Offbeat Humor

Click to enlarge.

Data Talks

This map overlays the approximate boundaries of African ethnic groups on top of the modern national boundaries of Africa. #maps #mapping #Africa pic.twitter.com/EjkxXJ53Ti

— Bryan Druzin (@BryanDruzin) February 13, 2023

 

An Alchemist's Laboratory, 1570 by Jan van der Straet (Dutch, 1523-1605)

An Alchemist's Laboratory, 1570 by Jan van der Straet (Dutch, 1523-1605)

Click to enlarge.

History

Bring back competitive dining clubs. Bring back marauding gourmands. pic.twitter.com/k7DbXmhc8i

— Bachman (@ElonBachman) February 20, 2023

 

Observational studies must address all four quadrants of the Rumsfeld matrix

I not infrequently disagree with Emily Oster, usually not because I believe she is necessarily wrong but because she often seems to lack confidence in her empirical approach.  I endorse the scientific method, while recognizing that it is more challenging than is often acknowledged.  We should be bold in our confidence in the scientific method even while acknowledging that the world is complex and does not easily let go of its secrets and mysteries.  

Through the Covid pandemic, there were many public health issues where the federal approach was either explicitly wrong based on past experience or at best on dubious grounds.  I would have wished Oster to have been more forthright and earlier in her criticism because the science warranted the criticism.  But she was generally mooted and conformed.  

So it is nice to find a piece by her with which I whole-heartedly endorse.  In this instance, the care that needs to be taken when concluding anything from observational studies.  They can be dispositive, clarifying and suggestive.  They are rarely sufficient to reach strong conclusions.

A question I get frequently: Why does my analysis often disagree with groups like the American Academy of Pediatrics or other national bodies, or other public health experts, or Andrew Huberman (lately I get that last one a lot)? The particular context is often in observational studies of topics in nutrition or development.

[snip]

The questioner essentially notes: the reason we know that the processed food groups differ a lot is that the authors can see the characteristics of individuals. But because they see these characteristics, they can adjust for them (using statistical tools). While it’s true that education levels are higher among those who eat less processed food, by adjusting for education we can come closer to comparing people with the same education level who eat different kinds of food.

However, in typical data you cannot observe and adjust for all differences. You do not see everything about people. Sometimes this is simply because our variables are rough: we see whether someone has a family income above or below the poverty line, but not any more details, and those details are important. There are also characteristics we almost never capture in data, like How much do you like exercise? or How healthy are your partner’s behaviors? or even Where is the closest farmers’ market? 

For both of these reasons, in nearly all examples, we worry about residual confounding. That’s the concern that there are still other important differences across groups that might drive the results. Most papers list this possibility in their “limitations” section. 

We all agree that this is a concern. Where we differ is in how much of a limitation we believe it to be. In my view, in these contexts (and in many others), residual confounding is so significant a factor that it is hopeless to try to learn causality from this type of observational data. 

Indeed.  And this is not inconsequential.  Whenever there is a disparate impact study, it is intended by design to control for confounding variables so that any variance in outcomes can be attributed to discrimination.  For decades it was both a matter of ideological faith that women were discriminated against in the market economy because they were women.  Infamously known as the Gender Pay Gap.

And for a decade or two now, we have known that the Gender Pay Gap is entirely an artifact of confounding variables.  Once you control for education achievement, hours worked, number of absences and duration of absences from the marketplace, field of endeavor, etc., there is no gap.  As economic theory would suggest, people, regardless of sex, are paid the same for the same type of work with the same experience.  

Think about the tens of thousands of hours of legislative and policy debate and the thousands of laws and regulations which have been passed or enacted to banish a problem that doesn't actually exist.  There is no Gender Pay Gap except in badly or inadequately designed observational studies in the context of an ideological conviction that any disparate impact must be attributable to conscious or unconscious discrimination.

What a waste of time.

Oster continues:

Conceptually, the gold standard for causality is a randomized controlled trial. In the canonical version of such a trial, researchers randomly allocate half of their participants to treatment and half to control. They then follow them over time and compare outcomes. The key is that because you randomly choose who is in the treatment group, you expect them, on average, to be the same as the control other than the presence of the treatment. So you can get a causal effect of treatment by comparing the groups.

