The problem with statistics

Our society relies extensively on statistics, which are an averaged measurement of effects over multiple instances.

However, there’s a problem with these: they suggest a truth, linguistically, that they do not actually indicate.

Those who specialize in statistics tend to be wary of their use except as a statistical assessment; they’re careful to point out that statistical data doesn’t decide an issue. Yet many of the sources of information we rely in on this busy modern society, especially those from the mass media, source their argument with statistics.

Here are some problems with statistics:

  • Cause/effect reversal (from Nietzsche): The old man thinks he has had a long life because he eats nothing but celery. Reality: it’s genetic, and he would have lived just as long eating steak.
  • Cause/correlation: Many people with green eyes die of heart disease. Therefore, the implied argument goes, green eyes lead to hart attacks. Reality: people with green eyes are clustered in an area of the country where people eat more fats, are of Irish heritage and spend many months indoors during the winter — any of these could contribute.
  • Single factor: Many of the people who eat lots of eggs have high cholesterol. Therefore, eggs cause high cholesterol. Reality: many of the people who eat lots of eggs also eat lots of bacon, causing high cholesterol.

As you can see, statistical correlation does not mean proof of cause.

Another illusion: that you can statistically compensate for the above, which requires knowledge of the causes of outcomes which, in any sample size of statistical importance, is impossible.

Using statistics, I can prove that brushing your teeth causes AIDS and being Caucasian makes you a pygmy anal rapist. Does it mean what is seems to mean, if you read the language as it falls on the page?

Here’s a great example:

In their book The Spirit Level: Why More Equal Societies Almost Always Do Better, Richard Wilkinson and Kate Pickett present a graph showing a 30-year widening gap between the incomes of the richest and the poorest. Using a brilliant compilation of statistics, they argue that inequality is the root cause of much unhappiness in our society.

The Guardian

More detail here:

Their book charts the level of health and social problems — as many as they could find reliable figures for — against the level of income inequality in 20 of the world’s richest nations, and in each of the 50 United States. They allocate a brief chapter to each problem, supplying graphs that display the evidence starkly and unarguably. What they find is that, in states and countries where there is a big gap between the incomes of rich and poor, mental illness, drug and alcohol abuse, obesity and teenage pregnancy are more common, the homicide rate is higher, life expectancy is shorter, and children’s educational performance and literacy scores are worse. The Scandinavian countries and Japan consistently come at the positive end of this spectrum. They have the smallest differences between higher and lower incomes, and the best record of psycho-social health. The countries with the widest gulf between rich and poor, and the highest incidence of most health and social problems, are Britain, America and Portugal.

The Times

Let’s assume they have not faked any of their data — a big assumption with these books that get debunked five years after their release, which is not soon enough to prevent their authors from carrying off a pile of loot and prestige.

Societies with more inequality have more problems. Sounds like simple cause/effect logic, doesn’t it? We mention more inequality, and then what sounds like the result: more problems. And then there’s some handwaving about statistically accounting for other factors and eliminating other sources.

But the one thing they didn’t address was that there are in fact also causes of inequality, and those causes could be the cause of the social problems mentioned. They have shown correlation; have they shown cause? They have shown a single factor; have they accounted for all?

No, they haven’t, because that wouldn’t sell books. We’d all love to hear that life not treating us fairly is the reason for all our problems; that way, we’re off the hook. If we’re crafty, we claim that life treats no one fairly, and that we all need to be treated with more equality; that gets us treated equally, and puts the weight of all the other people in society behind us. If one is mistreated, in theory, the rest rush to his or her aid. So with that solution we get to have our cake and eat it too.

Modern people are so stupefied by correlation arguments that when confronted with them, they panic and spout gibberish.

[I]nequality comes from people being in unequal situations.

See how this man’s brain has been cut off at the stem? The world is not considered as a factor. Only human motivations — we’re all equal, and the only difference is our situation. There are no factors that existed before us, no biological factors that limit us, and no real-world factors that restrain us. It’s just ol’ Inequality there (substitute “Satan” if you wish) that holds some of us back and gifts others — without them having worked for it, or even better being smart enough that they don’t need to slave away to get better results.

Isn’t that wishful thinking in a nutshell? It’s that mentality that the authors pander to, and instead of using a scientific proof, they create an implied statistical correlation and leave it at that. “Oh, that’s the answer! How easy!”

Yet making society more equal, throughout history, has never made the society what it needs to be, which is more organized, more creative, more intelligent and more disciplined — factors which lead to both greater wealth and greater social stability. (We call those “rising” cultures.)

Here’s another statement of the same idea:

Growing inequality in US cities could lead to widespread social unrest and increased mortality, says a new United Nations report on the urban environment.

In a survey of 120 major cities, New York was found to be the ninth most unequal in the world and Atlanta, New Orleans, Washington, and Miami had similar inequality levels to those of Nairobi, Kenya Abidjan and Ivory Coast. Many were above an internationally recognised acceptable “alert” line used to warn governments.

“High levels of inequality can lead to negative social, economic and political consequences that have a destabilising effect on societies,” said the report. “[They] create social and political fractures that can develop into social unrest and insecurity.”

