Comments, observations and thoughts from two bloggers on applied statistics, higher education and epidemiology. Joseph is an associate professor. Mark is a professional statistician and former math teacher.
Picking up on Joseph's recent post (which in turn picked up on Jared Bernstein's earlier post), specifically this part:
When the benefits of trade are broadly spread then everyone benefits. But capturing all of the benefits and then using that political power to seize additional benefits is a great way to get very powerful but it runs the risk of undermining the political calculation. After all, if trade is the way that the "rich get richer" and the "poor get even less" then it starts to look like a very bad deal.
Let's focus on the first sentence. This is a completely conventional assertion and it's entirely valid if you make certain standard (and always implicit) assumptions about what it means to benefit from a transaction. Unfortunately, the reasoning does not stand well when wee start tweaking that assumption.
A few years ago, we ran a post discussing different ways of viewing wealth. One of the approaches we covered was ordinal wealth, the idea that, in some situations, total wealth might be less important (or a less useful metric) than wealth-rank. In terms of social status, political power, and personal satisfaction, being the richest man in town with say $10 million in the bank might be preferable to having $15 million but not breaking the top five.
There is no obvious reason why absolute wealth should be a better central metric than ordinal wealth and other relative measures – you can almost certainly find cases where each is appropriate – but if we do allow for the possibility that relative measures might sometimes work better, all sorts of cherished economic truths start to look fairly shaky.
Maybe I'm missing something, but pretty much all the assurances we've heard about how enlightened self interest will keep us on the right track seem to assume that rational actors will seek to optimize absolute wealth. If the rich and powerful are more concerned with maximizing relative position, it's difficult to see where that enlightenment would come in.
Following up on the previous post, here is a bit of background on the recent, widely-mocked graphs from Thomas Friedman. Though it was the first that prompted the most derision, the second graph might actually merit a bit more attention.
At this point, it is helpful to have a bit of background on Friedman's ideas about modern pedagogy. Friedman
has enormous faith in the power of technology to revolutionize
education. Unfortunately, he appears to have no idea how antiquated his
view of educational technology is, or how badly his ideas have fared
in the past. Here's what we had to say on the subject back in 2013.
While pulling together some material for a MOOC thread, I came across
these two passages that illustrated how old much of today's cutting
edge educational thinking really is.
"
Experts in given fields broadcast lessons for pupils within the many
schoolrooms of the public school system, asking questions, suggesting
readings, making assignments, and conducting test. This mechanize is
education and leaves the local teacher only the tasks of preparing for
the broadcast and keeping order in the classroom."
For
relatively little money, the U.S. could rent space in an Egyptian
village, install two dozen computers and high-speed satellite Internet
access, hire a local teacher as a facilitator, and invite in any
Egyptian who wanted to take online courses with the best professors in
the world, subtitled in Arabic.
I know I've made this point before,
but there are a lot of relevant precedents to the MOOCs, and we would
have a more productive discussion (and be better protected against false
starts and hucksters) if people like Friedman would take some time to
study up on the history of the subject before writing their next column.
* If you have any interest in the MOOC debate, you really ought to read this Wikipedia article on Distance Learning.
The deeply flawed premise through which elites have long operated is that trade is a net plus for everyone as long as the winners compensate the losers. But in the real world, the winners both fail to do so and use their winnings to buy tax and deregulatory policies that further screw the losers.
Insofar as this point is correct, there is a very short-sighted dynamic going on in modern politics. After all, trade has the potential to greatly improve human life and has been a key element to the development of civilization. Just look at the Silk Road.
When the benefits of trade are broadly spread then everyone benefits. But capturing all of the benefits and then using that political power to seize additional benefits is a great way to get very powerful but it runs the risk of undermining the political calculation. After all, if trade is the way that the "rich get richer" and the "poor get even less" then it starts to look like a very bad deal.
There are stable outcomes that lead to everyone being worse off, and we should guard against them. For trade, I think we need to think very carefully about how we distribute the benefits from trade throughout society.
[the next few of these come in response to this Matt Taibbi piece in Rolling Stone brought to our attention by Andrew Gelman.]
There is a lot of silliness to unpack here. Friedman has always been really bad on technology. A once solid reporter Peter-Principled into the role of "deep thinker" columnist, he spends his time cheerfully regurgitating flawed conventional wisdom on the subject. This is particularly notable with one of his favorite topics, MOOCs, but more on that later.
For now, though, I want to briefly address a side issue that is been bothering me for a long time, the way Friedman and other writers in the field have come to use the term "moonshot."