Randomized trials are great but not always possible. A lot of what is done in public health and economics aims to estimate causal effects without randomized trials. The key to doing this is to isolate a source of randomness in some treatment, even if that randomization is not explicit.

[snip]

We can take this lens to the kind of observational data that we often consider. Let’s return to the processed food and cancer example. The approach in that paper was to compare people who ate a lot of processed food with those who ate less. Clearly, in raw terms, this would be unacceptable because there are huge differences across those groups. The authors argue, though, that once they control for those differences, they have mostly addressed this issue.

This argument comes down to: once I control for the variables I see, the choice about processed food is effectively random, or at least unrelated to other aspects of health.

I find this fundamentally unpalatable. Take two people who have the same level of income, the same education, and the same preexisting conditions, and one of them eats a lot of processed food and the other eats a lot of whole grains and fresh vegetables. I contend that those people are still different. That their choice of food isn’t effectively random — it’s related to other things about them, things we cannot see. Adding more and more controls doesn’t necessarily make this problem better. You’re isolating smaller and smaller groups, but still you have to ask why people are making different food choices.

Food is a huge part of our lives, and our choices about it are not especially random. Sure, it may be random whether I have a sandwich or a salad for lunch today, but whether I’m eating a bag of Cheetos or a tomato and avocado on whole-grain toast — that is simply not random and not unrelated to other health choices.

This is where, perhaps, I conceptually differ from others. I have to imagine that researchers doing this work do not hold this view. It must be that they think that once we adjust for the observed controls, the differences across people are random, or at least are unrelated to other elements of their health.    

She provides some good detail on illustrating examples.  She gets to the core issues:

The control sets we typically consider are incomplete. There are a lot of papers that report effectively only the first two bars in the graph above. But those simple observable controls are just not sufficient. The residual confounding is real and it is significant. 

It is all well and good to control for the known known confounding variables.  But we still have the unknown known, the unknown unknown, and the unknown unknown confounding variables in the Rumsfeld Matrix.

Oster is effectively pointing out that virtually all observational studies only at best address one of the four Rumsfeldian quadrants of epistemic uncertainty, the known knowns.  All the rest are terra incognita, rendering  observational studies of little usefulness.  

The question of whether a controlled effect in observational data is “causal” is inherently unanswerable. We are worried about differences between people that we cannot observe in the data. We can’t see them, so we must speculate about whether they are there. Based on a couple of decades of working intensely on these questions in both my research and my popular writing, I think they are almost always there. I think they are almost always important, and that a huge share of the correlations we see in observational data are not close to causal.

An Insight

For virtually all of human history, almost everyone alive (95%+) was subsistence farmer-level poor. When I was born, that had fallen to about half the world’s population. Today, we’re down to ~10%.

This “everything sucks now” worldview comes from a place of incredible privilege. https://t.co/K5DdSwdIXN pic.twitter.com/b1qGMSAkaE

— John Mozena (@johnmoz) February 21, 2023

 

Honesty and truth telling are too high a price to defeat conspiracy theorising

Over the past six years, there has been a mainstream media or chattering class tendency to deploy the accusation of conspiracy thinking as a weapon to be brandished against Republicans and conservatives.  From an epistemic point of view, that is interesting given how many conspiracies have ended up being proven.  

But there is a real underlying question.  Is conspiracy thinking more prevalent on the left or right.  My working hypothesis is that inclination towards conspiracy explanations is not so much a partisan issue as it is a balance of power issue.  The more secure a party feels, the less inclined to indulge conspiracy thinking.

You can predict that one party or the other might have greater sympathy to conspiracy thinking than the other, not because of their underlying philosophy but based on whether their position is strengthening or weakening.  Weakening parties are more  prone to conspiracy thinking.  

Two pieces from this morning that tie into the debate. The first is Americans face a rapidly encroaching 'emergency' CBDC power grab by Jordan Schachtel.  The subheading is The ruling class may pursue a Hail Mary pass to restore their control over the system.

The American financial system is threatening to come apart at the seams, and for the people who control the levers of power, the only way to patch things up may involve the installation of a monetary Social Credit Score system. In recent years, America’s fiat fractional reserve system has transformed into a faith-based credit system, and the people who use the dollar are losing confidence in a system that relies entirely upon their complete and total trust. Should our collective faith in the system continue to decline, the American ruling class will decide that their path forward involves regrasping full control of their confidence scheme through the implementation of a Central Bank Digital Currency (CBDC).

