A population is the full group a researcher wants to understand; a sample is the subset actually studied. Sampling methods — random, stratified, convenience, snowball — differ in how well they produce representative samples, which determines the external validity of findings. A major criticism of psychology research is over-reliance on WEIRD samples (Western, Educated, Industrialized, Rich, Democratic), limiting generalizability. Sample size affects both the precision of estimates and statistical power.
Compare the same research question studied with a convenience sample versus a nationally representative sample. Predict how conclusions might differ and why.
Every empirical psychology study begins with an implicit or explicit claim: "here is something true about people." But the people actually studied are almost never "people in general" — they are a specific subset, a sample, drawn from a larger population the researcher cares about. Understanding sampling means understanding the chain between the people in your study and the people you want to say something about.
The ideal is a probability sample — one in which every member of the target population has a known, nonzero chance of being selected. A simple random sample (like drawing names from a hat) achieves this for a defined list. A stratified random sample divides the population into subgroups (strata) and samples randomly within each, ensuring minority groups are adequately represented. These methods are expensive and logistically difficult, which is why most psychology research instead uses convenience samples: whoever is available. In practice, this usually means introductory psychology students, who participate for course credit. Such samples are fast and cheap but structurally unrepresentative.
The WEIRD problem names the systematic bias that results. Henrich, Heine, and Norenzayan documented that the typical psychology participant is Western, Educated, Industrialized, Rich, and Democratic — characteristics that correlate with unusual patterns on perceptual tasks, fairness judgments, and social cognition relative to the rest of the world. Research built on WEIRD samples generates findings that may be specific to that cultural context but are framed as universal. Recognizing this is not a condemnation of all prior research — it is a call to replicate studies across cultures before assuming universality.
There is also a critical distinction between random sampling and random assignment. Random sampling concerns who is recruited into a study — it affects external validity and generalizability. Random assignment concerns who receives which experimental condition — it affects internal validity and causal inference. A study can have one without the other. A nationally representative survey has excellent random sampling but no random assignment; a tightly controlled lab experiment might have perfect random assignment but only a student convenience sample. Both matter, but they answer different questions.
Finally, sample size and sample quality are not the same thing. A large sample from a biased sampling frame is still a biased sample. Power analysis determines how large a sample you need to detect an effect of a given size — but it assumes you have a reasonably representative sample. Before worrying about sample size, you need to worry about whether the people in your study resemble the people you want to draw conclusions about.