Research Methods & Study Design

About a quarter of every MCAT science section tests Scientific Inquiry & Reasoning Skills (SIRS) — your ability to evaluate how an experiment was done, not just recall facts. Mastering study design is the highest-leverage skill for passage-based questions.

Study designs, from weakest to strongest causal evidence

• Case report / case series — descriptive; hypothesis-generating only.
• Cross-sectional — measures exposure and outcome at one time; shows association, not causation; vulnerable to reverse causation.
• Case-control — starts from outcome, looks backward at exposure; efficient for rare diseases; reports odds ratios; prone to recall bias.
• Cohort — starts from exposure, follows forward to outcome; reports relative risk; can establish temporality.
• Randomized controlled trial (RCT) — random assignment balances confounders (known and unknown); the gold standard for causation. Blinding (single/double) controls for placebo and observer effects.

Variables and controls

• Independent variable is manipulated; dependent variable is measured.
• Confounding variable is associated with both exposure and outcome and distorts the relationship — controlled by randomization, restriction, matching, or statistical adjustment.
• A positive control confirms the assay can detect an effect; a negative control confirms the signal isn't an artifact.

Validity and bias

• Internal validity — does the study correctly measure the relationship within its sample? Threatened by selection bias, measurement bias, and confounding.
• External validity — do results generalize to other populations?
• Common biases: selection bias (non-representative sampling), recall bias (differential memory of exposure), observer/measurement bias (corrected by blinding), attrition bias (differential dropout).

How MCAT questions use this

You'll be given a figure or methods paragraph and asked: what would strengthen the conclusion?, which control is missing?, or what alternative explanation does the design fail to rule out? Always identify the design first, then its characteristic weakness.