet established,” not “the drug does not work.”
Signals to read closely
- Contrast and concession: “although,” “however,” “yet,” “admittedly.” The author's real position usually follows the concession.
- Hedges: “may,” “appears,” “preliminary,” “in some cases.” These limit how strong any inference may be.
- Loaded word choice: calling a study “preliminary” or a claim “contested” signals skepticism that an item may target.
- Numbers and comparisons: if group A improved by 30 percent and group B by 12 percent, you may infer A outperformed B, not that A is clinically superior.
Worked example
Passage: “The sealant was applied to permanent molars in 200 children; retention at 24 months was 84 percent, comparable to published rates for resin-based sealants, though the study did not include a control group.” Which conclusion is best supported? Answer: the sealant's short-term retention resembles existing products, but its advantage over no treatment is unproven. The 84 percent figure supports the comparison; the missing control group supports the second half. “The sealant prevents caries better than resin sealants” fails-retention is not caries prevention, and no comparison group exists.
Traps that cost marks
- Extreme language: “proves,” “always,” “never,” “the only.” Correct inferences are usually modest.
- Reversal: swapping cause and effect, or who did what to whom.
- Outside knowledge: true in the world but not derivable from the passage.
- The “could be true” option: several choices may be possible, but only one is required by the text.