History study
Primary and secondary sources: classify the evidence
Use the research question to distinguish a surviving record from a later interpretation, including scans, quotations and memories.
Primary and secondary sources are not two shelves labelled old and new. The distinction describes how a source relates to the subject you are investigating. A recent digital scan can preserve a much older primary record. A book printed long ago may interpret an even earlier event. Before choosing a label, identify the historical question and the exact material you are using.
This guide focuses on classification: deciding which kind of evidence an item supplies and explaining that decision. It is not a replacement for a full analysis of meaning, purpose or reliability. The classroom packets below are entirely fictional, with invented dates and situations rather than quotations from real archives. You can complete the exercises on paper. The photographs illustrate students studying; they do not show authentic historical evidence for the examples.
1. Primary and secondary sources answer different needs
In a history assignment, a primary source supplies direct evidence from people, activities or materials connected with the subject under study. A secondary source interprets, analyses or explains evidence about that subject. A participant’s letter and a historian’s later discussion can therefore do different jobs in the same investigation. You may need both rather than treating one category as the correct choice for every paragraph.
The OpenStax introduction to historical sources distinguishes original materials from later accounts and explains the contextual value of secondary research. Use that broad distinction, but do not turn examples into rigid format rules. Letters, images, objects and recordings can supply primary evidence. Books and websites can contain several kinds of material, including reproductions of earlier sources.
Start a classification note with two short statements: “I am investigating…” and “The item I am classifying is…” A question about the opening of a school garden is different from one about how that opening was remembered fifty years later. Likewise, an old newsletter inside a modern archive page is not identical to the page’s newly written description. Naming both the question and the item prevents several misleading shortcuts.
2. Sort an original classroom packet without guessing
Imagine a fictional school garden opened in 1974. Your question is: “How did pupils describe the opening at the time?” Packet A is a school newsletter produced in 1974 with a report by a pupil who attended. Packet B is a historian’s 2024 chapter that compares newsletters and school records. Packet C is a recorded 2018 interview with a former pupil remembering the opening. These are invented teaching materials, not real documents.
For this question, A is a primary account from the period, while B is a secondary interpretation. C is first-hand testimony recalled later: it can be treated as a primary oral-history source, but its recording date and retrospective nature must remain visible. It is not a statement recorded in 1974. The distinction between evidence from the time and later memory matters even when both fall within the broad primary category.
Packet D is a 2024 scan of A on an archive website. Classify the reproduced newsletter separately from the modern catalogue description. Packet E is an image forwarded without a creator, date or identifiable origin. Do not award E a primary label because the paper looks old. Record what is missing and say that its relationship to the question cannot yet be established. An honest uncertainty is more defensible than a completed but unsupported worksheet box.
3. Separate the event date from the copy date
Several dates may belong to one item. For the fictional newsletter, 1974 is the creation date of the original report, and 2024 is the date of digitisation or publication on a website. The screen on which you read it is newer still. Those later dates do not by themselves turn the reproduced report into a secondary account of the opening. They describe how the record reached you.
Verify that the object really is a reproduction of the identified item. Look for a catalogue reference, creator, original date, page information and an indication of missing or cropped material. A caption claiming “school newsletter, 1974” is useful information to check, not proof that every visible sentence belonged to the original. A translation or transcription may also introduce choices or errors without automatically changing the source’s basic role.
Write your label precisely: “A digital reproduction of a 1974 primary newsletter account, accessed through a 2024 archive page.” If you are using the page’s new explanation of the garden’s importance, say so separately. Keeping the layers apart helps a reader understand which words came from the earlier record and which came from a modern editor. Do not describe a scan as the original physical sheet you handled.
4. Classify the quoted passage and its container separately
Suppose the 2024 historian’s chapter quotes two sentences from the 1974 newsletter. The chapter remains a secondary interpretation of the opening. The quoted passage reproduces primary material within it. A book does not become entirely primary because one paragraph contains an old quotation. Equally, the original passage does not lose its historical relationship simply because a later author reproduced it.
