Write Field Notes That Can Support an Observational-Series Pitch
Write Field Notes That Can Support an Observational-Series Pitch
You come back from a promising morning — a repair shop, a market hall, a small kitchen — with three lines in a notebook and an interest that arrived faster than the record did. By the time you reach a keyboard, the afternoon has become a series idea with a shape and a tone. That is normal. It is not a failure of discipline. The problem starts when the shape begins borrowing authority from the afternoon.
A research visit is not a sample of an episode. It is one encounter, on one day, seen from one place, by a person who was visibly standing there. Whether the pitch you write in three months can be traced back to something that happened depends almost entirely on what you do with the notebook in the next hour — and on whether you keep the difference between a thing you saw, a thing you were told, and a thing you concluded. The archived state-history guide How to Write Field Notes, from the Utah Division of State History, treats description, interpretation and reflection as separate parts of the record and recommends expanding your jottings while the context is still available (history.utah.gov; source check 9 September 2026). It is a fieldwork teaching resource from a state history office. It is not a television protocol, a consent rule, or evidence that one morning in one shop stands for anything beyond that shop.
Everything below uses one invented case so the method has something concrete to work on. No visit took place. No real shop, customer or technician is described. The details are stipulated once and then treated as fixed, which is the same discipline you would owe a real record.
Keep the first page, and write down the conditions around it
The page you wrote on site is the only thing you have that was made at the time. Do not retype it into something smoother and delete the original. Copy it forward exactly as it stands — shorthand, question marks, the words you spelled wrong — and treat that copy as the spine of the file. Everything else you add is an addition to it.
Then write a header around it, because a note without conditions cannot be assessed later. A useful header names the site, the date, the arrival and departure times, the purpose of the visit, what you were doing, and what you could not perceive from where you were. It also says which times are known and which are estimates. "Arrived 9:40 a.m., checked my phone" and "left around 11:15, a guess" are both usable; a single unmarked "9:40–11:15" flattens a checked fact and a guess into one thing.
Here is the invented record.
Field record — 3 February 2026 Site: a small appliance repair shop, market town. Name kept out of shared material. Arrived 9:40 a.m. (checked phone). Left about 11:15 a.m. (estimate). Purpose: exploratory. Possible observational series about shops that keep old machines running. What I did: stood on the customer side of the counter, roughly three metres from the bench. Did not move around. Limits: a radio played at the bench all morning. I could not hear most counter conversation. No recording was made.
Original page, written 9:52 a.m. Kettle back again? She talks, he puts the soldering iron down. Heard "...not paying..." Can't hear the rest — radio. She leaves with a box under her arm. 9:54 he's on the same board.
Two things in that jotting deserve attention before anything gets expanded. The word "again" is followed by a question mark: I did not know whether the kettle had been in before, and I wrote my uncertainty down. Good. The phrase "not paying" is in quotation marks because I wrote it as it was said, at 9:52, mid-sentence. That distinction is worth a great deal later.
The header also has to hold what your own presence might have done. If the technician glanced at you and then spoke more quietly, that is an observation, and it belongs on the page. "He spoke more quietly after glancing at me" is a description. "He was guarded because I was there" is an inference, and you cannot know how the exchange would have gone without you. Record the possibility; do not settle it.
A notebook and a camera are also not the same kind of witness. A camera records whatever was in front of it, including things you failed to notice. Your page records what you decided to write down, which is exactly why the conditions around the page matter so much.
Expand what happened before you say what it meant
Expand while the encounter is still available to you, and date the expansion. The gap between the jotting and the expansion — five hours, three days — becomes part of the record's strength, and a reader can weigh it later. Expansion is not the moment to convert a shorthand page into a story. It is the moment to add the sequence back in: who moved what, who interrupted which task, who left the room, in what order, over how long.
Then give the interpretation its own labelled space. The state-history guidance separates description, interpretation and reflection for the same reason a lab notebook separates a reading from a conclusion. "She stepped away after the request" and "she rejected the request" do different work. The first is something you could have photographed. The second is a claim about her intention, and your record supports it only if she said so.
The invented evening expansion might read:
Description (the same day, written 4:20–5:00 p.m.) A woman in a green coat came to the counter carrying a kettle and set it down. The technician put his soldering iron down, wiped his hands on a cloth, and came to the counter. They spoke. At one point I heard the words "not paying." The rest was lost to the radio and the distance. At 9:54 he was back at the bench with the same board; by then she had left, carrying a flat package under her arm. Correction, added 4 February: the page of 9:52 says "box." The "flat package" above is my later reading of the same object. The original wording stays as it is.
Interpretation (the same evening, mine) The counter may be where this shop decides what it will take on, and the bench may be waiting on that decision.
Reflection (the same evening) I stood where a customer stands, so I could see the counter and not the bench. Next time I would ask to sit where the work is.
Keep unresolved alternatives unresolved rather than compressing them into a cleaner scene. If the technician went to the back of the shop after she left, and you heard a drawer open but could not see which one, write both possibilities and mark the limit. A record that admits two readings is more useful than one that picks the better-sounding one and hides the choice.
Match your quotation marks to what you can actually support
Speech is where field notes most often inflate. There are three tiers, and they are not interchangeable. First, a recording you have made and are permitted to use — the strongest tier, and a separate conversation about consent, not something this article settles. Second, a contemporaneous note: you wrote the words while or immediately after they were said, at the site, and the page shows it. Third, recollection, hours or days later, which is real evidence about your memory and not about their words.
