Zurück

Types of Research: Purposes, Methods and Designs

Matthieu Saussaye

The main types of research, grouped by purpose, are exploratory, descriptive, correlational, explanatory and causal (experimental). Research is also classified by the data it uses (qualitative, quantitative or mixed), by where the data comes from (primary or secondary), by its aim (basic or applied) and by timing (cross-sectional or longitudinal). One study always has a place on each of these dimensions at once.

The short version:

  • Exploratory: what is going on? Used when the problem is not yet defined. Produces hypotheses.

  • Descriptive: how many, who, how often? Measures a population or behavior accurately.

  • Correlational: do two things move together? Measures association without changing anything.

  • Explanatory: why and how does a pattern happen? Ranks and unpacks explanations.

  • Causal or experimental: does changing X change Y? Tests a cause by changing it deliberately.

  • Qualitative vs quantitative: words and meanings versus numbers and measurement. Most real projects use both.

  • Primary vs secondary: data you collect yourself versus data someone else already collected.

The confusion online comes from mixing these dimensions in a single list, as if 'qualitative' and 'causal' were alternatives. They are not. A study can be qualitative and exploratory, or quantitative and causal. Pick the purpose first, then the methods that serve it.

Types of research by purpose, compared

Purpose is the dimension that matters most, because it decides what you will be able to conclude.

Type

Core question

Typical methods

Sample

What you can conclude

Business example

Exploratory

What is going on?

Interviews, open-ended surveys, desk research

Small, varied

These themes and hypotheses exist

Why are trial users not converting?

Descriptive

How many, who, how often?

Structured surveys, analytics, observation

Large, representative

X% of the population does Y

What share of customers use the mobile app weekly?

Correlational

Do two variables move together?

Survey or records analysis, correlation and regression

Large

X and Y are associated, with this strength

Do customers who use more features renew more?

Explanatory

Why and how does this pattern occur?

Driver analysis, mixed methods, follow-up interviews

Medium to large

This is the best-supported explanation

Why did satisfaction fall among enterprise accounts?

Causal / experimental

Does changing X change Y?

Randomized experiments, A/B tests, survey experiments

Randomized groups

Changing X changes Y by this amount

Does a guided setup raise satisfaction?

Each purpose has its own detailed guide: exploratory research, descriptive research, explanatory research and causal research.

The other ways research is classified

These dimensions sit alongside purpose. Every study has a value on each.

Qualitative, quantitative and mixed methods

Qualitative research collects words, images and observations to understand meanings, experiences and reasons. Interviews, focus groups and open-ended questions are typical. Quantitative research collects numbers to measure and compare: structured surveys, experiments, analytics. Mixed methods combine both, for example a survey followed by interviews that explain its results. For the detailed comparison, see qualitative vs quantitative research.

Primary and secondary research

Primary research is data you collect for your own question: your survey, your interviews, your experiment. Secondary research reuses data collected by someone else: government statistics, industry reports, published studies, your own CRM records collected for another purpose. Secondary research is cheaper and faster; primary research fits your question exactly. Most projects start with secondary research and move to primary research for what is missing.

Basic and applied research

Basic (fundamental) research aims to understand how something works, without an immediate use in mind. Most academic science starts here. Applied research aims to solve a specific practical problem, such as whether a new package design will sell better. Almost all market research is applied.

Cross-sectional and longitudinal research

Cross-sectional research measures at one point in time: a single survey wave. Longitudinal research measures repeatedly over time, either with new samples each time (trend studies, trackers) or with the same people (panel studies). Only longitudinal designs can show change.

Laboratory and field research

Laboratory research takes place in a controlled setting, which makes it easier to isolate causes. Field research takes place in the real world, such as a store, a website or a workplace, which makes results more realistic but harder to control. Online surveys and A/B tests are a kind of field research.

Inductive and deductive research

Inductive research builds a theory from observations, which is typical of exploratory and qualitative work. Deductive research starts from a theory or hypothesis and tests it with data, which is typical of causal and quantitative work.

Research types named after their method

Many lists also include types that are really methods. They are worth knowing, as long as you remember each one still serves a purpose from the table above.

  • Survey research: structured questions to a sample. Usually descriptive, but can be exploratory (open questions) or causal (randomized versions).

  • Case study research: an in-depth look at one organization, person or event. Usually exploratory or explanatory.

  • Ethnographic research: extended observation of people in their own environment, such as shadowing a team at work. Exploratory and qualitative.

  • Observational research: recording behavior without intervening, in person or through analytics. Descriptive or correlational.

  • Action research: a team studies its own practice while changing it, in repeated cycles of plan, act, observe and reflect. Applied by nature.

  • Historical research: analysis of past records and documents to understand how something developed. Usually secondary.

  • Meta-analysis: statistical combination of results from many earlier studies. Secondary and quantitative.

Which type of research do you need?

Start from the decision, then the question. The table maps common business questions to the type that answers them.

