A career filled with tricky sales pitches has reached its apex for President Trump. With Americans dubious of his handling of the coronavirus pandemic and nervous about trying to return to normal too quickly, Trump hopes to convince the country that some places or some states can begin to scale back social distancing measures that have helped contain the virus’s spread.
To do so, he and Vice President Pence on Thursday unveiled data they hoped would bolster their case that the United States is on the mend. Unfortunately, those data were, not for the first time, misleading and incomplete.
Trump began the day’s coronavirus news conference by articulating the steps that had been taken to combat the virus and the positive results that had followed. Part of his defense of rolling back containment measures has consistently been that some places in the country have only been affected by the virus to a limited degree. On Thursday, he offered a statistic aimed in the same direction.
“Our experts say the curve has flattened and the peak in new cases is behind us,” Trump said, referring to efforts to tamp down the surge in infections. “Nationwide, more than 850 counties, or nearly 30 percent of our country, have reported no new cases in the last seven days.”
That first figure, the more than 850 counties one, is accurate. Data from Johns Hopkins University as of April 15 indicate that 905 counties had added no new confirmed infections since April 8. But more than half of those counties didn’t have any cases in the first place. Of counties where the virus had been identified by April 8, more than 2,500 saw new cases over the next week.
For the most part, only a modest number of cases were added in those counties. In 762 of them, the number of cases grew by no more than 25 percent from April 8 to 15. In 421, the number of cases doubled — about the same number of counties as the ones in which no additional cases had been identified over that week.
Most of the counties in which the number of cases grew by a large amount had relatively few cases in the first place. The counties in which there were cases on April 8 and no more on April 15 averaged only 4.6 confirmed cases, suggesting that the lack of growth was less a function of containment measures than of slower spread of the virus overall.
Those figures aside, Trump’s presentation of the 850 counties as “30 percent of the country” is far from the mark — just as far as when task force member Deborah Birx said late last month that nearly 40 percent of the country hadn’t seen a significant growth in new cases. As a function of the number of states, maybe. As a function of population? Of actual Americans? Only about 7 percent fell into Birx’s category.
Trump’s number is even smaller. The counties with no cases reported from April 8 to 15 — including those that have reported no cases at all — make up less than 6 percent of the population. Counties in which more than 94 percent of the population live saw new cases crop up over that seven-day period.
When it was his turn to speak, Pence took a different tack. The task force has repeatedly pointed to the Centers for Disease Control and Prevention’s influenza-illness tracking data as an indicator of how the coronavirus has spread in the United States. Pence presented a series of maps, intending to show how ILI data indicated that the virus was receding.
He showed this one, from the week ending March 28. Areas with darker reds are ones with a higher detected level of flu-like activity. Green states saw less activity of that type during the week.
He then showed this map, from the week ending April 4. A lot more green.
In a third map, showing the week ending April 11 (which isn’t yet public) the green had spread farther. Pence didn’t make the point explicitly, but he didn’t need to: By this metric that they are tracking, things are improving.
If you stop and look at the map, though, some oddities jump out. Like Michigan, which has been a center of concern for some time. Recent data suggest a reduction in new cases there, but the maps above suggest that there were only low levels of infection in the state in late March, at a time when the task force itself was warning about the threat posed in the state.
Pulling out weekly ILI data and comparing it to the same weeks from the difficult 2017-2018 flu season, you can see a lot of variance. In many places, including coronavirus hots pots like New York and Louisiana, the ILI tracker is seeing more activity later than what was experienced two years ago. But, again, Michigan never even hit the highest tracking level this year. Florida’s current activity is almost nonexistent. Do Pence’s maps actually show what he claims they do?
Notice Rhode Island on the chart above. Just this week, Birx indicated that the state was having a difficult time, given that it is located between the hot spots of Boston and New York City. Yet it’s green on Pence’s maps, and it’s most recent ILI data shows the level of activity near the lower bound.
As of nearly two weeks ago, mind you. If we overlay the weekly increase in new coronavirus cases onto the ILI charts, the picture gets confusing, quickly. In a lot of places, the ILI activity sank even as coronavirus cases spiked. So perhaps ILI is a leading indicator? But in some places, both were elevated at the same time. What does that tell us?
Look at the row showing Kansas to Maine. Two states with high levels of activity as of April 4; two with low. All with big spikes in coronavirus cases nonetheless.
A map with a lot of green seems encouraging, but what’s it actually telling us?
The odd thing about this is that there actually are data that suggest the country has turned something of a corner. We took state-level and national data on the change in the number of confirmed cases by day and made an interactive showing how each location’s new case totals compare to its peak.
In most states and nationally, the number of new cases each day has declined in recent days. Sometimes the decline is dramatic.
Show data for
Why not make a chart that looks like this? Why not present data that actually make the point that some places are having more success than others and that governors should then feel free to reconsider the scale of distancing measures? Why elevate data that equate number of counties with a percentage of the country or which substitute aesthetically reassuring maps for more nuanced ways of presenting more direct evidence of the point?
The answers to those questions aren’t clear. All we can say with certainty is that this is not the first time Trump has been accused of presenting questionable data to make a sale.
