Did you know that two versions of the same interview can look equally official while carrying very different evidence? A history project might have an audio file, notes made during the conversation and a tidy paragraph written three days later. Every version may be useful. The mistake begins when a student treats a polished sentence remembered afterwards as the speaker’s exact words.
An interview recording preserves an audio or video trace of what was captured at that time. Contemporaneous notes are observations written during or immediately alongside the event; they may be incomplete or paraphrased. Reconstructed notes are an account written later from memory, possibly assisted by other material. The immediate rule is simple: identify the kind of record, keep the original separate, verify direct quotations against an available recording, and label uncertain memories as uncertain. Do not invent exact words, dates or numbers to make the project sound authoritative.
Why this is a source-quality question rather than a handwriting question
Beautiful handwriting does not make a reconstructed account more direct evidence than a messy recording. Nor is an audio file automatically a perfect representation of the whole event. A microphone may miss a question, ambient noise may obscure a phrase, and a participant may misremember the event they describe. What the source lets us claim depends on what it records, how it was created and what was omitted.
The Library of Congress Veterans History Project guidance explains why recordings and transcripts should be preserved together: a transcript makes speech searchable, but does not capture every vocal feature that can remain in audio, and time-indexing makes passages easier to locate. For school history, that suggests a practical hierarchy of work: preserve the available original under appropriate permissions, make a traceable index or transcript, and retain the distinction between speaker statements and the student’s interpretation.
Consent matters first. Recording a classmate, family member or community participant should follow the school’s directions and applicable privacy and permission requirements. If recording is not permitted, the project can still be valuable, but notes must be labelled for what they are. A teacher should not punish a learner for not having a recording when none was authorised.
Three kinds of records and what each supports
- Original recording: supports careful checking of captured wording, sequence and some audible features, subject to sound quality, completeness and consent.
- Notes made during the conversation: support an account of what the note-taker selected or understood at the time, but usually not a verbatim quotation unless words were deliberately checked.
- Notes reconstructed later: support a transparent summary of what the student remembers. They may be useful for planning a follow-up question but are weaker evidence for precise wording, dates or quantities.
Keep the creation dates separate. An interview might have happened on Monday, the notes been written on Wednesday and the essay been submitted Friday. The date of writing the notes is not necessarily the date of the historical event discussed. Similarly, ‘at 02:14’ in an audio file means two minutes and fourteen seconds into that file, not 2:14 in the afternoon.
Worked example: a fictional community reading-corner interview
Imagine a fictional, consented student interview about a neighbourhood reading-corner project. In this invented example, an interviewer records ten minutes of speech and later drafts notes. These are not the words of a real volunteer, and no real interview is being represented.
At recording time code 02:14, the interviewee says: ‘I think we started the week after the exams, but I can’t remember the date.’ At 04:32, the interviewee says: ‘Some days maybe twenty children came. Other days it was much quieter.’ The student later writes: ‘The reading corner opened on 12 November and served twenty children every day.’ That is an attractive sentence. It is also unsupported by these recorded remarks.
- Claim 1: ‘Opened on 12 November.’ The recording does not establish a precise calendar date. Correct label: date not confirmed.
- Claim 2: ‘Served twenty children every day.’ The speaker described variation and uncertainty. Correct repair: ‘The volunteer recalled that attendance sometimes reached about twenty children, with quieter days too.’
- Claim 3: The student remembers that the interviewee smiled when discussing children. If video shows that, it may be a supported observed detail; if only audio exists, the smile cannot be verified from that audio alone.
- Claim 4: ‘At 04:32 the speaker mentioned attendance.’ This is a time code within the hypothetical recording, not the clock time at which the reading corner operated.
A small arithmetic check illustrates the time-code distinction. From 02:14 to 04:32 there are 4 × 60 + 32 − (2 × 60 + 14) = 272 − 134 = 138 seconds, or 2 minutes 18 seconds between the two moments in the recording. That elapsed recording interval tells us nothing about how many days passed between opening the reading corner and its busier afternoons. Accurate computation is useful only when the quantities have the right meaning.
Now write the corrected project sentence: ‘In a recorded interview, a volunteer recalled that the reading corner began sometime after the exams, though the date was uncertain. The volunteer also recalled days with roughly twenty children and other quieter days.’ The sentence is less precise than the invented one, but stronger as evidence. It carries the uncertainty honestly.
Why students accidentally upgrade memories into facts
A tidy narrative can erase uncertainty
Students often want a beginning, middle and end. ‘I think’, ‘perhaps’ and ‘some days’ can feel untidy, so they disappear during editing. But those phrases change the strength of a claim. Removing them is not mere grammar correction; it can turn a memory into a false exact record.
A transcript can look primary even when it is edited
An edited transcript may be a highly useful derivative source. Yet its punctuation, omissions, expansion of acronyms and identification of speakers involve decisions. If words are inaudible, use a transparent notation such as ‘[inaudible]’ rather than manufacture plausible wording. If a teacher requests a clean reading copy, retain an unaltered copy or a clear audit trail wherever the approved process allows.
