MCAT common mistakes
MCAT · Understand the traps and how to avoid them.
Reading a passage once for gist, then answering questions by recalling general knowledge instead of locating the sentence or data point that decides the answer.
Under time pressure, test-takers default to memory because re-reading feels slow, but MCAT questions are deliberately written so that plausible prior knowledge conflicts with what the passage actually states.
What to do instead
For every question, name the passage location (paragraph, figure, or table) that supports your choice before selecting. If you cannot point to it, treat the question as unanswered and return to it. Practice by annotating a one-line function for each paragraph during the first read.
Misreading the question stem's task word - answering 'which is LEAST likely' or 'which does NOT support' as if it asked for the true statement.
Stems are long and the negation is often buried mid-sentence; the brain tends to encode the topic, not the logical operator.
What to do instead
Circle or mentally bold the operator (EXCEPT, NOT, LEAST, WEAKENS) and restate the task in your own words before evaluating options. On flagged review, re-read the stem first, not the choices.
Treating the four answer choices as independent and evaluating each on its own merits rather than as a set of contrasts.
Choices are engineered so that two are near-misses differing by one word (a direction, a magnitude, a causal vs. correlational claim), and isolated evaluation makes both feel correct.
What to do instead
After identifying the two strongest contenders, state aloud the single dimension on which they differ and check that dimension against the passage or data. Eliminate by contrast, not by familiarity.
Solving physics and chemistry problems by formula-matching before writing down the knowns, unknowns, and units.
Passages supply extraneous numbers, so the first formula that 'fits the topic' often uses the wrong quantities; unit mismatches then go unnoticed.
What to do instead
Before computing, write a one-line inventory: given quantities with units, target quantity with units, and the principle that links them. Estimate the answer's order of magnitude first so a units or exponent error is caught immediately.
Defaulting to a remembered equation or mechanism instead of the passage's stated conditions when they conflict (for example, assuming ideal behavior, standard temperature, or a stated equilibrium when the passage describes otherwise).
Coursework rewards rapid recall, and many passages deliberately place a system outside textbook-standard assumptions to test whether the examinee reads the setup.
What to do instead
Highlight any condition words in the passage - 'nonideal', 'at equilibrium', 'in the presence of', 'assuming complete dissociation' - and check each against your recalled model before applying it.
Interpreting experimental results as causal when the design is correlational or lacks a control, and choosing the answer that states a mechanism.
Answer choices that offer a tidy causal story feel more satisfying than ones that merely describe an association, and the design details are easy to skim.
What to do instead
For every study described, explicitly label it: randomized experiment, observational, or quasi-experimental. Only randomized designs with a proper control license causal language; otherwise select the associational option.
Ignoring base rates and sample-size information and answering probability or prevalence questions from the sensitivity/specificity numbers alone.
Sensitivity and specificity are the salient numbers in the table, while prevalence sits in a sentence or a footnote and is treated as background.
What to do instead
When a question involves a diagnostic test, write out the 2x2 table with a hypothetical population of 1,000 at the stated prevalence, then compute the requested conditional probability from the table.
Applying a psychological or sociological concept by label rather than by its defining features - e.g., calling any group-influenced behavior 'conformity' regardless of whether the response was public or private.
Social-science terms are familiar from everyday language, so surface similarity to a scenario substitutes for checking the concept's operational definition.
What to do instead
For each social-science question, write the two or three necessary conditions of the concept, then test the scenario against each condition. Distinguish paired concepts explicitly: conformity vs. obedience, prejudice vs. discrimination, classical vs. operant conditioning.
Mixing levels of analysis - explaining a population-level pattern with an individual-level mechanism, or vice versa - and selecting the distractor that sounds sociologically sophisticated.
Both micro and macro explanations appear among the choices, and the stem's level is often implied rather than stated.
What to do instead
Before answering, decide whether the question asks about individuals, groups, or societies, and eliminate options pitched at the wrong level. Watch for the words 'individual', 'population', 'institution', and 'society' in the stem.
Running out of time because every question receives equal effort, leaving whole passages unread at the end.
Early questions feel solvable with more thought, so minutes are sunk into a few hard items while easy later items go unseen.
What to do instead
Set a per-passage time budget during practice and enforce a hard stop: answer, flag, and move on. Build in a final two-to-three-minute sweep for flagged items, and never leave a blank - eliminate and guess rather than run out of clock.