Now watch the invented expansion do the wrong thing. An earlier version of that evening paragraph read:
She said, "I'm not paying for that again," and put the kettle down.
Nothing in the record supports those quoted words. I heard three words. The rest — the completion of the clause, the "again," the tidiness of the whole line — was assembled at 5 p.m. from a fragment and an impression, and then dressed in quotation marks that promised a speaker had said exactly those words in that order. The revision keeps what the page can carry:
I heard the words "not paying" at the time; they are on the original page from 9:52. I cannot supply the rest of the sentence. As a recollection, written the same evening and clearly mine: my impression was that she did not want to pay again for work on the kettle. I cannot support that from anything I heard.
Note what the correction does not do. It does not throw away the lead. The impression is still there, dated, marked as an impression, and it still points somewhere — it just no longer masquerades as speech. That is the difference between a weak record and a dishonest one.
Corrections follow the same rule. If you decide two days later that the flat package was more likely a folded carton, add a dated line saying so and leave the first wording standing. An invisible replacement makes the earlier record look more precise than it was, and the next person reading the file has no way to see that a judgment changed.
This matters downstream because quotation marks carry a promise. On a pitch page, they tell a commissioner that a named or unnamed person said those words. If that promise fails once, every other claim in the same document becomes negotiable.
Trace one pitch claim back through the layers
Now take a claim you would like to make and walk it backwards through the record, word by word.
The tempting version writes itself: returns constantly derail the work at the bench. Every word of that is doing something the record did not authorize. "Constantly" needs a frequency, and the file contains one return. "Returns," plural, has one instance. "Derail" suggests work abandoned or lost; what the two clock readings show is a pause of about two minutes, after which the technician picked up the same board. And the fact that the kettle had been in before — the question mark in the jotting — has quietly hardened into a series about regulars. A guess with a question mark became a premise with teeth, somewhere between the notebook and the pitch page.
The narrow version is shorter and much harder to attack:
One morning, one counter return paused bench work for about two minutes, and the same board was resumed.
From there, a pitch sentence can stay honest and still be interesting: We want to follow how one small repair shop decides what to take on at the counter, and what that decision does to the bench. On a single morning we saw one return stop work for about two minutes; how typical that is remains open.
That leaves the shape of a research question without asserting an outcome, and it hands the commissioner something real to fund: not a promise that conflict happens weekly, but a question the visit could not settle. One observed interruption cannot establish a pattern, and a note about a single event is not footage of an episode. The follow-up question is the deliverable here — how often does a counter conversation interrupt bench work, and who decides when it is worth stopping? — and further time in the shop, or a conversation with the people who run it, is what would answer it. That is a decision for after the pitch, not something the pitch page should answer on the shop's behalf.
While you are doing this, sort your material into what can travel and what stays home. Names, faces that would identify someone, the street address, anything a person said in a moment they did not expect to be documented — none of that goes into shared pitch material unless it is genuinely needed and someone with the standing to agree has agreed. The customer in the green coat becomes "a customer." The shop becomes "a small appliance repair shop in a market town." Your most dramatic interpretation stays in your own file, where it may be valuable later, and the limited observation plus the open question go forward in its place.
Where the file should end for now
A field record is finished enough when it can hand three things to the next reader without a translator.
Dated observation, 3 February 2026. By 9:52 a.m., when I wrote the page, a customer had set a kettle on the repair counter. The technician put down his soldering iron and came to the counter. The exchange lasted about two minutes; a radio at the bench meant I could not hear it. Her departure is not timed on the page: at 9:54 he was back at the same board, and she had already left with the package.
Interpretation, added the same evening, mine alone. The counter may be where this shop's real scheduling happens, and the bench may be downstream of it.
Question worth checking. How often does a counter conversation stop work at the bench, and who decides when it is worth stopping?
Frequently asked questions
Why keep the original field page and write conditions around it?
The page written on site is the only record made at the time; copy it forward exactly rather than retyping it into something smoother and deleting the original. A header should name the site, date, arrival and departure times, purpose, what you did, what you could not perceive, and which times are known versus estimated. Without those conditions, a reader later cannot assess the note.
What is the difference between description, interpretation and reflection?
Description is what could have been photographed, interpretation is a claim about meaning or intention, and reflection is your own position or next move. They should get separate labelled spaces. 'She stepped away after the request' is description; 'she rejected the request' is an interpretation that the record supports only if she said so.
How should quotations be handled if you only caught a fragment?
There are three tiers: a permitted recording, a contemporaneous note written at the site, and later recollection. Quotation marks carry a promise that someone said those words in that order. If you heard only 'not paying,' you can keep that fragment as a contemporaneous note, but you cannot supply the rest of the sentence or present an evening impression as quoted speech. Corrections should be dated and the original wording left standing.
How can one visit support an honest pitch claim?
Trace the claim backwards through the record. 'Returns constantly derail work' overstates a file with one return, a pause of about two minutes, and the same board resumed. A narrower pitch can say you want to follow how the shop decides what to take on and what that decision does to the bench, while noting that one observed interruption cannot establish a pattern and a note is not footage of an episode.
What should stay out of shared pitch material, and what is the source guide's role?
Names, identifying faces, street addresses, and statements made in moments people did not expect to be documented should stay out unless genuinely needed and someone with standing has agreed. The archived state-history guide on field notes treats description, interpretation and reflection as separate and recommends expanding jottings while context is available; it is a fieldwork teaching resource, not a television protocol, consent rule, or evidence that one morning stands for anything beyond that shop.