A later summary may mix sources
A learner may see a poster dated 12 November and then unconsciously attribute the date to the interviewee. Both sources may matter, but the claim should state its provenance: ‘The poster carries 12 November; the volunteer could not recall the opening date.’ The poster may still need its own date and authenticity checks. Do not make the interview say what another document says.
A learner diagnostic: test the evidence before the grammar
Give learners four labels: heard in the recording, written during the conversation, remembered later, and inferred from other evidence. Present five statements from an invented interview folder and ask them to label each statement’s origin before editing its wording. Then ask which could responsibly be placed inside quotation marks.
- Statement A: ‘Around the time the school holidays began,’ audible at 03:01 of the recording. Directly heard, assuming the audio is clear.
- Statement B: ‘Twenty visitors daily,’ written in notes three days later without a recording match. Reconstructed or inferred; not a verified direct quotation.
- Statement C: ‘The interview was conducted on Tuesday,’ shown by a dated consented recording log. A documentation claim, not necessarily a claim about the historical event.
- Statement D: ‘The speaker felt proud,’ written by the student after hearing an enthusiastic tone. Interpretation, not verified testimony unless the speaker explicitly said it.
- Statement E: ‘The opening took place at 9 a.m.,’ derived from a separately dated schedule. Attribute to that schedule, not to the interview.
The diagnostic reveals whether the learner mistakes one document for another, treats quotation marks as decoration, or changes uncertain language into certainty. Instruction can then address the actual problem rather than simply telling the pupil to ‘add more evidence’.
Progressive practice: guided to independent source criticism
Guided: identify the source of one sentence
Provide a 20-second invented transcript and a one-sentence student summary. Highlight phrases that are exact, paraphrased or missing. Ask: ‘Which words did the speaker actually say?’ The learner should point to a time code rather than answer from memory.
Supported: repair the overstatement
Suppose a hypothetical interviewee says ‘Perhaps five or six neighbours helped.’ A student writes ‘Six neighbours helped.’ Have the learner change it to ‘The interviewee estimated that five or six neighbours helped.’ Discuss why six is possible but not uniquely established, and why rounding this uncertainty to one number is not warranted.
Independent: triangulate without merging voices
Give the learner an audio excerpt, handwritten notes and a poster. Require a short paragraph that attributes each claim to the correct source, explains one disagreement and states what further confirmation would help. The answer should not be graded by how many dates it includes but by whether its claims can be traced to evidence.
A useful next-step routine for families and teachers
A parent can ask, ‘Is that exactly what the person said, or the meaning you remember?’ That is more productive than accusing a child of making something up. Encourage the learner to locate the audio passage if one exists, or to change the sentence to a clearly labelled paraphrase. If the date or name matters, look for a relevant document or seek a permission-respecting clarification.
A teacher can provide a compact evidence ledger: claim, source type, original source location, precise wording or summary, level of certainty and possible corroboration. In a group project, ask a second student to trace one randomly chosen claim back to its source. This tests whether the method works without relying on the writer’s confidence.
Progress is observable when a student can preserve uncertainty, distinguish time codes from event times, avoid invented quotes and repair a claim after a source check. Students do not need flawless transcripts. They need an honest route from evidence to words.
Limits and ethical boundaries
Original recordings may contain private details or material unsuitable for sharing. Use only information the participant has agreed may be used, follow school instructions, and do not publish sensitive details for a better mark. A recording can document a recollection without proving the remembered event happened exactly as described. A transcript, whether human- or software-generated, still requires appropriate checking against audio if used for exact quotations.
Frequently asked questions
Can I quote from handwritten notes if there was no recording?
You can report what your notes captured, but avoid claiming an exact verbatim quote unless the wording was carefully recorded and your method is clear. A cautious paraphrase with attribution is often stronger.
Can I fix grammar in a transcript?
For a readable edited transcript, the school’s format may allow light edits. Do not silently change meaning, remove qualifications or present an edited reconstruction as an untouched verbatim document. Keep a traceable version or edit note where permitted.
What if the recording and the interviewee’s later correction disagree?
Document both versions and their dates. A later correction is important evidence, but should not erase the original recording. Explain what was clarified and whether it changes the project’s conclusion.
Does having an audio file make an interview unbiased?
No. Questions can be leading, memories can be imperfect, microphones can miss context and participants can have partial perspectives. Good source criticism examines these limits rather than treating technology as a guarantee.
Related reading and a more transferable way to learn
A helpful companion is how to compare eyewitness accounts from different viewpoints. The separate historical-photograph evidence lesson shows why another kind of source also needs context. For diagnosis-first learning methods see the eduKate Education Hub and the Clementi tutorial reference, where learners move from explanation to independent checking.
Remember what was recorded, what was inferred and what is still unknown
History is not improved by making every memory certain. It improves when the reader can see where a statement came from, which parts were recorded, which were remembered and what evidence could change the interpretation. That is the core habit an interview project can teach long after the assignment is over.