Primary and Secondary Data Sources
One line each is all the supplied teaching source gives these two terms, and the two lines are not symmetrical. This page works out what that asymmetry commits you to once you have to decide where your evidence will actually come from.
Two definitions, one line each
Both terms appear in the terminology table of the supplied teaching source, alongside terms such as descriptive, explanatory, inductive and deductive — the whole table is reproduced in the research terminology glossary. They are set out there as vocabulary, not as a decision framework, and the entirety of the source's treatment of these two terms is reproduced below.
THE SOURCE DEFINITIONS IN FULL
| Term | Definition as printed | Associated terms given |
|---|---|---|
| Primary data | Data directly observed or gathered by the researcher engaged in the research project. These sources are relevant, timely and reliable. | None listed |
| Secondary data | Use of existing data gathered by other people or organisations for other purposes. These sources are interpreted by others. | None listed |
That is the complete treatment of both terms in the supplied material. Every other statement on this page is either drawn from another section of the source, or flagged as practice.
Read the two rows again as a pair. Primary data is credited with three properties: relevance, timeliness and reliability. Secondary data is credited with none, and is given a caveat instead. Nothing elsewhere in the supplied material qualifies either statement.
The interpretation layer you inherit
The one substantive warning the source gives about secondary sources is that they are interpreted by others. That single clause carries more than it appears to, because interpretation happens long before anyone writes an analysis.
Someone decided what was worth recording. Someone chose the categories, drew the boundaries, and decided what fell outside them. Someone decided when to stop collecting. All of those decisions arrive attached to the dataset, and none of them were made with your research question in mind.
Three things that travel with second-hand evidence
It was collected to answer something else
The source's own definition is explicit that secondary data was gathered for other purposes. The fields that matter most to you are often the ones the original collector had no reason to capture precisely.
Someone else drew the boundaries
Category schemes encode judgments — which contract types are distinct, where a project ends and operations begin, what counts as a variation. You inherit those judgments whole and usually cannot recover the underlying detail.
It stopped when they stopped
Existing data has a collection window you did not choose. The source's own question set asks how current the information needs to be; a secondary source answers that question for you, and the answer may not be yes.
Where the choice sits in the research process
Sourcing is not a step of its own in the supplied material. The seven-stage account of the research process moves from research question to data collection directly, and the question of provenance is folded into the research question step, where you are told to ask yourself whether you will use primary and/or secondary sources of data.
- Research question fixed
- Decide primary and/or secondary
- Source and collect
- Analyse and interpret
- Theorise against the literature
The and/or in the source's own phrasing is the useful part. It is not framed as a binary. The realistic reading is that most designs use both, with secondary material establishing the context and primary collection answering the specific question — which is exactly the pattern the exemplar designs show.
Two of those questions are easy to skip and expensive to skip. Authority to collect points directly at Step 6: seeking ethics clearance and at whatever internal approval your organisation requires before you may ask its people anything. Currency determines whether an existing dataset is usable at all. Both are best answered while the question is still being written, in Step 4.
What the source leaves out
That gap matters more here than almost anywhere else in the syllabus, because sourcing decisions are the ones that cannot be reversed late. An unfit analysis can be redone; data you were never permitted to collect cannot be recovered in the final month.
A worked example: secondary establishes, primary answers
One chapter of the exemplar volume of master's project management research studies the effective management of infrastructure projects in a remote region. Its design is a clean demonstration of the two source types doing different jobs.
How the two source types divide the work in that design
Existing material sets the frame
A literature review is used to determine the challenges facing such projects and the strategies available. The output is a defined list of candidate challenges — thirteen of them — which the researcher did not have to generate from scratch.
Primary collection tests the frame against this context
Data is then collected from core project team members, by online survey and by interview. The people involved in the actual delivery are asked about the challenges the literature identified.
The finding is the comparison
Of the thirteen challenges identified from the literature, nine scored high or extreme. The contribution is not the list and not the scores separately; it is what happens when the second-hand frame meets first-hand evidence.
How the exemplar designs combine the two
Reading across the seventeen research designs in the exemplar volume, an observable pattern emerges: almost every one pairs a review of existing published material with one or two primary collection methods, and survey and interview dominate the primary side. That is an observation about that volume, not a published statistic.