The Guardian

Inequality is a red herring. Statistically speaking, unequal places have more problems. But what if there is a cause of inequality? Supposing there are two options:

  1. Inequality is caused by society.
  2. Inequality is caused by different levels of ability.

If it’s the former case, and we accept that belief as religion, then the UN makes sense: people who are capable are being oppressed. Yet that obviously makes no sense, because the amount of effort required to suppress capable people causes nations to collapse.

Instead, it’s more likely the answer is “B” — that we have differing abilities, and so differing fortunes, and that emerges as “inequality.” But equality isn’t a fix, because the people are unequal. So campaigning for inequality only produces more unrest, and more incapable people, who eventually create social unrest and then the society collapses. Heck, it’s how most societies have collapsed through history.

But their statistical view doesn’t look this deeply. It only checks off factors: inequality yes, a big fat mess, yes also. Then, in some nitwit sense of logic, it assumes there’s an arrow between the former and the latter. “Well this must be the cause! Great Scott!”

And then some evidence for column B:

American political views aren’t so binary, yet the happiness divide seems to be real. Previous studies, including a 2006 survey from Pew Research Center have found the same general trend, much to the delight of conservative pundits like George Will, who noted that “liberalism is a complicated and exacting, not to say grim and scolding, creed.”

The authors of the Pew study suggested income, religion and ideology all played a role in shaping the happiness divide.

To add some ammo to these explanations, Napier and Tost conducted a series of surveys on political attitudes of Americans and citizens of 8 Western countries, using previously collected data. Their results affirmed the “conservatives are happy, liberals are mad” findings of previous polls, but income, education, religion and other demographic variables couldn’t explain the happiness gap.

However, when the authors instead grouped people by their “rationalisation of inequality,” the differences between conservatives and liberals dissolved. Republican or Democrat, people not bothered by social or economic disparities tend to be happy.

This trend held for non-Americans, as well. Right-wingers in the Czech Republic, Germany, New Zealand, Norway, Slovakia, Spain, Sweden and Switzerland were all happier than liberals, on average. And the poorer – and presumably more unequal – a country, the greater the happiness divide.

New Scientist

Liberalism is a tantrum against reality. It assumes that society is like a parent, and if we didn’t get what we wanted for Christmas, it was just that parent being unfair. Conservatives are happier because they cast aside this neurotic outlook. They recognize that thinking of society as a parent makes one disempowered to do anything but misbehave. Viewing society instead as a series of natural forces allows one to do what is necessary to enjoy it and survive well.

And to cap it off:

As a general rule, the more unequal a place is, the more Democratic; the more equal, the more Republican. The gap between rich and poor in Washington is nearly twice as great as in strongly Republican Charlotte, N.C.; and more than twice as great as in Republican-leaning Phoenix, Fort Worth, Indianapolis and Anaheim.

My fellow conservatives and Republicans have tended not to worry very much about the widening of income inequalities. As long as there exists equality of opportunity — as long as everybody’s income is rising — who cares if some people get rich faster than others? Societies that try too hard to enforce equality deny important freedoms and inhibit wealth-creating enterprise. Individuals who worry overmuch about inequality can succumb to life-distorting envy and resentment.

New York Times

Are people Republican because they’re in an equal place, and people Democratic because they’re in unequal ones? Possibly. Or possibly there’s not much of a link between the two at all that we can derive from this data. However, one thing does pop up that seems worth investigating: if Democratic places are not more equal, are Democratic policies failing?

Starting with, of course, that “life-distorting envy and resentment” — that sounds like the tantrum I was describing above. A tantrum against parents and the world at large that re-affirms the inability of the individual to make positive change. Maybe the authors of The Spirit Level should consider that harder than their inaccurate conclusions about cause and effect from their statistics.

(As a side note, I found this amusement:

Two Brown University economists have created a new data set explaining differences in the world’s current per capita gross domestic products (GDPs). In a National Bureau of Economic Research working paper, Louis Putterman and David N. Weil introduce a “World Migration Matrix” showing that inequality among countries can be largely explained by where the ancestors of each country’s people lived some 500 years ago. “What matters is the history of the people who live in a country today, more than the history of the country itself,” they say.

When Putterman and Weil used the matrix to investigate the effects of the post-1500 population movements on income differences today, the results were “almost breathtaking.” The power of regional origins is illustrated by the fact that in a 125-country regression, 44 percent of the variance in current per capita GDP is accounted for by entering only the share of the population’s ancestors that lived in Europe in 1500.

eScience News

Interesting in that it uses the same approach, but looks at origins of biological individuals, not socially-applied equalizing forces.)

From this, we see the ultimate problem with statistics: reliance on them feeds a mentality that believes all actions have a cause in their symptoms, because there’s an implied parent-ness to this view of society that assumes the individual is inert (but equal). Statistics do not explain reality. They show us details, not conclusions. Too much reliance on them reverses the process of cognition, so that instead of thinking from cause to effect, we’re thinking from effect to cause and assuming the roles are reversed. That alone could lead us to mental disease, if the ensuing social problems caused by well-intentioned but delusional people did not.

When we start thinking in cause/effect logic, we are freed from the negativity of statistics. In turn, we start to encounter a world we can finally understand and manipulate. Society by its very nature of a utilitarian, bureaucratic approach tends to reduce its thinking to statistics. Only knowledge of cause/effect logic can reverse this destructive pattern.

Share on FacebookShare on RedditTweet about this on TwitterShare on LinkedIn