If we are talking about Apollo program as a template for projects to advance science and technology, it needs to mean "do something very big on a very aggressive schedule by spending huge amounts of money as fast as you can." Of course, there were other contributing factors, but if you want the bullet point explanation, that's it. For a relatively brief period, political and economic conditions lined up so that the LBJ administration was able to convince the country to let it spend somewhere close to 5% of the federal budget in direct and indirect funding for what was, in the short term, almost entirely a symbolic accomplishment. Even under ideal conditions that was a tough sell.
Did the program eventually pay for itself? Possibly many times over. Most scientific research provides a good return on investment. In terms of immediate payout however, we showed the world that we could beat the Russians to the moon, we produced a genuinely inspirational moment for the nation, and we provided work for as many engineers as the country could supply. That was about it.
The pour-as-much-money-as-possible-as-fast-as-possible-onto-the-problem model does not work equally well in all situations and it may not be the best approach overall, but it was the model for most of the achievements that the "visionaries" of today are so fond of invoking.
Today, "moonshot" has come to mean pick some cool-sounding futuristic project, hype it with a comically vague proposal, a few neat 3-D graphics, and the inevitable TED Talk, then proceed with some half-assed R&D, making sure to give overblown press conferences for anything that can be packaged as an advance. If it all possible, combine the story with a profile of a visionary Silicon Valley billionaire.
The purpose of today's "moonshots" is to make us all feel excited about living on the verge of a bright and wonderful future without actually having to do any of the work or make any of the sacrifices required to bring that future to fruition.
Aesthetically, I much prefer the first, but I think the second might do a better job conveying information. I have a feeling that I ought to be able to draw a connection between these and graphical representations of more traditional data. Maybe something will come to me over the weekend.
I've been making the point for a while now that the evangelical movement that I grew up with in the Bible Belt is radically different from the evangelical movement of today. I was aware that something was changing for a while, but the nature and the extent of the change crystallized for me when I readd this 2004 article from Slate:
Next Stop, Bethlehem?
By David Sarno
The Polar Express is the tale of a boy's dreamlike train ride to the North Pole to meet Santa Claus. Like all stories worth knowing, it's rich enough in image and feeling to accommodate many interpretations. Chris Van Allsburg, the author of the book, calls his story a celebration of childhood wonder and imagination. William Broyles Jr., one of the screenwriters of this year's film version, calls it a kind of Odyssey in which a hero undertakes a mythic, perilous journey of self-discovery. And Paul Lauer, who is a key player in the film's marketing apparatus, sees The Polar Express as a parable for the importance of faith in Jesus Christ.
Lauer's firm, Motive Entertainment, is best known for coordinating the faith-based marketing of The Passion of the Christ. Motive helped spread early word of mouth about the filmby holding screenings for church groups and talking the movie up to religious leaders. When The Passion took in a stunning $370 million at the box office, making it the highest-grossing R-rated film in history, Lauer and his cohorts got a lot of the credit. Earlier this year, Motive was hired by Warner Bros. to promote The Polar Express to Christians. But wait, is The Polar Express an evangelical film?
You'd certainly think so, considering the expansive campaign of preview screenings, radio promotion, DVDs, and online resources that Lauer unfurled in the Christian media this fall. This Polar Express downloads page includes endorsements from pastors and links to church and parenting resources hosted by the Christian media outlet HomeWord. There are suggestions for faith-building activities and a family Bible-study guide that notes, for example, the Boy's Christ-like struggle to get the Girl a train ticket. "The Boy risked it all to recover the ticket," the guide observes. "Jesus gave His all to save us from the penalty of our sins."
HomeWord Radio, which claims to reach more than a million Christian parents daily, broadcast three shows promoting the film. At one point, the show's host wondered excitedly if the movie "might turn out to be one of the more effective witnessing tools in modern times." Motive also produced a promotional package that was syndicated to over 100 radio stations in which Christian recording artists like Amy Grant, Steven Curtis Chapman, and Avalon talked about the movie as they exited preview screenings.
…
Some audience members—and a few Christian film critics—would argue that Santa Claus isn't necessarily a stand-in for Jesus Christ. Last month, Lauer told the Mobile Register that he sees The Polar Express as a parable, "not a movie about belief in God." But when Lauer speaks to a Christian audience, he tells a different story. Lauer told HomeWord Radio that when he asked Robert Zemeckis about all the biblical parallels he was seeing in the film, the director "winked and said, 'Nothing in a movie this big ends up in the script by accident.' " (Zemeckis was traveling and wasn't available for comment.)