A U.S. CBDC would do much more than simply implement a fully digital version of the U.S. dollar. This system could provide authorities with an almost unlimited digital toolkit to both surveil and censor citizens. A CBDC is advertised as making the system more “efficient” and helping to deliver monetary power to the unbanked. However, it would also give shadowy bureaucrats the power to swipe a “criminal’s” life savings, instantly distribute funds to allies of the system, among an almost infinite series of additional authoritarian instruments.

I have never heard of Schachtel before.  I would not have read the article except for this post.  It sounds very conspiratorial. 

However, the experience with Canada and Covid demonstrates that governments are perfectly comfortable with abridging human rights when it suits them.  Protest the lockdowns and your bank account gets shut.  

In addition, when something feared does actually happens, it becomes more difficult to dismiss as a loony conspiracy.  When factions within the regulatory agencies begin to illegally coordinate with private sector banks in order to abridge civil rights laws (such as the second amendment right to bear arms), then the wilder positions taken on a sheen of legitimacy.

The second piece I note this morning is also supportive of the legitimacy of conspiratorial thinking.  From It's OK if the King does it by Yassine Meskhout.  

Remember Seattle's CHAZ/CHOP? After the place was cleared, a bunch of local businesses and property owners sued the city and recently all reached a settlement. One part that definitely didn't help Seattle were tens of thousands of deleted text messages:

The city of Seattle has settled a lawsuit that took aim at officials’ handling of the three-week Capitol Hill Organized Protests and further ensnared the former mayor and police chief, among others, in a scandal over thousands of deleted text messages. The Seattle City Attorney’s Office filed notice of a settlement Wednesday in U.S. District Court, just three weeks after a federal judge levied severe legal sanctions against the city for deleting texts between high-ranking officials during the protests and zone that sprung up around them, known as CHOP.

[...]

Attorneys for the more than a dozen businesses that sued the city, led by Seattle developer Hunters Capital, sent a series of letters to the city in July 2020 — after another lawsuit over the violent police response to the protests — demanding that any evidence pertaining to the city’s alleged support and encouragement of the zone’s creation be retained, according to the court docket and pleadings.

U.S. District Judge Thomas Zilly concluded last month that officials ignored the notifications, sending the so-called Hunters Capital lawsuit to trial on two of five claims and dismissing three others. In doing so, Zilly issued a blistering order that leveled crippling sanctions against the city for the deletion of tens of thousands of text messages from city phones sent between former Mayor Jenny Durkan, former police Chief Carmen Best, fire Chief Harold Scoggins and four other ranking city officials during the protests.

The judge found significant evidence that the destruction of CHOP evidence was intentional and that officials tried for months to hide the text deletions from opposing attorneys.

Meskhout goes on chapter and verse where government entities basically abuse citizen rights and then escape scot free.  Did Seattle city government collude with Antifa against the interests of citizens and the Cities obligations?  That sure sounds like a conspiratorial position but apparently it was also the reality.  The evidence being the effort of City government to hide the evidence.  

Likewise with the current case over Harvard discrimination against Whites and Asians in which a judge has sought to hide the joking communication between the federal regulator and the admissions officers where they are joking about the absurdity of just how talented are Asian applicants.

We are all trying to navigate what is real, or at least usefully true, information.  And we keep finding that government, the academy, and the mainstream media are perfectly comfortable trying to censor or suppress information which does not conform with their interests.  If trust were greater, and more warranted, we probably would not have so many conspiracy theories.  

But when bad actions are increasingly prevalent and the propaganda against truth more persistent, the harder it is for responsible citizens to discern reality and the more prone everyone becomes to entertaining conspiratorial thinking.  Not because they are conspiracy fanatics but because epistemically it is warranted.

Finally, there is a good discussion at This Just In: Conspiracy Theorists Not Quite as Kooky as Previously Reported by Jesse Walker.  The subheading is Greetings from the second International Conspiracy Theory Symposium, where one of the most cited findings in the field has been debunked.

If you believe that Princess Diana was assassinated, you almost certainly do not also believe that she is secretly still alive.

