How to Mislead with Numbers

As someone who made their living interpreting numbers and data, it bothers me to see things presented that are factually true but not in the appropriate context. Since the beginnings of the pandemic, the media has been doing this non stop. Feeding a worried public numbers and data without any meaning. Generating scary click-bait headlines to keep you in a state of fear. And it still goes on.

Case in point, saw this headline on the twitter site of the Guardian “First Thing: US in race against Delta virus as Covid cases rise in nearly half of the states.” Going to the actual article reveals little additional information. The first paragraph repeats the headline while adding the scary line “fears of another surge.” Who’s afraid? Well they don’t tell you. The rest of the article is on vaccines and why people aren’t taking them which really has nothing to do with the first part but they’ve already set their hooks into you. Cases are rising! Delta is on the way! Another surge is coming!

Factually, this is true. Cases were up last week in a lot of places. By my count, they were up in more than half of the states but if you use different sources, you can get different information.

Let me also be clear in that I do think there we are seeing an underlying increase in cases here in the US. There are plenty of signals suggesting that but are we really in a “race against Delta”?

You can’t determine a trend from a single data point. A weekly increase can be just noise in who and how many were tested. If cases were to double in a week, then that’s a pretty strong signal that something is up but that isn’t what is going on here.

So now I’m going to do what the quote above suggests but it is also to put some of these state increases into context.

• Nebraska – in percentage terms, cases were up 32% last week which puts Nebraska in fourth place in the US. The actual number of cases went from 206 to 271. In late March, Nebraska has a week with 3694 cases. Call me when we get to that level.
• In New Hampshire, cases increased from 116 to 135. I don’t think an extra 19 cases is really “racing against Delta”.
• In Iowa, cases when from 522 to 594. Back in the early spring, weekly cases peaked at 3788 and that was still way below the winter totals. Long way to go still.
• In Maryland, cases increased from 418 to 454. Is an extra 36 cases going to break the system?
• In DC, the increase was from 59 to 64.
• In Wyoming, the numbers went from 418 to 445.
• In Wisconsin, the increase was from 530 to 560. 30 extra cases? Is that a “race”.
• In Delaware, the increase was from 179 to 189. 10 whole cases!

Yes, I was cherry picking the states with small numbers. But do those numbers represent an emerging problem? I’d argue no. So when you say that cases are up but don’t put the increase in context with where we’ve been, it is misleading.

The state with the largest increase last week was Louisiana where cases increased from 1925 to 2726 which is a 42% increase. Last year, in the same week, they had 8520 cases and had seen cases jump by 74% and 58% the previous two weeks. In the next week, cases again jumped from 8520 to 13514. That’s a surge by the way.

Yes, cases were up last week. Right now, there is no evidence suggesting that we’ll see a “surge” like we did last summer. Until we’ve actually established a true trend, headlines like the one above do a lot more harm than good.