Siddhartha Mukherjee on a common post hoc abuse of statistical analysis

Siddhartha Mukherjee:

then a second instinct takes over: Why not try to find the people for whom the drug did work?…

But it’s also a treacherous seduction…

Perhaps the most stinging reminder of these pitfalls comes from a timeless paper published by the statistician Richard Peto. In 1988, Peto and colleagues had finished an enormous randomized trial on 17,000 patients that proved the benefit of aspirin after a heart attack. The Lancet agreed to publish the data, but with a catch: The editors wanted to determine which patients had benefited the most. Older or younger subjects? Men or women?

Peto, a statistical rigorist, refused — such analyses would inevitably lead to artifactual conclusions — but the editors persisted, declining to advance the paper otherwise. Peto sent the paper back, but with a prank buried inside. The clinical subgroups were there, as requested — but he had inserted an additional one: “The patients were subdivided into 12 … groups according to their medieval astrological birth signs.” When the tongue-in-cheek zodiac subgroups were analyzed, Geminis and Libras were found to have no benefit from aspirin, but the drug “produced halving of risk if you were born under Capricorn.” Peto now insisted that the “astrological subgroups” also be included in the paper — in part to serve as a moral lesson for posterity. I’ve often thought of Peto’s paper as required reading for every medical student.

 

Jonathan Haidt on the decline of liberal arts education

Jonathan Haidt:

Today’s identity politics . . . teaches the exact opposite of what we think a liberal arts education should be. When I was at Yale in the 1980s, I was given so many tools for understanding the world. By the time I graduated, I could think about things as a utilitarian or as a Kantian, as a Freudian or a behaviorist, as a computer scientist or as a humanist. I was given many lenses to apply to any given question or problem.

But what do we do now? Many students are given just one lens—power. Here’s your lens, kid. Look at everything through this lens. Everything is about power. Every situation is analyzed in terms of the bad people acting to preserve their power and privilege over the good people. This is not an education. This is induction into a cult. It’s a fundamentalist religion. It’s a paranoid worldview that separates people from each other and sends them down the road to alienation, anxiety and intellectual impotence. . . .

Is moving to free college good for the health of higher education?

There is much talk about moving to free college in the US. What impact would that have? Some evidence of possible consequences is provided in The End of Free College in England: Implications for Quality, Enrolments, and Equity by Richard Murphy, Judith Scott-Clayton, and Gillian Wyness:

This paper examines the consequences of charging tuition fees on university quality, enrolments, and equity. To do so, we study the English higher education system which has, in just two decades, moved from a free college system to one in which tuition fees are among the highest in the world. Our findings suggest that England’s shift has resulted in increased funding per head, rising enrolments, and a narrowing of the participation gap between advantaged and disadvantaged students. In contrast to other systems with high tuition fees, the English system is distinct in that its income-contingent loan system keeps university free at the point of entry, and provides students with comparatively generous assistance for living expenses. We conclude that tuition fees, at least in the English case supported their goals of increasing quality, quantity, and equity in higher education.

WSJ reports that Americans losing faith in the value of college

Josh Mitchell and Douglas Belkin of the Wall Street Journal report that Americans are losing faith in the value of college, as represented in these survey results:
The most shocking results to me are that 30% of college degree holders do not think a college degree is worth it, as do over half of men and over half of 18-34 year olds.

Andrew Gelman and Hal Varian on Instrumental Variables

Andrew Gelmen

The trick: how to think about IV’s without getting too confused

Suppose z is your instrument, T is your treatment, and y is your outcome. So the causal model is z -> T -> y. The trick is to think of (T,y) as a joint outcome and to think of the effect of z on each. For example, an increase of 1 in z is associated with an increase of 0.8 in T and an increase of 10 in y. The usual “instrumental variables” summary is to just say the estimated effect of T on y is 10/0.8=12.5, but I’d rather just keep it separate and report the effects on T and y separately…

If there’s any problem with the simple correlation, I see the same problems with the more elaborate analysis–the pair of correlations which is given the label “instrumental variables analysis.” I’m not opposed to instrumental variables in general, but when I get stuck, I find it extremely helpful to go back and see what I’ve learned from separately thinking about the correlation of z with T, and the correlation of z with y. Since that’s ultimately what instrumental variables analysis is doing.

Hal Varian adds

You have to assume that the only way that z affects Y is through the treatment, T. So the IV model is
T = az + e
y = bT + d
It follows that
E(y|z) = b E(T|z) + E(d|z)
Now if we
1) assume E(d|z) = 0
2) verify that E(T|z) != 0
we can solve for b by division. Of course, assumption 1 is untestable.

An extreme case is a purely randomized experiment, where e=0 and z is a coin flip.

Scott Alexander on what passes for socialism these days

Scott Alexander

Most of the policies being mooted by the supposedly socialist left today – Medicare-for-all, better social safety nets, et cetera – are well within the bounds of neoliberalism – ie private property and capitalist economies should exist, but the state should help poor people. “Socialism” should be reserved for systems that end private property and nationalize practically everything. I’m worried that people will use the success of neoliberal systems in eg Sweden to justify socialism, and then, socialism having been justified, promote actual-dictionary-definition socialism. To a first approximation, Sweden is an example of capitalists proving socialism isn’t necessary; Venezuela is an example of socialism actually happening.

Louisa Lim on elite turnover 68 years after the Chinese revolution

Louisa Lim

Research by the University of Sydney’s David Goodman has found that around 84% of today’s elite are direct descendants of the elite from pre-1949. This suggests that six decades of Communism may not have a dramatic impact upon the elites, who have the advantage of decades of capital accumulation — including economic, cultural and social capital — which have apparently continued to benefit them under the party-state system…

Matt Ridley on socialism  

Matt Ridley

Remember that we have run two very careful randomised controlled trials to see if full-blown socialism or half-hearted free enterprise works better. One in the Korean peninsula, the other in Germany. And the results were unambiguous. Socialism was a humanitarian catastrophe.

Communism was not really a new or radical idea, even in 1917. It was simply a clever repackaging of the old, old story that the king knows best. That the state should decide how to plan and run society. It matters not whether his name is Rameses or Augustus or Suleiman or Henry or Napoleon or Adolf or Vladimir or Josef or Mao or Fidel or Kim or Hugo. It’s the same recipe…

Is it not bizarre, after the 20th century, that people are so forgiving of the state and so mistrustful of the market?…