There are two factors that make it particularly hard to track the spread of new coronavirus infections. The first is that symptoms can appear up to two weeks after a person first contracts the virus, making it much harder to establish the point of infection. The second is that some people don’t show symptoms at all, making it harder to find new cases in the first place.

Given that, tracing the outbreak of coronavirus cases apparently linked to the White House — an outbreak that has now spread to nearly three dozen people — would have been difficult under the best conditions. That the White House doesn’t appear to be particularly eager to aid in that effort simply introduces another layer of difficulty, obscuring the public’s ability to understand how President Trump and his team contracted the virus and where it may have spread.

Let’s therefore walk through what we do know.

It’s useful to first establish a baseline of what one might expect from a typical infection. Using data from the Centers for Disease Control and Prevention, MIT and research published by the Annals of Internal Medicine, we put together this rough look at how infections progress.

In a typical case, symptoms appear five to seven days in. Individuals may be contagious for two or three days before symptoms appear and are contagious for a period afterward. In all of this, conditions vary: The CDC says that more severe cases can be contagious for as long as 20 days after the onset of symptoms, for example, though most people are not contagious after 10 days. Symptoms themselves can resolve within days, leading most people to feel fine after a week.

Those who are infected may test negative for the virus between 17 and 20 days after diagnosis. Notice, though, that they’re also more likely than not to test negative before symptoms kick in: Until Day 4, two-thirds of tests return negative, according to one study, despite the patient probably being both contagious and about to show symptoms.

We can take that estimated time frame and build out an (admittedly complex) grid contrasting known White House infections to what we’d expect from a typical course of infection.

The New York Times reported Friday that several White House staffers had tested positive for the virus “nearly three weeks ago,” putting that cluster somewhere in the vicinity of Sept. 20 or 21. It’s not clear whether they had direct contact with the president or the first lady.

The first person to test positive from Trump’s political circle was Republican National Committee Chairwoman Ronna McDaniel. As The Washington Post reported Thursday, McDaniel began feeling sick Sept. 28, testing positive Sept. 30. (For this assessment, we use the onset of symptoms, when available, as the point at which we work backward to try to identify infection.) Her diagnosis became public Oct. 2 after she informed the White House.

McDaniel had participated in a fundraiser with Trump at his hotel in Washington on Sept. 25. If hers was a typical case, that might well mean that she was contagious at the time. Neither she nor Trump wore a mask at the event.

Trump and the first lady both tested positive for the virus Oct. 1, announcing it early in the morning Oct. 2. His schedule was busy over the preceding week, including events in multiple states. It’s the events centered around the White House and the debate Sept. 29 that have attracted the most attention however. Those included:

  • The Sept. 25 fundraiser.
  • His Sept. 26 announcement that he was nominating Amy Coney Barrett to sit on the Supreme Court. The announcement was preceded by meetings inside the White House at which attendees were not masked. After the announcement, he flew to Pennsylvania for a rally.
  • A session Sept. 27 to prepare for the debate, followed by a reception for Gold Star families that evening.
  • The debate itself was Sept. 29 in Cleveland.

It’s not clear when Trump began feeling symptoms. It seems possible that it was on Wednesday, Sept. 30, when he cut off a rally speech in Minnesota after an unusually brief period. (That his doctor cleared him to participate in public events Saturday suggests that he might have been symptomatic Sept. 30, given the 10-day window identified by the CDC.) We know that he experienced symptoms Oct. 2 and Oct. 3, according to his doctor, though his doctor says no symptoms have manifested since.

You can see from the above chart that Trump could have been infected at the Sept. 25 fundraiser — particularly if he was symptomatic Sept. 30. It could have occurred earlier or later, too, given the variability of how the illness progresses. He might have been contagious over the next few days — or not.

Given when he tested positive, Trump was almost certainly contagious at the debate Sept. 29 and on the day following. If he had been tested the day of the debate, though, it would probably have come back negative if he contracted the virus Sept. 25.

Trump’s adviser Hope Hicks was also experiencing symptoms Sept. 30, and tested positive the following day. She wasn’t at the events surrounding the Barrett nomination, but was at debate prep and flew to the debate with Trump. (The identification of attendance at these events is from this Times report.)

A number of people who’d participated in debate preparations or attended the Barrett nomination events tested positive the day after Trump and Hicks. They include Trump adviser Kellyanne Conway, campaign manager Bill Stepien and Sens. Thom Tillis (R-N.C.) and Mike Lee (R-Utah). Conway, Tillis and Lee were all at the Rose Garden ceremony announcing Barrett and at meetings and receptions inside the White House that same day. Stepien wasn’t, though he traveled to the debate with Trump on Air Force One.

On Oct. 2, New York Times reporter Michael Shear also tested positive. He’d traveled to Pennsylvania with Trump for the rally Sept. 26 and was experiencing symptoms six days later.

The following day, two more people tested positive. One was Sen. Ron Johnson (R-Wis.), who wasn’t at any of the White House events. The other was former New Jersey governor Chris Christie, who remains in the hospital. Christie was at the Barrett events and participated in debate prep Sept. 27. But on the day of the debate, four days before his positive test, he’d tested negative.

Three more senior White House staffers also tested positive over the next three days. Trump’s personal assistant Nicholas Luna tested positive Sunday. Press secretary Kayleigh McEnany tested positive Monday, and Stephen Miller, a senior adviser to Trump, tested positive Tuesday. Both McEnany and Miller said that they’d tested negative each day since Oct. 1 before their positive tests, a good demonstration of the lag time between infection and detectability.

On Monday, there was another prominent positive test: Adm. Charles Ray, vice commandant of the United States Coast Guard, tested positive for the virus after showing symptoms the day before. Ray had attended the Gold Star families event at the White House, were no masks or social distancing were evident.

Again, that Ray attended this event doesn’t mean that he contracted the virus there, any more than it’s necessarily the case, based on what we know, that any of the events at the White House were the ones at which people were infected. It’s extremely unlikely that there were dozens of independent infections among people who work together at the same time, but it’s not impossible.

This uncertainty, of course, is one reason that the administration is not enthusiastic about helping in a contact-tracing program. Without documenting who might have been exposed to the virus at the White House, there’s at least some deniability, however implausible.

More important, though, is the effect of the virus on Trump himself. We’re left to guess when he might have contracted it and from whom, in an effort to parse out where it might have spread and how long both he and those with whom he comes into contact are at risk. Amazingly, we know more about Trump’s taxes at this point than we do about his health, particularly given the ongoing questions about his sudden trip to the hospital a year ago.

Learning the pattern of Trump’s tests for the virus over the past few weeks would offer a great deal of information that could clarify the situation. But it might also make clear that it was the president’s lax approach to containing the virus that contributed to his infection or that he knowingly put people at risk of infection.

For the White House, it seems, uncertainty is better than certainty when the certainty may be politically problematic — even if it also means that the health of other people might be at risk.