This is a spectacular example of getting the pertinent details of the story right and yet completely missing the point. In another piece, the understatement of “Santa Claus isn't necessarily a stand-in for Jesus Christ” would be sharply comic but Sarno seems to be completely oblivious to the joke.
I know we overuse the clip of the minister gunning down Santa in the middle of a children's sermon, but it illustrates an important point.
Over the past few years the evangelical movement has abandoned the majority of its most deeply held theological beliefs (think of how doctrinal differences with Catholics and, even more notably, Mormons have been put aside). It is not at all coincidental the beliefs that were abandoned were uniformly inconvenient from a political standpoint. The conservative movement has both weaponized and secularized the evangelical movement with remarkable success.
Traditionally, evangelicals were more concerned with the potential corruption of their own religion (frequently to the point of paranoia) than with what others were practicing. Christmas was a particularly hot-button issue. In the eyes of several good Southern Baptist ministers, the holiday had become unacceptably commercial, cultural rather than religious, and, in many ways, pagan. Most of the music, imagery, and traditions had nothing to do with the nativity, the "reason for the season." Often, this general hostility toward secular Christmas celebrations focused on Santa Claus.
Like many religious practices, the no-Santa rule could look a bit silly when viewed from the outside, but there's nothing unreasonable about adherents of a particular faith wanting to maintain what they see as the original meaning of a religious holiday. Growing up, I found these attitudes and the little lectures that often accompanied them painfully annoying, but, even though I disagreed, I could see where they were coming from from a theological standpoint.
Now, evangelicalism is a religious movement stripped of its religious elements. There is no scriptural foundation for tax cuts for the rich, deregulating greenhouse gases, or Donald Trump, but those are the defining issue of the movement of today.
Of course, evangelicals are not monolithic. There are many within the movement, some in positions of authority, who object to these obvious deviations from their original core principles. There are indications that the resistance is gaining momentum, and it is entirely possible that in a few years we will have to rethink our assumptions about evangelical Christians and politics. For now, though, this is a cultural (social reactionary) and political (far right) movement, not a religious one, and trying to think of it in any terms that these is misguided.
I have extremely mixed emotions about "faithless elector" push. I see it as a distraction, it is very probably doomed to fail, and there is a "two wrongs make a right" feeling about the whole thing (just imagine how we would react if the situation were reversed).
My take on this (which has, if anything, been sharpened by recent events) is that we have been negligent about both maintaining the democratic process and protecting it from its opponents. I am more interested in fixing that process for the long-term than in changing the immediate outcome. As of next year, two of our last three presidents will have come into office losing the popular vote. Thanks to gerrymandering, the House has not represented the democratic will of the people for years. The Republicans decision to trash Constitutional norms and long-standing traditions has had a analogous affect on the judiciary. And none of this takes into account the blatant voter suppression efforts of the past few years.
At best, the faithless elector talk might serve as an effective protest against these anti-democratic efforts; at worst, it will simply drain the oxygen from the debate.
On a more personal note there are damned few people out there who are looking to me for normative statements. I've been doing a lot of reflecting as the election, thinking about the role of a statistics blogger in this whole mess. How do we make sure that we are adding value and not simply increasing the noise? One way is for each of us to ask him or herself what unique contributions we bring to the table. If we are all just trying to get our opinions on the record, we might as well all go over to Twitter.
Particularly for a statistics blogger, I believe personal experience and special knowledge can be extraordinarily useful. The combination of an analytic background and an informed perspective can go a long way toward bringing fresh insights and spotting potential problems in conventional arguments. There are few things more valuable for a statistician than having a good sense of what groups should or should not be aggregated and what relationships should or should not be treated as stable.
In my case, that experience includes growing up in the Bible Belt and spending pretty much all of my formative years arguing with fundamentalists. That has given me a strong feel for how evangelicals think. It has also made me more alert to the tremendous, in some cases cataclysmic, changes that have taken place in the movement over the past few years.
The current configuration of the evangelical movement is unstable. Just to be clear, I am not saying that we are about to see a radical shift toward either liberalism or toward the distrust of politics historically associated with groups like the Southern Baptists. When I am saying is that there are great tensions in the movement, that Trump has heightened those tensions, and that while we may not see huge changes in the way we think about religion and politics over the next few years, it would be foolish to rule out the possibility.
