Research Methods: A Plain-Language Guide

Research methods are the procedures and techniques used to collect and analyze data. The choice of method depends on the research question: what you want to know determines how you should look for it.

This guide covers the major categories of research methods used in social science, education, psychology, business, and related fields, with enough detail to understand the differences and choose appropriately.

Quantitative vs Qualitative Research Methods

The most fundamental distinction in research methods is between quantitative and qualitative approaches:

Most research questions can be addressed by either approach, but they produce different types of knowledge. Quantitative research tells you how much, how often, and whether differences are real. Qualitative research tells you why and what something means to the people involved.

Mixed methods research combines both approaches in the same study, typically using qualitative methods to explain or contextualize quantitative findings.

Experimental vs Observational Methods

Within quantitative research, the most important distinction is between experimental and observational designs:

The key principle: correlation does not imply causation. Only randomized experiments with adequate controls can establish that one variable causes another. Observational studies can identify associations, suggest mechanisms, and generate hypotheses, but alternative explanations remain possible.

Major Research Methods

Experiments

Randomized controlled trial (RCT). Participants are randomly assigned to conditions (treatment vs. control). Random assignment ensures that the groups are equivalent before the intervention, making it possible to attribute any difference in outcomes to the treatment. RCTs are considered the gold standard for causal inference in medicine, psychology, and policy evaluation.

Laboratory experiment. Conditions are controlled in a laboratory setting to isolate the variable of interest. High internal validity (the study measures what it claims to measure) but potentially low external validity (findings may not generalize to real-world settings).

Field experiment. An experiment conducted in a natural setting. Real-world context increases external validity but introduces more sources of variation than a laboratory.

Surveys and Questionnaires

A survey collects self-reported data from participants using structured questions. Surveys are efficient for collecting data from large samples and can measure attitudes, behaviors, and experiences. Key considerations:

• Sampling: A representative sample is necessary for results to generalize to a population. Convenience samples (e.g., college students, Amazon Mechanical Turk) limit generalizability.
• Response bias: Participants may underreport stigmatized behaviors or overreport desirable ones.
• Cross-sectional vs. longitudinal: A one-time survey measures a snapshot; repeated surveys track change over time.

Interviews

Qualitative interviews involve an in-depth conversation between researcher and participant. They produce rich, detailed data about experiences, reasoning, and meaning.

• Structured: Every participant is asked the same questions in the same order. Allows comparison across participants.
• Semi-structured: A guide with core questions, but the interviewer can probe and explore responses. Most common in qualitative research.
• Unstructured: Conversational; follows the participant's lead. Used in exploratory or ethnographic work.

Focus Groups

A moderated discussion with 6 to 10 participants. Useful for understanding group dynamics, shared norms, and how people discuss topics socially. Not appropriate when individual views are the primary interest, since social pressure can suppress minority opinions.

Observation

The researcher observes behavior in a natural or structured setting.

• Participant observation: The researcher is embedded in the setting being studied. Common in ethnography and sociology.
• Non-participant observation: The researcher observes without participating. Can be structured (coding specific behaviors) or unstructured.
• Systematic behavioral observation: Observers code predefined behaviors according to a structured protocol. Used in developmental psychology and classroom research.

Case Studies

In-depth examination of a single case (an individual, organization, event, or policy). Case studies are useful when the boundary between the case and its context is unclear and when you need rich, contextual understanding. Not appropriate for generalizing to a population.

Secondary Data Analysis

Analysis of data collected by others for a different purpose (e.g., government databases, administrative records, previously collected survey data). Efficient and allows access to large, high-quality datasets but limits the researcher to variables that were already collected.

Common secondary data sources: U.S. Census, National Health Interview Survey, General Social Survey, Add Health, PISA, administrative school records.

Systematic Review and Meta-Analysis

A systematic review identifies, evaluates, and synthesizes all existing research on a question using a pre-specified protocol. A meta-analysis quantitatively combines effect sizes from multiple studies to produce a pooled estimate.

Systematic reviews and meta-analyses sit at the top of the evidence hierarchy for many fields because they synthesize the entire body of evidence rather than relying on any single study.

How to Choose a Research Method

The choice of method flows from the research question:

Practical constraints also matter: available time, funding, access to participants, and institutional requirements all shape the choice of method. A question that ideally calls for an RCT may need to be addressed with an observational study if random assignment is not feasible.

Validity and Reliability

Two concepts apply across all research methods:

Validity: Does the study measure what it claims to measure?

• Internal validity: Are the causal claims justified by the design?
• External validity: Do the findings generalize beyond the study sample or setting?
• Construct validity: Does the measure accurately capture the theoretical construct?

Reliability: Would the study produce the same results if repeated?

• Test-retest reliability: Consistency over time
• Inter-rater reliability: Consistency between different raters or coders
• Internal consistency: Items in a scale measuring the same construct

A measurement can be reliable without being valid (consistently measuring the wrong thing). A measurement cannot be valid if it is not reliable.

Frequently Asked Questions

What are the main types of research methods? The main categories are quantitative (collecting numerical data to measure relationships and test hypotheses) and qualitative (collecting textual or observational data to understand experiences and meaning). Within quantitative research, methods include experiments, surveys, and secondary data analysis. Within qualitative research, methods include interviews, focus groups, observation, and case studies. Mixed methods research combines both approaches.

What is the difference between research methods and research methodology? Research methods are the specific techniques used to collect and analyze data (e.g., survey, experiment, interview). Research methodology is the broader philosophical framework that guides the choice of methods, including assumptions about knowledge (epistemology) and what counts as valid evidence. In a thesis or dissertation, the methodology chapter explains both why you chose a particular method and what assumptions underlie that choice.

Which research method is best? There is no single best research method. The best method is the one that best answers the research question given practical constraints. If you want to establish causation, a randomized experiment is best. If you want to understand lived experience, in-depth interviews are better. Most research questions can be addressed in multiple ways, and the choice involves trade-offs between internal validity, external validity, feasibility, and depth.

What is quantitative vs qualitative research? Quantitative research collects numerical data and uses statistical analysis to measure relationships, compare groups, and test hypotheses. It typically uses large samples and aims to generalize findings to a population. Qualitative research collects non-numerical data (text, images, observations) and uses interpretive analysis to understand experiences, meanings, and processes. It typically uses small samples studied in depth. The two approaches answer different types of questions and are often used together.

What is a hypothesis in research methods? A hypothesis is a testable prediction about the relationship between variables. In quantitative research, hypotheses are stated before data collection and specify the expected direction of the relationship (e.g., "students who receive tutoring will have higher grades than students who do not"). A null hypothesis states that there is no relationship. The researcher then collects data and uses statistical tests to determine whether the evidence is consistent with or against the hypothesis. Qualitative research typically does not state formal hypotheses, instead developing propositions inductively from the data.

For related research guidance, see primary vs secondary sources, how to write a literature review, and how to write a research paper introduction.