Lies, damn lies, and statistics
Lies, damn lies, and statistics
What percent of conversations regarding a contested issue are actually focused on assessing the evidence and determining the truth?
It’s not large. In fact, such discussions are rare as hen’s teeth.
To invent a statistic, which is what everybody does most of the time, scientists and “experts” have determined that the percent of honest conversations held by people with open minds is 0.0000397.
Precisely.
We all do it to some extent. We develop heuristic models of how the world works and then fit the data, or invent it, in order to prove our points. We may have arrived at those conclusions through deep analysis or by absorbing a gestalt, but it is the rare instance when people honestly examine new evidence and reassess whether our view is correct.
That’s why you see very few people change their minds, and why it happens over a longer period of time. The number of times somebody says “Yeah, you’re right” in the midst of a conversation is near zero.
As a practical matter most of us develop opinions based upon the information that surrounds us, shoved in our faces by others or simply repeated endlessly. We have no reason to doubt it, unless we develop them, and a general consensus develops. A few people will always dissent, but people need REASONS to doubt and it takes time for that doubt to develop. This is generally as true for you as for anyone else, because the alternative is being paralyzed by doubt all the time.
This is why propaganda is both so dangerous and so effective. By creating an echo chamber, people are led to believe that a consensus exists and that there is little reason to doubt it. Dissenters are simply outliers, and outliers exist for everything.
Statistics are especially useful as propaganda, because they lend an air of scientific credibility in the same way that quoting a Bible verse to a believer does. The statistic (or verse) may not actually support the point, but we rarely do a deep dive. We tend to trust authorities.
So here is a good example of how we can be led astray, and how propagandists can use statistics to reinforce their points in a way that requires people to do a deep dive to see why they are misleading.
First, there has been no noticeable increase in days over 100 degrees per year in Phoenix over the last 2 decades – one of the hottest parts of the state. Indeed, you will notice that 2023 is so far a below average year. pic.twitter.com/F2xeOS8dQa
— Cherry "Born Maskless And Staying That Way" Davis (@DACDAC4DAC) July 21, 2023
The chart is tweeted out by the Financial Times’ “Chief Data Reporter.” You have a credible source, a great credential, and a scientific looking chart. Bam! Very impressive. And it reinforces what everybody else is saying too!
Must be not only true, but look at the data! It is shocking. Lots of other people are saying the same shocking thing too! People are dying from heat all of a sudden. A big spike. Scary sh!t, man.
Notice how a huge spike started about 5 years ago? This is very troubling indeed.
But wait…are temperatures really spiking THAT much in just 5-6 years? I guess it must be so, because look at the data. Except other data doesn’t seem to quite fit…but that other data wasn’t provided by the Financial Times. It came from some schlub. Conspiracy theorist.
We can’t listen to conspiracy theorists in the midst of an emergency, can we? They are spreaders of misinformation. Better have Facebook take a look at that. Maybe the good folks at Homeland Security and the information fascists in the EU look into how we can limit misinformation.
Look at that spike in deaths! We can’t afford misinformation in an emergency.
Third, the EPA also provides a clue as to what is happening. In its Technical Paper from 2017, it admonished coroners that "it has been well documented that many deaths associated with extreme heat are not identified…by the medical examiner."
Well, let's go back to the data! pic.twitter.com/cypDRGlU5D
— Cherry "Born Maskless And Staying That Way" Davis (@DACDAC4DAC) July 21, 2023
But why 5-6 years? That really does seem a bit strange if you think about it. Climate change is measured in decades, not centuries, and the temperature shift is 1-2 degrees over the next century they say. Call it 5 degrees even. Will that cause a major spike in deaths?
Look at the jump from 2015 to 2016, and the HUGE spike from 16 to 17 to 18 to 19… That seems very weird, if you think about it.
So don’t. This is an emergency.
An emergency so great that the government started telling doctors that they were undercounting deaths from heat around the same time the spike in deaths occurred.
Nice government folks looking out for us. They warn us that deaths from heat were undercounted, and the coroners fixed that! And see! More deaths from heat.
Just look at the numbers!
Of course, John knows this. But he has a mission to paint as grim a picture as possible. Many people won't dig into the numbers. They will just accept them. John knows this as well.
This is where statistics lie to serve a particular agenda. You can thank data experts like John.
— Cherry "Born Maskless And Staying That Way" Davis (@DACDAC4DAC) July 21, 2023
Of course, the real answer is that the government wanted more deaths from heat counted, and they got more deaths from heat counted. The change in numbers? Who knows? Could be more, could be less, might be the same. Because they changed how they count them.
For all I know deaths from heat were undercounted before, or overcounted, or neither. Because we set a convention for how to assess that and that assessment sets the standard. Was it hyperthermia? Was it dehydration? Was it a heart attack caused by extra stress on the heart? Was it…? You set a standard, use that standard, and that is the number.
Changing the standards makes the statistics incomparable. It may make them better or worse, but certainly they are not comparable.
I won’t assess the motives of the EPA and others for suggesting a change in how the deaths from heat are accounted, because I don’t know. Clearly there could be reasons for making a change that are totally legitimate, and I won’t look for boogymen unless I have a reason.
But John Burn-Murdoch of the Financial Times? He is clearly trying to lie to you. He is a data analyst. He knows that he is lying with statistics.
We see this all the time, especially during the past few years with COVID, climate change, and tons of other contested issues. It is propaganda, not journalism or science.
As I started this essay I suggested that this was pretty normal, and it is. We should expect it when dealing with people who are engaged in contested issues. But generally we rely on somebody, somewhere, to provide data that is at least somewhat reliable. The media and scientists used to perform this function, if not perfectly, at least with enough disagreement and integrity that we could piece together some dataset to argue over that we could generally agree on.
Not anymore. Scientists and journalists are just as likely to engage in propaganda as anybody else, and it takes a deeper and deeper dive than ever to get a sense of what might be true and what the uncertainty levels are.
That’s why it is vital to keep bringing up these failures in truth-telling made by these journalists who are really advocates. If people don’t change their minds right away, and they almost never do, it is our job to undermine the credibility of the “amen chorus” and propagandists. People will have their own “Aha!” moments.
We all get red-pilled at different moments.
Post a Comment