Research methods · 15 pages
Research methods & design
Clear, standards-grounded explainers for the core building blocks of empirical research — variables, hypotheses, study designs, sampling, and data-collection methods. Each page leads with a concise definition and links across to the wider CASRAI standards and dictionary.
Browse the topic
All 15 research methods & design pages
Independent vs dependent variable
The difference is that the independent variable (IV) is the factor a researcher deliberately changes or varies — the presumed cause — while the dependent variable (DV) is the outcome that is measured to see whether it responds. The DV is called "dependent" because its value is expected to depend on the IV. In a dose–response study, the drug dose is the IV and the measured symptom level is the DV.
ComparisonInductive vs deductive reasoning
The difference is direction and certainty. Inductive reasoning works bottom-up, moving from specific observations to a broader, probable generalisation or theory. Deductive reasoning works top-down, moving from a general theory or premise to a specific prediction that is logically certain if the premises are true. Induction is associated with theory-building and qualitative work; deduction with theory-testing and quantitative work. Many studies combine the two in an abductive cycle.
DefinitionIndependent variable
An independent variable is the factor a researcher manipulates, varies or selects in order to test its effect on something else. It is the presumed cause in a cause-and-effect relationship, standing independently of the outcome being measured. In an experiment on study time and test scores, study time is the independent variable. It is also called the predictor, explanatory or treatment variable, and is conventionally plotted on the x-axis.
DefinitionDependent variable
A dependent variable is the outcome that a researcher measures to see whether it changes in response to the independent variable. It is called "dependent" because its value is expected to depend on the factor being manipulated — it is the presumed effect in a cause-and-effect relationship. In a study of study time and test scores, the test score is the dependent variable. It is also called the outcome, response or criterion variable.
DefinitionConfounding variable
A confounding variable is an extraneous variable that correlates with both the independent and the dependent variable, creating a false or distorted impression of the relationship between them. Because it offers an alternative explanation for the observed effect, it threatens internal validity and can make a non-causal link look causal. This is the third-variable problem. Researchers control confounders through randomisation, restriction, matching and statistical adjustment.
DefinitionControl group
A control group is the group in an experiment that does not receive the treatment or intervention being tested. It provides a baseline, so researchers can compare it with the experimental (treatment) group and judge whether the treatment actually caused any difference in the outcome. Without a control group, you cannot rule out that the change would have happened anyway. Control groups are central to randomised controlled trials, often using a placebo.
GuideData collection methods
Data collection methods are the systematic techniques researchers use to gather and measure information to answer a research question. They divide first into primary methods, which gather new data directly (surveys, experiments, interviews, observation), and secondary methods, which reuse existing data. They divide again by data type: quantitative methods produce numbers, while qualitative methods produce text or imagery. The right choice depends on the question, and every method must guard validity and reliability.
DefinitionNull hypothesis
A null hypothesis (H0) is a statement that there is no effect, no difference, or no relationship between variables — the default assumption a statistical test sets out to challenge. Researchers gather evidence to decide whether to reject H0 in favour of an alternative hypothesis. You never prove the null true; you only fail to reject it when the evidence is weak.
DefinitionAlternative hypothesis
An alternative hypothesis (H1 or Ha) states that there is an effect, difference, or relationship between the variables under study — the claim a researcher expects to find. It is tested against the null hypothesis of no effect: if the evidence is strong enough to reject the null, support shifts to the alternative. It can be directional (predicting which way) or non-directional.
DefinitionResearch hypothesis
A research hypothesis is a clear, specific, testable and falsifiable prediction about the relationship between variables, usually derived from theory or earlier findings. A strong one names the variables, states the expected relationship and can be supported or refuted by evidence. It differs from a research question, which asks rather than predicts, and from a statistical hypothesis (H0 or H1).
DefinitionQuestionnaire
A questionnaire is a research instrument consisting of a standardised set of questions used to gather data from respondents. It can mix open-ended questions, which invite free-text answers, and closed-ended questions, which offer fixed options such as Likert scales. Standardising the wording lets researchers compare responses across people. A questionnaire is the tool; a survey is the wider study that uses it.
DefinitionCross-sectional study
A cross-sectional study is an observational design that measures variables at a single point in time across a population or sample, producing a snapshot. It is well suited to estimating prevalence and identifying associations between variables, and is usually quick and inexpensive. Because it captures only one moment, it cannot establish time order and so cannot reliably show causation. Longitudinal studies follow subjects over time instead.
DefinitionStratified sampling
Stratified sampling is a probability method that splits a population into mutually exclusive subgroups, called strata, that share a relevant characteristic, then randomly samples within each stratum — often in proportion to its size. By guaranteeing every subgroup is represented, it improves representativeness and precision compared with simple random sampling when the strata differ on the variable of interest. It differs from cluster sampling.
GuideSampling methods
Sampling methods are the strategies researchers use to choose which members of a population to include in a study. They fall into two families: probability sampling, where every unit has a known, non-zero chance of selection (simple random, systematic, stratified, cluster), and non-probability sampling, which relies on judgement or convenience (convenience, purposive, quota, snowball). Probability methods support generalisation; non-probability methods are quicker but more prone to bias.
DefinitionControl variable
A control variable is a variable that is held constant (controlled) throughout an experiment so it does not affect the outcome. Controlling such variables isolates the relationship between the independent and dependent variables, strengthening internal validity. A control variable is not the same as the control group, and an uncontrolled one can become a confounding variable.