Be clear about what you actually consulted. If you only read the quotation inside the chapter, do not claim to have inspected the complete newsletter. The selection may omit surrounding paragraphs, images or qualifications. Record the chapter’s author and page, the earlier source identification it supplies, and the limits of your access. Use your teacher’s required citation style for a source quoted within another source.
For a classification task, a useful answer is “The historian’s explanation is secondary; the embedded newsletter extract is primary material reproduced within it.” This is a distinction between layers, not an excuse to count one chapter as several independently consulted sources. To check a wider claim, locate the original record where possible or explain that your conclusion depends on a selected extract.
5. Keep classification separate from reliability
Primary does not mean true, complete or unbiased. Our fictional school newsletter might celebrate the garden because its writers wanted to promote the school. That makes it useful evidence of the account pupils were encouraged to publish. It does not establish what every pupil thought, how every part of the opening unfolded, or whether all reported details were accurate. The classification answers a different question from those judgments.
Secondary does not mean inferior. The later historian may compare several independent records, identify a mistaken date and explain an omission. A well-supported secondary account can therefore be more useful for checking a particular fact than one contemporary report. You still need to inspect the historian’s evidence and reasoning rather than accepting the account simply because it is scholarly or recent.
After labelling a source, ask what claim it can support and what you need to verify. Our guide to analysing a primary source addresses the next task: interpreting specific details, context and perspective. Keep that deeper analysis distinct from a worksheet asking only for type and a short justification. Neither a category label nor a long summary substitutes for evidence.
6. Change the research question and test the label
Now investigate a different question: “How did school history books in 2024 present the garden opening?” The same 2024 chapter becomes direct evidence of a 2024 representation. You can treat it as primary material for studying that representation, while it remains secondary for reconstructing the 1974 opening. There is no contradiction once you state the two different research questions.
This does not mean that every item can be called anything you like. The connection must be explained and supported. The 2024 chapter directly shows what its author published in 2024; it does not directly show every reader’s reaction. To investigate reception, you would need evidence about readers as well. A changed question alters the evidential role, not the document’s creator, date or contents.
Apply the same care to fictional writing. A novel published in 1990 and set in 1700 is not a contemporary record of 1700 just because a character writes diary entries. It could be primary evidence for a question about 1990 historical fiction. It is not automatically a sound secondary historical study of 1700 either. A task that offers only two boxes may need a qualification or a question to your teacher rather than a forced label.
7. Practise with new items and a stated question
Try this separate fictional investigation: “How did a town advertise its summer concert in 1986?” Item one is a surviving 1986 poster issued by the organisers. Item two is a 2025 research article comparing posters and council records. Item three is a scan of the poster in a 2025 online collection with a newly written caption. Item four is an unattributed picture that someone describes as an old concert poster.
For that question, the identified 1986 poster is primary material, and the 2025 analytical article is secondary. The scan reproduces primary material; the modern caption is a separate layer whose information and role need checking. The unattributed image cannot yet be classified confidently. State which missing details would help: original creator, date, collection reference or enough context to identify the event.
Change the question to “How did the online collection present local music in 2025?” The selection and caption on that collection page now provide primary evidence of its 2025 presentation. Explain the change in one sentence. Then compare your answer with a classmate’s: if your labels differ, first check whether you classified the same item for the same question. Many apparent disagreements come from silently choosing different layers or subjects.
8. Write a bounded answer and check any digital help
A concise classification sentence can identify the item, question, category and reason: “For studying how organisers advertised the 1986 concert, the identified poster is primary because it was produced as part of that advertising.” Add a separate sentence about uncertainty where needed. Do not turn the reason into a promise of accuracy, and do not invent provenance to make the answer sound complete.
If an AI labels every website secondary, test it with the scan example. Ask it to distinguish the reproduced record from the modern commentary and explain which research question it used. An optional explanation in Lirno Tutor may help with unfamiliar terms, but verify names, dates and attributions against the supplied material. AI can misread an image, invent a reference or apply a format shortcut confidently.
Keep private names and sensitive records out of uploads, use only material you are allowed to share, and follow school rules for assistance. Lirno does not guarantee correctness, grades, mastery or permission to use AI in assessed work. You do not need an app to complete this classification method: the essential step is to make the question, material and reason explicit.