A paper looks finished long before it is.
It already has headings. Introduction. Methods. Results. Discussion. The Results heading is especially confident. It sits there as if numbers were a matter of arriving at the right page.
By Week 4, we had not arrived.
The stroke file was still somewhere else. There were no coefficients to graph, mourn, or accidentally treat as a personality. From the outside, the week looked empty. I had pictured a cancelled laboratory: the instruments waiting, me waiting, perhaps another pass through papers I already knew.
Instead I wrote Methods.
This is a less glamorous sentence than it sounds. A methods section is the part of a paper most people skip. It is also the only part you can honestly finish before the data exist. You can say what a month is. You can say which weather series you will use. You can say that a count of hot nights is not the same thing as a measure of how hot those nights were. You can say, in advance, that a monthly pattern is not a daily trigger.
Hogan had already written the weather paragraph. He had also opened a shared manuscript, which is a quieter kind of collaboration than sending one another competing drafts. My job was the rest of the instructions: how the months would be assembled, how pollution would enter if it entered at all, why a city-wide population number is not the group of patients still at risk of a first event, and what the models would be allowed to ask.
It is a peculiar kind of writing. You are describing an analysis that has not happened. You have to sound as if you know what you will do, because later numbers need somewhere honest to sit. You must not sound as if you already know what the numbers will say.
Hogan has been trying to teach me that second part as a writing problem, not only a weather problem. Student prose wants to arrive with every paper it has ever met. Methods prose has to put the tools on the table and then stop.
Roro’s section was still a gap with his name on it. The health data are his to construct and to describe. I could leave a bridge sentence. I could not invent the dictionary.
Meanwhile the Results heading sat there with nothing under it.
I used to think white space in a paper meant you had fallen behind.
I am less sure now. It can also mean you have refused to get ahead of a file that has not yet become a file.
Around that blank I checked the parts that could be checked. The temperature months ran from January 2013 through December 2023. Influenza summaries covered most of those months and not quite all of them. The missing influenza months would stay missing. They would never be coded as a polite zero.
None of that is an answer. It is a refusal to pretend that waiting is the same as running.
Week 3 had already taught me that waiting is full of verbs. Week 4 added a less glamorous one: write. Write the instructions carefully enough that, when there is finally something to run, we will not have to invent the rules in the same afternoon as the numbers.
Week 4 ended with a methods section that was longer than it had been, and a results section that was exactly as empty as the data.
The waiting was not empty.
It had a heading in it, and nothing underneath, on purpose.