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

Classify designs from descriptions and decide which causal inferences their structure actually supports.

  • Confuses the directionality of case-control and cohort designs
  • Believes a strong association in a cross-sectional study supports causation

Sampling Methods (Random, Stratified, Cluster, Convenience)

Different sampling strategies introduce different threats to generalizability and selection bias.

  • Conflates simple random sampling with stratified random sampling
  • Confuses the external validity threat of convenience sampling with an internal validity problem

Variables (Independent, Dependent, Control, Confounding)

A true confounder must be linked to both exposure and outcome — not just one.

  • Identifies any outcome-associated variable as a confounder without checking its association with the exposure
  • Conflates control variables with confounding variables

Measures of Central Tendency (Mean, Median, Mode)

Skewed data pulls the mean away from the median in a predictable, testable direction.

  • Reverses the direction of mean shift in right-skewed distributions
  • Defaults to the mean as the best central tendency measure regardless of data distribution

Measures of Spread (Variance, SD, Range, IQR)

Choosing between SD and IQR depends on whether the distribution is symmetric or skewed.

  • Treats variance and standard deviation as equivalent measures with the same units
  • Applies SD to skewed data instead of IQR

Reliability vs Validity

Reliability guarantees consistency; validity does not automatically follow from it.

  • Assumes reliability guarantees validity
  • Conflates test-retest reliability with inter-rater reliability

Types of Bias (Selection, Recall, Observer, Hawthorne)

Recognize the dominant bias in a design and match it to the correct mitigation strategy.

  • Conflates observer bias (researcher-side) with the Hawthorne effect (participant-side)
  • Applies blinding as the remedy for selection bias instead of randomization

Statistical Significance and p-Values

The p-value is not the probability the null is true — it assumes the null is true.

  • Misinterprets the p-value as the probability that the null hypothesis is true
  • Equates statistical significance with clinical or practical importance
See how your Anki deck covers these topics.Upload for a free audit →

Confidence Intervals

For ratio measures, a CI excluding 1.0 signals statistical significance at the chosen alpha.

  • Misinterprets a 95% CI as a 95% probability statement about a single computed interval
  • Reverses the significance interpretation when a CI for a ratio excludes 1.0

Type I and Type II Errors; Statistical Power

Lowering alpha reduces false positives but increases false negatives and reduces power.

  • Swaps the definitions of Type I and Type II errors
  • Believes reducing α simultaneously reduces both Type I and Type II errors

Effect Size and Clinical Significance

Large samples can make trivially small differences statistically significant — effect size cuts through that.

  • Conflates statistical significance (p-value) with effect magnitude or clinical importance
  • Assumes large-sample statistical significance implies clinical or practical significance

See how your Anki deck covers Research Methods and Biostatistics.

Upload your deck for a free audit →