Links - Misinformation in Various Forms

Imperfection is required.

Today’s set of links will be an attempt to connect some dots among various posts along the lines of misinformation. Each should be read in its entirety and considered independently from one another. However, I think the connections among them is information itself.

The first is Arnold Kling writing specifically on misinformation as it relates to dissent.

The TL;DR for this essay:

The authority to police disinformation will inevitably be used to crush dissent.

Crushing dissent is bad.

Therefore, we should not give anyone the authority to police misinformation.

To dissent is to express a belief or opinion or claim that differs from . . . conventional wisdom; the Party Line; elite opinion; the scientific consensus.

Misinformation is bad, and it ideally would not exist. However, the devil is certainly in the details. Eliminating it is extremely fraught. As Kling points out, any attempts at doing so eventually results in quashing of dissent. Without dissent, there is no progress, and the best we can hope for is stagnation through oppression.

As he concludes,

Anyone can be wrong. The challenge is to arrive at beliefs that are more likely to be right. Trusting experts is a heuristic that is often a good choice. But it does not follow that dissent should be stamped out.

Dissent from an information standpoint is a method of revealed viewpoint. Often this comes with emotion attached, but this has signal value as well as noise. Dissent can be a rejection of the authority either because of differing preferences, beliefs, or values while often being some combination of these. It offers insight even when that insight should be rejected.

Similarly, insight can be lost in good-faith attempts to improve information quality. That brings us to DYNOMIGHT arguing that we should present study results with lots of apparently misleading or superfluous digits.

He makes a great point, which certainly qualifies as explicitly counter conventional wisdom. One framing of his point is that precision has multiple angles to it. Consider:

I beg you, do not listen to these people. Write all the digits.

Of course, they’re right that when you write 5.67664%, the final digits say nothing about how much better left-handed students might be at whatever was tested. With only 446 students, it’s impossible to estimate more than 2-3 digits with any accuracy.

But the argument isn’t just that those digits are unnecessary. It’s that they’re misleading. Supposedly, printing that many digits implies that the true population difference must be somewhere between 5.676635 and 5.676645%. Which is almost certainly untrue.

But… who says it implies that? If your point is that there’s lots of uncertainty, then add a confidence interval or write “±” whatever. Destroying information is a weird way to show an uncertainty band.

And deleting digits does destroy information. Yes it does. Yes it does! Important information.

He goes on to show that the lack of all those “extra” digits, which certainly do have a downside of potentially implying greater prediction accuracy vis-à-vis precision, leave us woefully blind to the studied data's underlying results. Many times this could be dismissed as not important, but often it will be and in unexpected ways. Using his stylized example, we could perhaps see that the study itself doesn't pass the sniff test as it doesn't comply with reasonable assumptions about the real world much less known facts like the proportion of lefthanded people.

The bottom line is that context matters. If you don’t understand what your seeing, you can easily be led astray. And that is exactly the point Scott Sumner is making when he describes a mischievous equation.

GDP = C + I + G + (X - M)

…

Many people wrongly assume that imports reduce GDP, because they appear in the equation with a minus sign.

Many people wrongly assume that if consumers become pessimistic and decide to spend less while saving more, this will cause GDP to fall. That’s because consumer spending appears in the equation with a positive sign.

Now to be clear, it is possible to write down plausible models where a decision to import more causes GDP to go down. And it is possible to write down plausible models where a decision to save more causes GDP to go down. I view these models as almost entirely useless, but some economists see them as having some validity.

…

I wish textbooks did not include the GDP = C + I + G + (X - M) equation. I fear that it leads to negative knowledge. I wish that 90% of Macro 101 was classical economics where the economy is considered to be on the production possibilities curve, i.e., where more of one type of output means less of another type of output. I wish the other 10% of macro were business cycle theory where students were taught that occasional monetary policy mistakes lead to fluctuations in NGDP. And these NGDP fluctuations impact the labor market due to sticky nominal wages and impact the financial system due to nominal debt contracts.

This is not a pedantic point. It is a crucial one as many policymakers and powerful influencers of policy have been misusing/misunderstanding this equation leading them to promote anti-free trade ideas. These mistakes therefore have material, negative real-world consequences for us all.

Similarly, when the media give us a biased perspective or when a dissenting view is not allowed in polite or official company, we get a warped perspective with, again, negative knock-on effects. Which brings us to the last link, also from Scott Sumner.

Two famous 90s stars tragically died by drowning, Cranberries singer Dolores O’Riordan and television star Matthew Perry. Both drownings were drug induced. Only one is labeled as such and remembered as a drug related death. As Sumner writes,

My concern here is with the media. If voters are to make intelligent decisions about drug policy, it is essential that the media not become an arm of government propaganda. Thus far, they have failed to provide objective information on the effects of drug use, as they report the consequences of illegal drug use in a radically different fashion from the way [they] report on the consequences of legal drugs such as alcohol. Please, just give us the facts.

The sin of misinformation is all too easy to commit. As a solution to misinformation, it is preferrable that we build resilience rather than attempting resistance. Or, as Arnold Kling might say, when it comes to information, make it easy to fix rather than hard to break.