That may sound obvious, but there are parts of the academy where it flies in the face of conventional wisdom. In 2012, a much-cited paper in the journal Social Psychological and Personality Science seemed to show that people willing to reject the official story of Di's death—that she had been killed in a car accident—weren't very choosy about which alternative they embraced: "the more participants believed that Princess Diana faked her own death, the more they believed that she was murdered." 

[snip]

The press couldn't resist the idea of a kook so divorced from common sense that he thinks someone could be both alive and dead. The study became a staple of pop-science pieces on conspiracy theories, and of pop-intellectual writing by figures such as Cass Sunstein. And when other experimenters followed up on the paper, they replicated its results.

"Journalists love it," declared Jan-Willem van Prooijen, a psychologist from VU Amsterdam, as he addressed the International Conspiracy Theory Symposium at the University of Miami this past weekend. "It's a cool finding. There's just one problem: It's not true."

Van Prooijen is not the first scholar to challenge this idea. Last year, for example, the philosopher Kurtis Hagen noted that the original study did not measure people's beliefs so much as the degree of credence they gave to different possibilities: Rather than simply endorsing or rejecting each theory, participants were asked to rate each story's plausibility on a seven-point scale, an approach that gave room to entertain the ideas as suspicions without embracing them as full-fledged beliefs. But van Prooijen was discussing a more fundamental problem. The whole phenomenon, he told the Miami audience, could just be a statistical artifact.

Most people, after all, don't believe that Diana was assassinated or that she faked her death. If you're just looking at the overall numbers, that huge correlation between the participants who disbelieve both stories could create the illusion of a correlation where participants believe both. So van Prooijen and four colleagues ran their own series of experiments, this time paying closer attention to who was endorsing and rejecting each yarn.

The results, which will soon appear in the journal Psychological Science, showed that people who endorsed one conspiracy story were generally less likely, not more likely, to endorse an apparently contradictory narrative. There were a few exceptions, but these involved questions where, on closer examination, the theories weren't necessarily contradictory after all. For example: After the first experiment showed people maintaining that pharmaceutical companies were both obstructing research to find a cancer cure and withholding a cure they already possessed, the authors realized that these could be reconciled if you believe Big Pharma is hiding a cure for one type of cancer and blocking research on another. Whatever else you might think of that belief system, it is not as irrational as the Schrödinger's Princess scenario.

Interesting throughout.  I would add that there is a framing issue.

There is some controversial issue and I think the most likely explanation might be X.  That belief that X is the most probable explanation might contradict an equally held belief in Y on some other issue.  

This overlooks that I may not be particularly wed to explanation X.  It may be the most likely explanation but still not be a likely explanation.  

For example, I might consider four explanations for the current inflationary and banking crises.  1) They are happening because President Biden is no longer mentally competent and the administration is drifting; 2) The crises are due to rank incompetence of Yellen; 3) The crises are the basis for Democrats to seize more executive power and control over the populace; and 4) It is just a coincidence that these crises are happening at the same time.  

I don't think any one of these is a viable explanation.  I suspect that there are elements of each in play at the same time as well as additional causal factors.  If I were asked to rank each by probability of being true, it might look like 1) 25%, 2) 30%, 3) 15%, and 4) 10%.  I don't think these are good explanations but if I were to choose a single cause I would go with Yellen's incompetence caused the problem.  Even though I think there is only a 30% chance of that explanation being true.  

Am I conspiracy thinker?  I would argue not.  I look for facts and where we are dealing with a complex, evolving, and chaotic system, I recognize that there might not be a comprehendible causal explanation.  All there might be is a range of causal scenarios which have some probability of being an accurate explanation and all of which have some possibility of being true.  

I see wonderful things

Slow motion of an enormous lightning strike!
That is Zeus?

pic.twitter.com/jot13DbD0W

— The Figen (@TheFigen_) February 20, 2023

 

Offbeat Humor

Click to enlarge.

Data Talks

Studies and meta-analyses typically find that men have more general knowledge than women.

A variation on this is the male proclivity for ranking and listing things by subject or classification — a pleasure that seems to elude most women. pic.twitter.com/AwkE6i9kWz

— Monitoring Bias (@monitoringbias) February 14, 2023