THREE EXEMPLAR DESIGNS AND HOW THEY SPLIT THEIR SOURCING
| Subject of the study | Existing-material component | Primary component |
|---|---|---|
| Project risk management practices in defence small and medium enterprises | A parallel review of major defence project performance data, used for context | A survey, with findings mapped to a risk management maturity model |
| Impact of modular construction on liquefied natural gas megaprojects | Examination of published works to identify differences in execution strategies and known causes of overruns | Surveying and interviewing experienced industry participants |
| Process improvement on a remote construction project | Review of existing literature on process improvement | A qualitative, opinion-based survey of professionals working in the sector |
Designs paraphrased from published abstracts in the exemplar volume. The full set is discussed in the worked exemplars page.
The lesson for a practising project manager is a scoping one. If your organisation already holds performance data, cost records, register extracts or closed-out lessons, that material can do the framing work at almost no cost — leaving your limited primary collection budget for the specific thing nobody else has asked. The seventeen designs are examined in detail in seventeen worked research designs in project management.
Testing a secondary source before you rely on it
The supplied source supplies no test. The checklist below builds one from its own sourcing questions, plus the interpretation warning attached to the secondary data definition. Treat the first four items as source-anchored and the rest as practice.
Before an existing dataset becomes your evidence
- You can state what the data was originally collected for, and it is not what you are using it for.
- You know how current it is, and you have decided how current your question requires it to be.
- You know how much of it there is, and whether that is enough for the claim you intend to make.
- You have the authority to use it, and that authority is recorded rather than assumed.
- You can see the category definitions the original collector used, not just the category names.
- You know where collection stopped and why, and whether anything systematic sits outside the boundary.
- You can describe the dataset's provenance in your methodology chapter in three sentences without hedging.
Common mistakes
What to carry forward
- The distinction the source draws is provenance and purpose: gathered by you for this project, or gathered by others for something else.
- Primary data is credited with relevance, timeliness and reliability. Nothing qualifies that claim, so treat those three words as design objectives.
- Secondary data arrives already interpreted — in its purpose, its categories and its collection window.
- The source's own sourcing questions cover sufficiency, currency and authority to collect. Answer all three while the question is still being written.
- The supplied material has almost nothing on data collection generally and no test for evaluating an existing dataset. You will need to source that elsewhere and cite it.
- The dominant pattern in the exemplar volume is existing material for the frame, primary collection for the answer.
Frequently asked questions
Is a literature review secondary data?
The supplied source does not say. It defines secondary data as existing data gathered by others for other purposes, and separately describes the literature review as the theoretical framework of the project. Many methodology chapters do treat published studies as secondary sources, but you cannot rest that claim on this teaching material — cite a research methods text if you want to make it.
Why does the source call primary data reliable without qualification?
It states relevance, timeliness and reliability as properties of the category, and offers no conditions. The most defensible reading is that it describes what primary collection can deliver when the instrument is well designed and the collection window is short. Nothing in the material licenses the assumption that any primary dataset is reliable simply because you gathered it.
Can a single study use both types?
Yes, and the source's own phrasing asks whether you will use primary and/or secondary sources of data. Across the exemplar volume of master's research, nearly every design pairs a review of existing published material with one or two primary collection methods. That combination is the norm rather than the exception.
What does the source say about evaluating the quality of an existing dataset?
Nothing. There is no quality test, no sampling guidance and no instrument design guidance anywhere in the supplied material — the entire prose on data collection runs to three sentences. Any evaluation framework you apply is yours to source and to justify.
Do I need approval to use my own organisation's project data?
The source raises the question rather than answering it, asking what authority you need to collect the data or information. It also defers substantive ethics treatment to a later topic that is not part of the supplied material. Settle internal permission in writing before you build a design that depends on the dataset.
References and source attribution
- Bryman, A. 2016, Social Research Methods, 5th ed., Oxford University Press, Oxford.
- Naoum, S. G. 2013, Dissertation Research & Writing for Construction Students, 3rd ed., Routledge. (Week 2 reading, pp. 1-6, in the supplied teaching source.)
- O'Leary, Z. 2017, The Essential Guide to Doing Your Research Project, 3rd ed., Sage Publications, London.
- An edited volume of master's degree research projects in project management, 2018, academic publisher, 378 pp. (design descriptions paraphrased from published abstracts).
- The supplied teaching source: teaching notes and slide decks on introduction to research methods, terminology table; and on developing a research topic, sections 4 and 5. Institution, unit code and authors withheld.
Suggested questions for Ask KEVOS
- Which parts of my research question need primary collection and which can be answered from existing records?
- Show me the exact definitions of primary and secondary data in the supplied source.
- Draft a provenance paragraph for a methodology chapter describing an internal project dataset.
- What questions does the supplied source tell me to ask about sourcing before I collect anything?
- How do the exemplar research designs split their work between existing material and primary collection?
- What approvals should I settle before designing a study around my organisation's own delivery data?