Sunday, December 31, 2017

Why understanding context when asking a question is important

In a recent forum discussion, one person asked if Chile celebrated January 1 as a holiday. It seemed to many who responded to be a rather odd question, given the historical and cultural contexts. But then someone had to bring up the point that "well some countries don't observe a national holiday on January 1."

Yes, that's true, but there are a few things that are odd about the question, and even more infuriating about that "well some countries" response.

1. This is a question that is reliably google-able. From the Wikipedia page that comes up as the first result, we get.
Public Holidays in Chile:
January 1 New Year's Day Año Nuevo Blue laws forbid opening of most services.


Looking down the list for the first Chilean (e.g., *.cl) webpage, we find feriados.cl, which says:
Monday, January 1 New Year's Day
Inalienable
Secular
And what does "Inalienable" mean? If one does a little bit of googling (this time in Spanish), one finds that it means basically everything shuts down, except emergency services, hotels, gas stations, and family-run businesses. But even without that knowledge, a contextual assessment (plus a working knowledge of English) will inform you that New Year's Day is a pretty big holiday in Chile.

So yeah. That's easy to see.

2. The "not every country celebrates New Year's Day" is one of those facts that is true, but is useless when said out of context. What do I mean?

It is true that there are countries that don't celebrate some sort of holiday on 1 January. Strangely, a quick google search didn't pull up anything, so I ended up using the holiday information from https://www.timeanddate.com/holidays/ to build a world map that showed the countries that do and don't have some sort of national observance on 1 January:

Basically, blue countries are those that celebrate New Year's Day on 1 January. (They might celebrate a cultural new year at some other time during the calendar year, but on 1 January, there is a celebration of an "international" New Year's Day.) Countries in yellow celebrate something in addition to - or in place of - New Year's Day.

The nine countries in red don't have any official celebrations listed for 1 January.

Note a few things:
i) All countries in South America celebrate 1 January as a holiday.
ii) All countries in the entire New World celebrate 1 January as a holidary.
iii) All countries that used to be part of the Spanish Empire celebrate 1 January as a holiday.
iv) Most of the countries that don't celebrate 1 January as a holiday are concentrated around one part of the globe.

So there is no nation in the geographic or cultural contexts in which Chile finds itself that don't celebrate 1 January as a holiday. In fact, since most countries in the entire world do celebrate 1 January as a holiday, saying "well some countries don't celebrate it" is true, but basically useless in most contexts.

In sum, we learned that Chile does have a national holiday on 1 January. The name of that national holiday - in English - is New Year's Day. And saying "well some countries don't have a holiday on 1 January" is true, but basically useless information in many contexts.

Happy New Year!

Thursday, September 07, 2017

Hurricane Irma, warm oceans, and expanding the Saffir-Simpson Hurricane Scale

Back in 2011, I wrote about the current five-category hurricane system that the US uses (known as the Saffir-Simpson Hurricane Scale), noting that the foundational logic of the scale was based on structural engineering questions:
a former NOAA hurricane center administrator and co-inventor of the SSHS that, "there is no reason for a Category 6 on the Saffir-Simpson Scale because it is designed to measure the potential damage of a hurricane to manmade structures. If the wind speed of the hurricane is above 155 mph (249 km/h), then the damage to a building will be 'serious no matter how well it's engineered'."
The current scale tops out at a "Category 5," which is any sustained wind speed above 155 mph. However, if one uses the threshold values for Categories 1 through 5 to develop a regression equation, it is possible to extend this relationship ever outward. Specifically, a revised category scale would be something like this:
Category 1: <95mph
Category 2: 96-110mph
Category 3: 111-130mph
Category 4: 131-150mph
Category 5: 151-175mph
Category 6: 176-205mph
Category 7: 206-235mph
Back in 2011, Hurricane Camille had sustained wind speeds of 175 mph, which is what prompted me to write that post. Currently, Hurricane Irma is reported as having sustained wind speeds of 185 mph, making it the strongest Atlantic hurricane in recorded history. However, based on the current hurricane scale, both Camille and Irma are classified as Category 5, even though Irma is obviously far stronger than Camille (which was - itself - a massive hurricane).

Indeed, the current system is fundamentally limited and fundamentally limiting, since one loses any sense of comparative scale once you enter "Category 5." And what would it hurt to look at adding a "Category 6," especially if warming waters are known to lead to stronger and more sustained hurricanes? Indeed, with warming oceans, hurricanes that will reach sustained wind speeds between 175 and 205 mph will not be theoretical. Indeed, Hurricane Irma is proof-positive that such hurricanes can and will form.

But so what? Why would that matter?

