MCAT topics → Research Methods and Biostatistics → Study Designs (Cross-Sectional, Case-Control, Cohort, RCT)

Study Designs (Cross-Sectional, Case-Control, Cohort, RCT)

MCAT trap: Confuses the directionality of case-control and cohort designs. Case-control studies start with disease status (cases vs controls) and look backward at prior exposures.

Study designs are one of the most reliably tested concepts in MCAT Research Methods — and the most consequential misconception is about causation: only RCTs reliably establish it. Prospective cohort studies do track exposure before outcome, which establishes temporality, but they cannot control for unknown confounders the way randomization does. Cross-sectional studies can't even establish temporality. The four core designs — cross-sectional, case-control, cohort, and RCT — each have a distinct logic: who gets recruited, when exposure and outcome are measured, and what direction time flows. The exam expects you to identify the design, evaluate its limitations, and know exactly what conclusions it can and cannot support.

The MCAT tests this at three levels. First, basic classification: read a passage description and name the design. Second, comparative reasoning: explain why one design is more appropriate than another, or identify a specific weakness like recall bias, selection bias, or inability to establish temporality. Third, causal inference: determine whether the study's conclusions are actually supported — and this is where most students lose points. A lot of passage questions will present a researcher's conclusion and ask you to evaluate it. If the study isn't an RCT, 'establishes causation' is almost never the right answer.

The tricky part is that these designs share surface features that students mix up. Cohort and case-control studies both involve exposed and unexposed groups OR diseased and non-diseased groups — but they run in opposite directions in time. Cross-sectional studies feel like they should be able to show causation if the correlation is strong, but they can't, ever. And even well-designed prospective cohort studies — which do track exposure before outcome — still can't control for unknown confounders the way randomization does. Keeping the directionality and the causal hierarchy straight is the whole game here.

Common misconceptions

Common mistake
Wrong: Case-control studies follow exposed vs unexposed groups forward to see who develops disease.
Right: Case-control studies start with disease status (cases vs controls) and look backward at prior exposures.
Case-control studies do NOT start with exposure groups and follow them forward — that's a cohort study. Case-control studies start by identifying people who already have the disease (cases) and people who don't (controls), then look backward to compare their prior exposures. The direction is: disease status first, exposure history second. If you remember that 'case' means diseased and the study works backward from there, you won't flip the two designs.
Common mistake
Wrong: A cross-sectional study can establish a causal relationship if the association is strong enough.
Right: Cross-sectional studies measure exposure and outcome simultaneously, so temporality cannot be established and causation cannot be inferred.
In a cross-sectional study, exposure and outcome are measured at the same point in time — you can't tell which came first. Without temporality (knowing exposure preceded outcome), you cannot establish causation, no matter how strong the association is. A strong correlation just means the two variables co-occur; it says nothing about which caused which or whether a third variable is driving both.
Common mistake
Wrong: Prospective cohort studies can establish causation because they follow participants forward in time.
Right: Only RCTs reliably establish causation because randomization controls for known and unknown confounders; cohort studies can only show association.
Prospective cohort studies do follow participants forward in time, which means you can establish that exposure came before the outcome — but that's not enough for causation. The missing piece is confounding control. Cohort studies can't ensure that exposed and unexposed groups are equivalent on all other variables, including ones the researchers didn't think to measure. Randomization in RCTs distributes both known and unknown confounders evenly across groups, which is why only RCTs can reliably support causal claims.
Common mistake
Wrong: Cohort studies are highly susceptible to recall bias because participants must remember past exposures.
Right: Recall bias is a primary concern in case-control studies, not cohort studies, because case-control designs ask participants to retrospectively recall exposures after disease status is known.
Recall bias happens when participants try to remember past exposures after already knowing their disease status — sick people dig harder into their memory for potential causes than healthy people do. This is exactly what case-control studies ask participants to do. Cohort studies measure exposure at the start, before anyone develops the outcome, so there's no disease-status-driven distortion in how exposures are recalled. Cohort studies have other limitations (like loss to follow-up), but recall bias is the case-control problem.
Free Deck audit

See if your Anki deck covers this topic.

Upload your deck →
Guided session

Stuck on this? An AI tutor that probes your understanding.

Start a session →

What the exam tests

  1. Given a study description, correctly classify it as cross-sectional, case-control, cohort, or RCT based on how participants are recruited and when exposure and outcome are measured.
  2. Compare the strengths and limitations of each design — including directionality (which direction time flows), ability to establish temporality, susceptibility to confounding, and which design is best suited for rare diseases, rare exposures, or causal claims.
  3. Read a passage describing a study and identify its specific methodological limitations, such as selection bias, recall bias, or inability to establish that exposure preceded outcome.
  4. Determine which causal conclusions a study's design can actually support — recognizing that only RCTs reliably establish causation because randomization eliminates both known and unknown confounders.

Can you avoid these mistakes?

A researcher identifies 200 patients with lung cancer and 200 patients without lung cancer, then surveys both groups about their smoking history over the past 20 years. What type of study is this, and what is its primary bias concern?
A study enrolls 5,000 healthy adults, records their dietary habits, and follows them for 10 years to see who develops type 2 diabetes. A cohort study like this finds a strong association between high sugar intake and diabetes. Can the researchers conclude that high sugar intake causes diabetes? Why or why not?
You're reading an MCAT passage where researchers surveyed a population at a single time point about both their current stress levels and their current blood pressure. They found a significant positive correlation. What design is this, and what specific limitation prevents a causal conclusion?
Rank the following from weakest to strongest in terms of ability to support causal inference, and explain the key reason for each ranking: cross-sectional study, prospective cohort study, randomized controlled trial, case-control study.

Related topics

Sampling Methods (Random, Stratified, Cluster, Convenience) Variables (Independent, Dependent, Control, Confounding) Measures of Central Tendency (Mean, Median, Mode) Measures of Spread (Variance, SD, Range, IQR)

See how your Anki deck covers this topic.

Upload your deck for a free audit →