← Back to Sociology

Doing research: the scientific method

Sociologists test ideas about society with evidence. Following the scientific method helps them move from hunches to reliable knowledge.

Sociology as a science

Sociologists use systematic methods to study social life, aiming for findings that others can check. Positivist approaches, following Comte and Durkheim, model sociology on natural science, seeking general patterns and causes. Interpretive approaches, following Weber, emphasise Verstehen, understanding the meanings people give to their actions. Most research combines both.

The research process

A typical research process involves asking a question, reviewing existing research, forming a hypothesis, choosing a design, collecting data, analysing results, and reporting conclusions, which others can then check and build upon. Research is cyclical: findings raise new questions.

Variables and hypotheses

A variable is a characteristic that can change, such as income or education. A hypothesis predicts a relationship between variables: the independent variable (the presumed cause) and the dependent variable (the presumed effect). 'More years of education lead to higher income' makes education independent and income dependent.

Operational definitions

Abstract concepts must be made measurable through operational definitions. 'Social class' might be measured by income, occupation, or education; 'religiosity' by frequency of worship or strength of belief. Different operational definitions can produce different results, so researchers must state them clearly.

Reliability and validity

Reliability means a measure gives consistent results, for example a survey question answered the same way on repeated occasions. Validity means it measures what it claims to measure: counting church attendance may not validly measure personal faith. A measure can be reliable without being valid, like a scale that consistently reads two kilograms too heavy.

Correlation and causation

A correlation is a relationship between variables, such as ice-cream sales and drowning rates both rising in summer. Correlation does not prove causation: a third variable, here hot weather, may explain both, producing a spurious relationship. To infer causation, researchers need correlation, correct time order, and the elimination of alternative explanations.

Samples and generalisation

Researchers usually study a sample rather than an entire population. A random sample, where every member has an equal chance of selection, allows findings to be generalised to the population within a margin of error. Biased samples, such as volunteers or online polls, can mislead. The 1936 Literary Digest poll wrongly predicted that Alf Landon would beat Franklin Roosevelt because its sample over-represented wealthier people.

Worked example

A researcher hypothesises that teenagers who eat family dinners regularly have better school grades. The independent variable is family-dinner frequency; the dependent variable is grade average. A survey finds a correlation, but family income could be a third variable affecting both dinners and grades. Controlling for income statistically, or using longitudinal data showing changes over time, helps test whether the relationship is causal.

Key takeaways

Sociology is systematic: positivist approaches seek patterns and causes, interpretive approaches seek meaning (Weber's Verstehen). Research moves from question to hypothesis, data, analysis, and reporting. Hypotheses link independent and dependent variables. Operational definitions make concepts measurable. Reliable measures are consistent; valid measures capture the intended concept. Correlation is not causation; spurious relationships arise from third variables, and causation needs correlation, time order, and ruling out alternatives. Random samples allow generalisation; biased samples, like the 1936 Literary Digest poll, mislead.

Flashcard

Positivism

Modelling sociology on natural science

Comte, Durkheim.

Check your understanding

A survey question gives the same answers every time but actually measures something different from what researchers intend. The measure is:

Take the unit quiz

← Previous lesson Next lesson →