Well, in the US, the SSHS is a widely known and used shorthand for hurricane strength. It's something that people latch on to when discussing preparedness measures and when making comparisons against past events. But if the maximum scale is effectively open-ended, the designation "Category 5" will be shared by a hurricane with wind speeds of 155 mph and another with winds speeds of 185 mph (like Hurricane Irma). And the simple fact is that wind speeds of 185 mph are fundamentally different than wind speeds of 155 mph, and placing both in the same open-ended category will not help with making short-hand comparisons that would be equivalent to comparing a Category 4 hurricane against a Category 3 hurricane.

The way we categorize natural phenomena is important, since it structures the way that we view and respond to the world, and if we continue to use a hurricane classification system whose comparative utility declines into a future that is expected to have stronger hurricanes, that can impact the type of public response given to future storms.

Monday, May 22, 2017

Some things to consider when thinking about political trends

I don't write so much these days on this blog, but that doesn't mean that I've stopped thinking about some of the themes that I have written about in the past. Today, I want to write a little about US national politics and trends. Specifically, I want to write a little about the US Presidential elections, and what it means for Democrats.

This urge was spurred on by an article I read at Vox ("What right-wing populism?"). Okay, the author makes some shortcuts by basically equating liberals with progressives and those with Democrats, but given the dominant duality of the US political system, that short-hand has become so commonplace that it is effectively taken as synonymous in many corners. However, he makes some good data-backed rhetorical points about the public wanting government spending even as they might say that they don't want government spending.

But this got me thinking about the "Trump Revolution" (of 2016) and the earlier "Reagan Revolution" (that happened in 1980). In both cases, the narrative was that working class/blue-collar voters moved from the Democrats to the Republicans. But in 2016, that story never rang true for me. After all, Hillary won the popular vote (and - indeed - the polls predicted the popular win quite accurately). And the increase in Republican votes between 2012 and 2016 was basically a rise of 2 million, but less than 1 million when Bush ran to his first popular vote victory in 2004. But - because the US population was lower in 2004 than 2016, this "less than 1 million than Bush in 2004" figure means that Bush actually won a greater percentage of the vote (29.06% of all voting-age Americans) than Trump (26.74% of all voting-age Americans). But let me unpack that a little bit, because those numbers seem too small.

What I wanted to do was to create an assessment of how many voting-age Americans did each party's candidate win in each POTUS election? Now, in every year, not all people vote (indeed, the average voter turnout for a POTUS election since 1940 is 56.3%). Therefore, if there is a year where the voter turnout is only 50% (like 1988), then a victory of 53.4% (which George H.W. Bush got) means that only 27.6% of voting-age Americans actually cast a vote for Papa Bush. Indeed, counted this way, most POTUS victories since 1940 were won with less than 1/3 of all voting-age Americans actually casting a ballot for the victor, save for four Presidents: FDR (1940, 34.4%), Ike (1952, 35.1%; 1956 35.0%), JFK (1964, 37.97%), and Tricky Dick (1972, 34.11%).

Okay, so what, though?

Well, if there was a major shift from Democrats to Republicans in 1980 with Reagan and 2016 with Trump, then there should have been a major shift in the share of voting-age Americans that the Republicans won in those years, and a consonant decline in Democrats compared to each previous election. With Reagan, we do see this:

Republicans: 26.24% (1976)     27.14% (1980)     +0.9%
Democrats: 27.36% (1976)     21.93% (1980)     -5.43%

But with Trump? Not so much:

Republicans: 26.37% (2012)     26.74% (2016)     +0.37%
Democrats: 28.53% (2012)     27.96% (2016)     -0.57%

So what's going on? Basically, the Republicans did gain more votes since the previous election, but 2016 was nothing like 1980. The change in Democratic vote-share in 2016 was nowhere near the enormous shift seen in 1980 moving away from Carter. And we see this in shifts in the popular vote from 2012:

Republicans: +2,050,000 votes compared to 2012
Democrats: -60,000 votes compared to 2012.

But, given the simple fact that Trump's share of voting-age Americans (26.74%) is basically the same as the average GOP vote-share since 2000 (26.78%) means that the power of Trump/Pence in the elections was not really any different from Bush/Cheney, McCain/Palin, or Romney/Ryan.

The only real difference is on the Democrats' side.

So 2016 isn't so much a story of conservative or right-wing America surging, but rather a story of liberal or left-wing American choosing to stay home. This, together with the Vox article, strongly suggest that - if liberal/progressive/Democratic Americans actually got out to vote - then there would be a dramatic across-the-board shift. Luckily for conservatives, the percentage of liberals/progressives who go and vote is lower than the percentage of voting conservatives.

(Note: All numbers are drawn from a simple set of calculations using voting statistics drawn from Wikipedia pages on presidential elections between 1940 and 2016.)