If your question is…

You need

Start with

We do not know why this is happening

Exploratory

10 to 20 interviews or an open-ended survey

How big is this market or segment?

Descriptive, often secondary first

Public statistics and industry reports, then a sizing survey

Who are our customers and what do they do?

Descriptive

A structured customer survey plus usage data

Is this metric linked to that one?

Correlational

Analysis of existing data or a survey with both measures

Why did this metric change?

Explanatory

Driver analysis with open follow-up questions

Will this change improve the result?

Causal

An A/B test or a split-sample survey experiment

Which concept, price or message works best?

Causal (survey experiment)

A monadic test, each respondent sees one version

How is this changing over time?

Descriptive, longitudinal

A tracker with identical questions each wave

How the types fit together in one project

Real projects move through several types in sequence. An illustrative product launch:

  1. Secondary, descriptive. Desk research on market size and competitors using public statistics and reports.

  2. Primary, exploratory, qualitative. Twenty interviews with target buyers to learn how they describe the problem and what they use today.

  3. Primary, descriptive, quantitative. A survey of 800 buyers to measure how common each problem is and how much they spend.

  4. Primary, causal, quantitative. A monadic concept and price test: each respondent sees one concept at one price.

  5. Primary, explanatory, mixed. Follow-up questions and driver analysis to understand why the winning concept won and what holds back the others.

  6. Longitudinal, descriptive. After launch, a quarterly tracker of awareness, trial and satisfaction.

The common thread: each step reduces uncertainty enough to justify the next, more expensive one.

Common mistakes when choosing a research type

  • Choosing the method before the purpose. 'Let's run a survey' before deciding whether you need to explore, describe or test leads to surveys with the wrong questions.

  • Treating exploratory findings as descriptive facts. Eight interviews can show that a problem exists, not how common it is.

  • Treating correlations as causes. Customers who use a feature may renew more because they were already your most engaged users.

  • Skipping secondary research. Paying to collect data that a statistics office or an industry body already publishes.

  • Running an experiment without a hypothesis. A test of 'something new' teaches less than a test of a specific prediction.

  • Changing questions in a tracker. It turns a longitudinal study into a series of unrelated snapshots.

Quick glossary of research types

Term

One-line definition

Exploratory research

Investigates an undefined problem to generate hypotheses

Descriptive research

Measures characteristics, frequencies and trends accurately

Correlational research

Measures whether and how strongly two variables are associated

Explanatory research

Explains why and how an observed pattern occurs

Causal / experimental research

Tests cause and effect by deliberately changing one variable

Qualitative research

Studies meanings and reasons through words and observation

Quantitative research

Measures and compares through numbers and statistics

Primary research

Data collected first-hand for the current question

Secondary research

Analysis of data previously collected by others

Longitudinal research

Repeated measurement over time

Applied research

Research aimed at solving a specific practical problem

Several research types, one platform

Most projects switch tools as they move from exploring to measuring to testing: an interview tool, a survey tool, a spreadsheet for coding open answers.

SmartInterview covers several of these stages in one place. It runs structured surveys and open-ended interviews by voice or text, with an AI interviewer that follows up on vague answers in the respondent's language. Respondents can be randomly assigned to versions for survey experiments, and open answers are coded automatically into themes, so exploratory depth and descriptive numbers come from the same fieldwork.

Start free and design your next study.

Frequently Asked Questions

What are the main types of research?

By purpose, the main types are exploratory, descriptive, correlational, explanatory and causal or experimental research. Research is also classified as qualitative or quantitative, primary or secondary, basic or applied, and cross-sectional or longitudinal. A single study has a place on each dimension.

What are the 3 basic types of research?

The three most often cited basic types by purpose are exploratory, descriptive and explanatory research. Exploratory research defines a problem, descriptive research measures it, and explanatory research explains why it happens. Many sources add causal or experimental research as a fourth.

What is the difference between qualitative and quantitative research?

Qualitative research collects words and observations to understand meanings and reasons, usually from a small number of people. Quantitative research collects numbers to measure and compare, usually from larger samples. Most practical projects combine both.

What is the difference between primary and secondary research?

Primary research collects new data for your own question, such as a survey or interviews. Secondary research analyzes data that already exists, such as government statistics or industry reports. Secondary research is faster and cheaper, primary research fits your question exactly.

Which type of research is best?

No type is best in general. The right type depends on your question: exploratory research when the problem is unclear, descriptive research to measure size and frequency, explanatory research to understand a pattern, and causal research to test whether an action will work.

Is a survey qualitative or quantitative research?

A survey can be either. Closed-ended questions such as ratings and multiple choice produce quantitative data, while open-ended questions produce qualitative data. Many surveys collect both, and coding open answers into themes makes the qualitative part countable.

Table of Contents

Verwandte Artikel

Kostenlos anmelden

Kostenlos anmelden

Kostenlos anmelden