Start with a question, not a pile of material
Before searching for sources, write down what the episode is trying to answer. “How did this local music space become important to its community?” gives you a clearer path than “the history of a venue.” Add two or three questions a listener would reasonably ask, then decide what the episode will not cover. A manageable scope makes it easier to spot gaps and harder to drift into claims you cannot support.
Next, sketch the episode’s likely structure: what listeners need to know first, which voices could help explain it, and what would make a satisfying ending. Treat this as a working plan, not a conclusion. Research may change the story. Note any assumptions you are bringing to the topic so you can test them rather than quietly building the episode around them.
Collect sources you can trace
Keep a simple research log as you gather material. For each useful item, record its title or description, creator, publication or recording date if available, where you found it, and the specific claim it might support. Save relevant page titles, document names or time markers for audio. A log saves you from trying to reconstruct a source after you have written the script.
Match the source to the question. An official document may help establish what an organization announced; it may not explain how people experienced that decision. A participant can describe what they saw or felt, but may not know every detail of a wider timeline. When a fact matters to the episode’s central point, look for independent corroboration where possible. Several accounts repeating the same original claim are not necessarily several separate confirmations.
Check claims at the level you plan to say them
Read your outline or draft and mark statements that a listener could reasonably ask you to prove. Pay particular attention to names, dates, locations, numbers, firsts, superlatives, cause-and-effect explanations and claims about a person’s motives. Check the exact wording you intend to use. Evidence that a project was influential, for example, may not justify calling it the first or the most influential.
Separate what is documented from what is remembered, interpreted or disputed. If two credible accounts disagree, do not combine them into a falsely tidy version. You might describe the difference, narrow the claim, or leave it out if it is not essential. Use language that fits the evidence: “The organizer recalls…” tells the listener something different from “Records show…”. Being precise about uncertainty can make a story clearer without making it less engaging.
Prepare interviews to test and deepen the story
Good interview questions invite detail rather than confirmation. Instead of asking, “That event changed the scene, didn’t it?”, ask what happened, who was present, and what the guest noticed changing afterward. Ask how they know a detail and whether there are documents, recordings or other people who could help verify it. Give guests room to distinguish firsthand experience from something they heard later.
Before recording, check names and pronunciations, explain the episode’s general focus, and make clear how the conversation may be used. Keep a short list of claims that need follow-up if they arise. An interview is not a substitute for fact-checking, and a vivid memory can still contain an error. Afterward, compare consequential details with your other material before building narration around them.
Do one final script pass for accuracy and clarity
Create a claim checklist from the recording script. Beside each factual statement, note its supporting source or mark it for revision. Check that quotations preserve the speaker’s meaning and that narration does not imply more than the evidence shows. If you plan to use archival audio, music or someone else’s recording, keep permission and attribution questions on a separate production checklist rather than assuming that finding a clip means you can use it.
Finally, read the script aloud. Spoken phrasing can hide leaps that look harmless on a page: an unclear “they,” an unexplained date, or a transition that makes coincidence sound like causation. Simplify where needed, and leave space to identify sources naturally in the episode. You do not need to recite your entire research log. You do need a clear path back to your evidence if a question comes up after recording.
