Mathwords logoMathwords

Data Collection Method — Definition & Example

A data collection method is the plan you use to get your data, such as asking people questions, watching what they do, or assigning them to try something. The method you choose decides what conclusions your data can support. In research generally, methods are also grouped by where the data come from: primary data you collect yourself (surveys and questionnaires, interviews, observation, measurements) and secondary data someone else already collected (existing records, government databases, published studies).

A data collection method is the design that determines how values of variables are obtained from individuals. The main designs are a census, which measures every individual in the population; observational studies, which measure variables without influencing the responses (a sample survey is the most common kind: it measures a subset chosen to represent the population); and experiments, which deliberately impose treatments on individuals and measure their responses.

How It Works

Ask two questions about any study. First, were individuals chosen at random from a larger population? If so, you can generalize results to that population. Second, did the researcher assign treatments at random? If so, you can draw cause-and-effect conclusions. If no treatment was imposed, the study is observational, and confounding variables may explain any association you see.

Example

Problem: A principal at a 1,200-student high school wants to study sleep and test scores. Classify each plan. (A) Randomly select 120 students and ask how many hours they sleep. (B) Compare test scores of students who already have a free first period with those who do not. (C) Randomly assign 60 of 120 volunteers to start school 1 hour later and the other 60 to keep the usual start, then compare scores.
Step 1: Plan A selects part of the population at random and asks questions, so it is a sample survey. It measures only sleep, not test scores, so on its own it cannot answer the principal's question.
Step 2: Plan B records groups that students formed themselves. No treatment is imposed, so it is an observational study. Students with a free first period may differ in other ways, such as workload, which could confound the comparison.
Step 3: Plan C assigns the treatment (later start) at random, so it is an experiment. Random assignment tends to balance other variables between groups, which supports a cause-and-effect conclusion for these volunteers.
Answer: A is a sample survey (of sleep only, so it cannot answer the question by itself), B is an observational study, and C is a randomized experiment. Only C can show that a later start causes a change in scores.

Why It Matters

AP Statistics questions, including free-response questions, often ask you to identify how data were collected and what conclusions that method supports (the scope of inference). Random sampling allows generalization to a population, while random assignment allows causal claims. Medical trials, election polls, and product testing all depend on getting this distinction right.

Common Mistakes

Mistake: Concluding that one variable causes another because an observational study found a strong association.
Correction: Without random assignment of treatments, confounding variables may explain the association. Only a well-designed experiment supports a cause-and-effect conclusion.

Related Terms

  • Population — The full group a census measures or a sample represents
  • Sample (Statistics) — The subset measured in a sample survey
  • Random Sample — Selection rule that lets survey results generalize to populations
  • Stratified Sampling — A random sampling design used within sample surveys
  • Convenience Sampling — A biased selection approach that undermines survey conclusions
  • Sample Size — How many individuals a chosen method actually measures
  • Observational Study — Design that records data without imposing treatments
  • Experimental Design — How an experiment assigns treatments and measures responses
  • Census — Design that measures every member of the population
  • Confounding Variable — Linked to both the explanatory and response variables; can fake cause and effect
  • Sample Survey — A survey of a sample, the most common observational design