A Texas Republican announced over the weekend that he plans to resign his post as a member of the Electoral College rather than cast a ballot for US President-elect Donald Trump, a man he deems "not biblically qualified for office."
Citing his Christian religion and his understanding of representative democracy, Art Sisneros wrote in a blog post that he could neither vote for Mr. Trump nor break his promise to do so by voting for anyone else. Texas does not require its 38 electors to vote in accordance with the state's presidential election results, but Mr. Sisneros says he made a pledge to the Texas GOP that his vote next month in Austin would follow the will of the general public.
"The reality is Trump will be our President, no matter what my decision is," Sisneros wrote. "Since I can’t in good conscience vote for Donald Trump, and yet have sinfully made a pledge that I would, the best option I see at this time is to resign my position as an Elector."
Sisneros spells out his position in more detail here and here.
There are a lot of over the air television stations available in Los Angeles. The last time I helped someone set up an antenna, we were able to pull in 170. Now, most of these are probably channels that you would never watch. English is the language of the plurality but not the majority. Quite a few (though possibly less than you might expect) are shopping channels and their are a few duplicates, stations that appear to or three times on multiple points of the dial.
As the familiar disclaimer reads, individual user experiences may vary. The number and reliability of the channels you will receive is primarily dependent on your location, quality of your antenna, and whether or not you are in the position to put it up on the roof. In some places, you can pull in well over 100 channels with a $.99 pair of rabbit ears. For most parts of town, you'll probably need to spring for a $30 amplified indoor antenna (or hook your TV up to and unamplified outdoor antenna if it still up there after all these years). If you happen to be deep in a canyon or on the wrong side of a mountain (which happens quite a bit around here) you might want to spend the $50-$100 for a good external amplified model.
Personally, I have found you hit a point of diminishing returns somewhere in the middle, and I have never spent more than 35 bucks for a set up.
Even with the best of setups, however, there are still likely to be a few low-powered channels that you never seem to pick up. These LPT the stations tend to collect the more low-budget entries. As compared with the top end of the spectrum exemplified by Weigel Broadcasting and the major studio affiliated channels that prominently feature older but still A-list material like M*A*S*H or Columbo or films from notable directors such as John Ford, Billy Wilder, or Mel Brooks, the poverty row stations generally rely on public domain material, justly forgotten bottom-of-the-catalog shows (such as the Barkleys), and a great deal of Canadian television.
Weigel is the undisputed King of the top end of the spectrum; Luken Broadcasting is arguably the king of poverty row (sort of like a modern day Monogram Studios). With Weigel you get something like NYPD Blue. With Luken, it's more likely to be Police Surgeon (if you were a serious television buff, you'd have heard of this one, but not in a good way).
Luken's networks are filled with absolutely the cheapest programming possible, but they do deserve credit for some real innovation, particularly when it comes to new themes for terrestrial superstations. In addition to Retro TV (a poor man's MeTV), Luken offers a family channel, and outdoors channel, a country music channel, and a gearhead channel. I don't believe I can pick up any of these channels and, even if I could, I probably wouldn't spend much time watching them (though I will admit to a morbid curiosity about whether police surgeon is as bad as everyone says), but it is good that they are out there somewhere.
As we have mentioned many times before, 21st century media has serious monopoly and monopsony issues. Putting aside YouTube for the moment (though that to brings up monopoly and monopsony issues), both content providers and consumers have to rely on a tiny group of major media and telecommunication companies, companies that are both ethically challenged and badly run. Over-the-air television offers an invaluable alternative, a way for independent companies to get stations direct to consumers, and it gives consumers a low or no cost media option.
One of the factors that contributed to the punched-in-the-gut feeling that so many people had immediately after the election was the seeming orthogonality of the data.
I'm using orthogonal in the broad rather than technical sense here (though I suspect both might apply) meaning to bring new information into the model. It wasn't just that the poll aggregators (with the partial exception of the outlier 538) were all telling us that the outcome was almost certain; we were also hearing exactly the same thing from pretty much everyone else, sources which supposedly had access to different information and were using a variety of approaches. A partial list included prediction markets, expert analyses, pseudo-exit polls (Slate's ill-fated, badly-thought-out Votcastr), and (from what we can infer) the consensus opinions within the campaigns themselves. All of these converged on exactly the same, completely wrong conclusion.
It is likely to take a great deal of hard work and deep digging to uncover exactly what went wrong here, but we can make some educated guesses:
The other data sources were never all that orthogonal (and possibly never all that good). For instance, even under ideal circumstances the predictive power of the markets was always overstated and overhyped, and presidential elections are nowhere near ideal circumstances.
