I have a problem with meta analysis. Slight untruth: I have many problems with meta analysis: I have done too many of them and the amount of subjectivity in what is supposed to be a systematic review and careful mathematical summation of pooled data is always more messy than it appears.

The problem I am talking about is effect size. This is defined as the difference in mean change divided by the pooled standard deviation. This gives an index: below 0.20 is usually seen as small, 0.4 as moderate and over 0.6 or 0.8 as large. This index has some utility for power calculations: if you have a scale and you think a clinical difference of X is appropriate you can estimate the pooled standard deviation from the literature (including meta-analysis) and estimate the number of participants you need for a trial.

But in meta analysis I lose something. I lose the measurement: all measurements have strengths and weaknesses. Many predate modern syndromes. Most measurements are proxy, and a series of measurements for the same syndrome ask different questions, have different predictive power, and different factor structures and internal consistency.

An effect size obscures all this. In medicine, this is known, and the psychologists and statisticians who want everything as effect size are held in suspicion, for we afraid that someone is sneaking a gorilla into a herd of goats.

In economics this is worse. The measurements are not merely proxy (how do you measure inequality? Life is more than money) but the assumptions behind the measurements are flawed, as Pendleton noted:

The Gini coefficient assigns a rating to a nation, between 0 and 1, where the lower the rating the more “equally” the nation’s wealth is distributed, while the higher the rating the more “unequally” the nation’s wealth is distributed. Keep in mind that logical premises can be played with to cook up all sorts of nonsensical conclusions. This is where the new church’s sleight of hand is at play. Consider that on the Gini scale, the 0 point, the most equal distribution of wealth, represents a nation where all of its citizens receive the exact same income. Let that premise simmer in your skull awhile.

The goddess Equality has beckoned to Corrado Gini from the far shore: all things must be equal in all ways. All fields must be leveled, all buildings razed, and all income curves must be flattened. The goddess spoke and Gini listened.

It should not take anyone a long time to figure out the flaw in Gini’s most fundamental assumption. Gullible eggheads and academics may be fooled but the rest of us have more sense between our ears. Equality of opportunity is not the same as equality of outcome. A society in which everyone receives equal pay would be the most unequal society imaginable.

Do we dare imagine a nation where an open-heart surgeon takes home the same pay as a janitor? One job requires nearly a decade of dedicated study, an unwavering hand, a stout set of nerves, and the ability to think quickly lest a patient die; the other job requires a broom. What may be termed equality of outcome would certainly not be equality to the surgeon who spent his youth learning the scalpel when he could have just taken up a broom and gotten paid the same wage. Now multiply the scale of this nightmare by several orders of magnitude: a pharmaceutical chemist taking home the wages of a street sweeper; an electrical engineer receiving the same pay as a prostitute. Nothing could be more unequal. Nothing could be more unfair.

Yet this is what Gini asks us to accept as his central premise. He asks us to accept the madness that “equality” means unequal work should make for equal wages. He asks us to embrace an idea of equality that benefits the most careless sweepers and penalizes the most capable surgeons. Not one shred of “equality” can be found in the Gini coefficient, and yet, if we were to research the mathematical formula in university textbooks; if we were to consult with professional economic analysts and think-tanks we would see a world that embraces the Gini coefficient as one of the many perfectly legitimate methods of evaluating a nation.

A nation where everyone has exactly the same income is a perfect communist state. In such, power is not related to income, but access: to (very rationed) goods, to power, and to the levers of terror.

Inequality of outcomes makes for a richer nation: it allows for charity, the sponsorship of the arts, and even truth and beauty.

A fully equal society is a fully converged society. It is the economy of hell, and many progressives seek such a fate. Do not be them.