More fun with numbers and the media

I’ve found USA TODAY’s “Coronavirus Watch” newsletter very useful during the pandemic. But today’s lede is a perfect example of the type of cherry picking I mentioned yesterday:

Is there cause for concern? Numbers are going up, right? Well, yes, but how much?

“Alaska and Arkansas more than doubled cases in the last week.” Okay, is that from 50 cases to 100 cases, or is it from 1000 to 2000 cases? Context matters.

“In Missouri, hospitalizations jumped by nearly 30% over the weekend.” Considering the map I included yesterday, this isn’t surprising, and a surge in hospitalizations is what we want to avoid so we don’t overwhelm our facilities and, more importantly, our healthcare workers. But is this from 100 to 130 people hospitalized, or from 1000 to 1300? Again, context matters.

“Mississippi’s fully vaccinated rate of 31% is the lowest in the nation.” Okay, no complaints with the reporting here; that’s a fact. It’s also unbelievably pathetic. Way to go, Mississippi.

Percentage changes, or deltas, are among the most misunderstood and easy to manipulate statistical measurements. Always ask what the underlying data is when you read something like these statements. Context matters.

When Montréal cops move to the Ontario countryside

I’ve never written much about my impressions of television shows or movies, largely because I don’t usually watch many of them. I used to see one movie a year in a theater, just to say I’d seen one, but I think I didn’t see any in 2019 and last year was a wash for obvious reasons.

I have been watching a bit more television lately, though, because you can’t play World of Warcraft every day (though I’ve certainly tried). Shows I’ve enjoyed during the pandemic have included The Mandalorian, Ted Lasso, For All Mankind, and a show my daughter tried to get me to watch for two years, Letterkenny. I feel bad for not listening to Erin on that one; Letterkenny is hilarious (though not family-friendly in case someone with younger kids is looking for something to watch with them). I’ve watched all nine seasons, six or seven episodes in each, at least twice now, and can’t wait for seasons 10 and 11 to drop late this year.

Letterkenny’s creator, Jared Keeso, was one of the leads in a police procedural set in Montréal, 19-2, which aired from 2014 to 2017 on Canada’s Bravo network and then on CTV. Keeso and Adrian Holmes play two street cops who get stuck with each other – not an unusual setup for a cop show, but this mismatch isn’t played for laughs. In fact, each episode gives you more reasons not to like nearly all of the characters, including Keeso’s Ben Chartier. It’s surprisingly deep and I’m enjoying it.

Interestingly, the English-language version of 19-2 is based on a French-Canadian series with the same name that aired from 2011 to 2015 on Télévision de Radio-Canada and was created by Réal Bossé. The first two seasons of the English version use the same scripts and characters, just translated, while the third season goes off in its own direction. Only one actor appears as the same character in both versions: Benz Antoine, who plays Tyler Joseph. Another regular in the French version, Catherine Bérubé, played a recurring paramedic character in the English series.

Because of Keeso’s connection to 19-2, a number of actors from that show have appeared on Letterkenny, often in roles very different from the ones in the police show. Adrian Holmes, who played Keeso’s partner Nick Barron on 19-2, played the “cousin’t” of Letterkenny’s bar owner Gail in a couple of episodes; Tyler Hynes, who played a rookie cop on 19-2 played Dierks, a rather slimy womanizer on Letterkenny; Alexander De Jordy was in season 3 of 19-2 and was a major character in the first two seasons of Letterkenny before moving on. And Dan Petronijevic, whose Jean-Marc Brouillard was a womanizing cheater and spouse abuser on 19-2, plays McMurray, a womanizer in an open marriage and the inability to keep quiet about it on Letterkenny. Well, maybe his character isn’t so different, except for the abuse.

The English version of 19-2 is available on Acorn TV. (J’essaie de voir la version française sur un service de streaming mais je n’ai pas trouvé de moyen de le faire. Il serait intéressant de regarder les épisodes de retour en anglais, puis en français.)

No “late surges” or “comebacks”: Elections are not football games

I touched on this yesterday, but since it doesn’t look like we’ll be moving away from mail-in voting any time soon, I think we need to reconsider how presidential elections are covered. A big reason why Trump and his supporters can claim that the election is being “stolen” is because the initial numbers on Election Day favored them, while mail-in votes have been significantly skewed toward Biden, so the final results keep inching toward a Biden victory.

That’s partially the president’s fault: he spent so much time deriding mail-in voting (despite doing it himself) that much of his base wouldn’t consider taking advantage of it for this election. In at least one case, Trump supporters held a protest in Michigan where they burned the absentee ballot applications they’d automatically received in the mail:

So we expected a “red mirage” on Election Night, and had a pretty good idea that the initial numbers would shift blue as the mail-in votes were counted. This was made worse by the legislatures in Wisconsin, Michigan, and Pennsylvania, who refused to allow early counting of those ballots, likely to support the Trump campaign’s narrative that votes counted after Election Day would be invalid and fraudulent.

(The Trump campaign is currently trying to have it both ways: In the states they’re slightly behind, they want all of the votes counted. In the ones they’re ahead in, they want to “stop the count.” Basically, the message is that the only valid vote is one for Trump.)

If elections have changed, why should we keep covering them the same way? The networks have an agreement between themselves not to call the results of individual states’ voting until the polls close in each state (but not necessarily the overall winner if it’s obvious they’ve collected the necessary electoral votes). I realize an informal agreement to not report preliminary election numbers until a winner can be confirmed would be extremely hard to enforce and might even be a violation of the First Amendment. But are clerks’ offices and secretary of state offices required by law to provide those preliminary results? If not, might they be able to take a day or two or even 72 hours to do all (or at least most, say 95%) of the counting before releasing results?

It would reduce the impression that elections are football games, with the score going back and forth until a last-minute comeback wins or falls short. After all, if a candidate ends up with more votes after they’ve all been counted, they always had the lead – always – not just after some imaginary surge at the “end.” Ironically, the “last-minute surge” is mostly made up of votes that were cast well in advance of Election Day, making that description of the changing numbers even more inaccurate.

I’m not blaming the media. We need them to tell us what’s going on, because they have access that the average citizen lacks. But it would be careless of them, especially the major television networks that still command a significant percentage of viewers on Election Day, to not consider that the way they cover elections might also be contributing to our current problems.