Special cases that confuse students
Journal articles
This is the most common point of confusion, because journal articles can be either primary or secondary.
An article that reports original research is considered a primary source that includes empirical studies, experiments, clinical trials, or any type of research where the authors collected and reported their own data. However, the literature review that is published in the same article is the secondary source, as it summarizes and provides an analysis and understanding of others' work. That’s why understanding of primary vs. secondary data sources is essential while citing any source.
Moreover, if the article is peer-reviewed, it is not necessarily a primary source, and it can be either. To be clear, peer review gives the quality check and not the category it belongs to. So you have to look for it yourself to determine whether the data is generated by the authors directly or not.
Autobiographies and memoirs
These are primary sources because they are based on the writer’s own experience. But they are written long after the events they describe, so they are shaped by memory, hindsight, and the writer’s wish to present themselves in a particular way. Treat them as firsthand evidence that needs careful handling.
Interviews and oral histories
A transcript or recording of an interview is a primary source, whether you conducted it yourself or found it in an archive. If a researcher quotes selected parts of that interview inside an argument of their own, that article is a secondary source.
Translations and edited collections
A translated novel or a published collection of a writer’s letters is still treated as a primary source, but it has passed through an editor’s or translator’s hands. Note which edition you used, because the choices made in translation and selection can affect your analysis.
There is no single answer. An organization’s own posts, statements, and pages are primary sources for research about that organization. A blog post summarizing a news story is a secondary source. Apply the three questions above rather than judging by the medium.
How to choose between primary and secondary sources
Knowing the difference is only half the job. The more practical question is which type to reach for at a given stage of a project. Five factors usually decide it.
What your research question needs. If you are making a claim about what happened, what was said, or what the data shows, you need primary sources, because that is the only kind of evidence that can settle the point. If you are positioning your work within a field, explaining why a debate exists, or establishing what is already known, secondary sources do that job better.
How well you know the topic. Starting with primary sources in an unfamiliar field is slow and often misleading, because you have no way of judging whether the document in front of you is typical or unusual. Read two or three good secondary sources first to learn the shape of the debate and the names that keep recurring, then go to the primary material with an idea of what you are looking for.
How much detail you need. A close reading, a discourse analysis, or an original data argument requires primary material. A background section, an introduction, or a short orientation piece rarely does.
How much time you have. Working through primary sources takes longer than it looks. Archival material may be uncatalogued, handwriting may be difficult, datasets may need cleaning, and none of it arrives pre-summarized. Secondary sources are written to be read quickly. Plan your schedule around that difference rather than discovering it a week before the deadline.
What you can actually get hold of. Check early whether the primary sources your question depends on are reachable, whether that means a digitized archive, a data repository, an interlibrary loan, or a visit. A question that requires material you cannot access is a question that needs reframing, and it is far cheaper to learn that in week one than in week six.