To make matters worse, whatever orthogonality these other sources once brought to the table had faded to nothing by the time we got to this election. Between their early successes and the ludicrous amount of attention they received, the poll aggregators' predictions increasingly dominated conventional wisdom and became the only input (direct or indirect) that mattered for all the other “independent” sources of information.
I suspect that we reached the point where (if you'll forgive a clumsy phrase) prediction markets and the rest were anti-orthogonal. By providing the illusion of independent confirmation of the flawed polling data and likely voter models, they actually made it more difficult to bring new information in the system. It is entirely possible that better informed (or at least less misinformed) voters might have acted very differently, which suggests that the consequences of this particular failure may have been high indeed.
If you watch this and you could swear you remember Johnny and Mr. Carlson discussing Pink Floyd, you're not imagining things. Hulu uses the DVD edit which cuts out almost all of the copyrighted music. .
As for my favorite line, it comes from the Buffy episode "Pangs" and it requires a bit of a set up (which is a pain because it makes it next to impossible to work into a conversation).
Buffy's luckless friend Xander had accidentally violated a native American grave yard and, in addition to freeing a vengeful spirit, was been cursed with all of the diseases Europeans brought to the Americas.
Spike: I just can't take all this mamby-pamby boo-hooing about the bloody Indians.
Willow: Uh, the preferred term is...
Spike: You won. All right? You came in and you killed them and you took their land. That's what conquering nations do. It's what Caesar did, and he's not goin' around saying, "I came, I conquered, I felt really bad about it." The history of the world is not people making friends. You had better weapons, and you massacred them. End of story.
Buffy: Well, I think the Spaniards actually did a lot of - Not that I don't like Spaniards.
Spike: Listen to you. How you gonna fight anyone with that attitude?
Willow: We don't wanna fight anyone.
Buffy: I just wanna have Thanksgiving.
Spike: Heh heh. Yeah... Good luck.
Willow: Well, if we could talk to him...
Spike: You exterminated his race. What could you possibly say that would make him feel better? It's kill or be killed here. Take your bloody pick.
Xander: Maybe it's the syphilis talking, but, some of that made sense.
Wednesday, November 23, 2016
This is Joseph
I hadn't been thinking about density like this, but here is a good point about LA:
LA suffers from the too-dense-but-not-quite-dense-enough problem (overall it's a very dense city, but with a kind of uniform density that is a bit difficult from a transportation perspective).
I have generally hold that density is a pure good for transit, but it seems obvious that "destinations" would make things work a lot better than thousands of point to point connections. The former (New York density) seems to make transit look super efficient. But if you are relatively dense everywhere, that almost makes a cab/uber style of transit look efficient.
An anonymous elector told Politico that the House election that would result from a Trump Electoral College defeat “would immediately blow up into a political firestorm in the U.S." and be a positive step toward galvanizing the public’s support for ditching the college. Another, former Democratic National Committee Vice Chairwoman Polly Baca, said she’d prefer that the Electoral College return to the model outlined by Alexander Hamilton in the Federalist Papers and become an independent body informed but not governed by actual vote tallies.
I am especially surprised by the second suggestion here: just how does this help the problem? You still have a potential difference between the popular vote and the electoral college. But now the college can go completely rogue and decide the election. How do we know that this version of the college, with what sound like unbound electors, will improve transparency. The problem that people seem to worry about is the discordance between the electoral college and the popular vote -- how does this help?
As for the first, the electoral college is currently in the constitution of the United States. Last time I looked, it would take staggering levels of public support to pass a constitutional amendment. Especially since any such amendment would obviously disadvantage whichever major party draws lots of support from rural and low population states. I like the general idea of reforming this system, but I am unclear if continuing to destroy the norms of governance is going to be a good plan.
This applies to a wide range of contexts, but, just to get the conversation started, let's say you're doing targeted marketing. You can come up with a fantastic mailing model, one that improves on the previous one in every conceivable way – – better response even if you mail somewhat deeper, more stable, and using data that are more reliable, cheaper, easier to work with – – and yet you will still get, at best, a lukewarm response from the executives. Invariably, you will be told something like this, "that's nice but we got so much more lift from the first model."
The trouble is that the improvement you see going from model to better model is almost always underwhelming compared to the improvement you see going from nothing to model. This curse can badly distort reputations and often leads to a kind of super Peter Principle where people are promoted to a level one step higher than their perceived level of competence which is much higher than their actual level of competence.
