Primary and secondary research
What separates them, when each is the right choice, and why the humanities use the same words to mean something different.
The distinction sounds obvious and produces more confusion than almost any other pair of terms in research, because two disciplines use the words to mean different things.
The social science definition
| Primary research | Secondary research | |
|---|---|---|
| What it is | Data you collect yourself | Data or sources that already exist |
| Examples | Interviews, surveys, experiments, observation | Published studies, official statistics, archives, existing datasets |
| Control | You decide what to measure | Limited to what others measured |
| Time | Months, including recruitment | Weeks, once you have access |
| Ethics | Full approval, often lengthy | Lighter process, sometimes none |
| Scale | Whatever you can manage alone | Potentially national datasets |
The humanities definition, which is different
In history, English and cultural studies, primary means the object of study itself and secondary means scholarly writing about it.
Studying Woolf's Mrs Dalloway
Primary: the novel, Woolf's diaries, contemporary reviews from 1925
Secondary: a 2019 journal article analysing the novel
Studying the 1926 General Strike
Primary: Cabinet papers, union records, newspapers from 1926
Secondary: a modern historian's account of the strike
A joint-honours student taking history and psychology modules will meet both definitions in the same term. Neither is wrong; check which your marker means.
When primary research is the right choice
- Your question concerns a population nobody has studied
- Existing data does not measure what you need
- You want to understand how people make sense of something, which requires asking them
- The context has changed enough that older data no longer applies
- Your programme requires it
Be realistic about time. Ethics approval commonly takes four to eight weeks, and recruitment routinely takes three times the estimate. A dissertation depending on interviewing NHS staff can lose a term to approvals before a word is written. Our guide to data collection methods covers piloting and the ethics timeline in more detail.
When secondary research is the right choice
- Good data already exists at a scale you could never collect
- Your question is about patterns across studies rather than a new population
- You have three months, not twelve
- Access to participants is genuinely unlikely
- The question is historical or documentary by nature
Sources worth knowing
| Source | Holds |
|---|---|
| UK Data Service | Major UK social surveys, free to registered students |
| Office for National Statistics | Census, labour market, health and crime statistics |
| NHS Digital | Health service activity and outcomes data |
| Understanding Society | Longitudinal household panel, tracking the same people over years |
| The National Archives | UK government records |
| Gov.uk statistics | Departmental releases across every policy area |
The honest trade-off
Secondary research is faster and lower-risk. It is not easier to do well, and it has a specific weakness worth naming in your limitations.
You are constrained by what someone else decided to measure. If the Labour Force Survey records employment status but not job satisfaction, no amount of analysis will produce a finding about satisfaction. Primary research lets you ask exactly what you want and costs you months.
Combining the two
Many strong dissertations do both, with secondary work establishing the pattern and primary work explaining it.
Secondary: national data shows attendance is lowest in the most deprived quintile
Primary: interviews with fifteen parents in one such area explore why
That combination is efficient. The scale comes from data you did not have to collect, and your primary work answers the question the numbers raise but cannot address.
How to decide
- Write the question first. Never choose the approach before the question.
- Search for existing data. Spend a full day on this. Students routinely collect data that already existed.
- Check your timeline honestly. Count backwards from submission through analysis, collection, recruitment and ethics. If it does not fit, secondary is not a compromise, it is the correct decision.
- Check the handbook. A minority of programmes require primary collection.
- Ask your supervisor. They know which datasets students in your department have used successfully.
There is no hierarchy here. A well-designed secondary analysis beats a rushed set of interviews with four participants recruited in the final fortnight, and markers know it.
Last updated: 30 July 2026 · Written by the AssignWise editorial team ·Spotted something wrong?
Frequently asked questions
Primary research collects new data yourself: interviews, surveys, experiments, observation. Secondary research analyses data or sources that already exist: published studies, official statistics, archives, existing datasets. Both are legitimate research.
It is faster and usually avoids ethics approval, but it is not easier to do well. You are limited to what someone else chose to measure, and the analytical work of building an argument from existing evidence is substantial.
Yes, and many strong dissertations do. Systematic reviews, policy analysis, archival work and secondary dataset analysis are all standard. Check your handbook, because a minority of programmes require primary data collection.
The object of study itself: a novel, a painting, a manuscript, a government record, a newspaper from the period. Scholarly writing about that object is secondary, which reverses the social science usage and confuses joint-honours students constantly.
Usually a lighter process, and sometimes none, if you use fully anonymised published data. You will still normally complete an ethics form. Restricted datasets have their own access conditions on top.
Want a second pair of eyes on it?
Send us your document for a fixed, no-obligation quote. We reply within one working hour.
Get Your Free Quote