I can’t not include statistics in this series and yet the subject is so large that I also don’t feel that I can do it justice within the space of a single post. Nevertheless, we have to talk about it because I don’t think it’s an exaggeration to say that in the last year we have seen the greatest explosion of statistical propaganda in history courtesy of the corona event. People’s lives in most western countries have been governed by those statistics to an incredible fashion as the increase or decrease in case numbers determined your likelihood of having a job next week or your ability to go out after a certain time of night or see a member of your family or visit a friend or any number of other things you would once have taken for granted but that could now only be granted to you if a particular number moved in the right direction.
One of the strangest things with the corona event was that everybody, including some notable public intellectuals who should have known better, simply took the statistics at face value. But the use of statistics as a propaganda tool has been in play for decades, perhaps centuries. How many people actually know, for example, how GDP or inflation or unemployment are calculated? These numbers are talked about all the time but I doubt more than 5% of society could give an adequate account of how they are measured. This is important because these measurements all have technical definitions and governments often change those technical definitions in order to make the numbers look better. For example, the number of hours needing to be worked in a week for somebody to be considered ‘employed’ has fallen over time. This change has the effect of reducing the unemployment rate, which is something that governments have a strong interest in. But this change in definition is not reflected in the number itself which just shows a simple percentage figure. If you compare the unemployment rate from now to the one from thirty years ago, it’s not really the same but we talk about it as if it is. Also hidden from the unemployment figure are the people who aren’t even looking for work. They are ‘unemployed’ in the everyday sense of the word but they are not included in the technical definition. Then there are the people who ‘underemployed’. They would like to work more but they can’t find that work. They are not happy with their employment status but the government considers them employed and therefore satisfied. It’s in all these hidden nuances where the propaganda value of statistics lies.
There’s a whole book to be written on the misuse of statistics in relation to the corona event but let’s just look at one issue: the definition of a case. In my book on the corona event, I noted the change in the definition of a case from the first SARS event to the corona event which is a move from a case being about symptoms that are diagnosed by a doctor to being about test results that are generated by a lab technician.
The word case has a general meaning that a layperson would understand. If I say “there were ten thousands cases of heart attack last year” that means ten thousand people had a heart attack. If I say “there were ten thousand cases of malaria” that means ten thousand people were sick with malaria. But, with the corona event, if I say “there were ten thousand cases of covid” that does not mean that ten thousand people were sick. It means ten thousand people returned a positive result to a PCR test. In this way, the case number is misleading. The public understanding of it does not reflect the way in which the number is generated. Given that the asymptomatic rate for tests is apparently around 50%, it’s misleading by a lot.
Consider also the change in the process required for somebody to become a case. With SARS, you became a case by going to hospital with what was presumably an acute illness. A doctor would then diagnose your symptoms and a public health bureaucrat would contact trace your movements to find a link with a prior case. With corona, you can have any severity of symptoms or even no symptoms. You can attend one of your friendly local testing centres where a person will put a stick up your nose and send it to a laboratory where a technician will run it through the process to determine a positive or negative result. That’s how you become a case. The procedure is completely different as is the definition. But we talk about cases of SARS, cases of influenza and cases of covid as if they were the comparable. That’s the danger of statistics.
As with all propaganda, the antidote is to know how things work in the real world. When it comes to statistics, that means you need to know the technical definition of the statistic (not the assumed folk meaning), the person or organisation who defines that meaning, the person or organisation who is responsible for collecting the data, the method of data collection and any mathematical transformations that are applied. That’s before you even get into the methods of data presentation (eg. graphs) or any actual methods of statistical analysis. Of course, that’s a lot of work and most people are never going to spend the time to do that work. Hence, the power of statistics as a propaganda tool.
Here is a paradigm example of statistical propaganda that came to my attention just yesterday.
Note the wording of the headline “The US economy lost 140,000 jobs in December. All of them were held by women.” This is an astonishing claim. Its meaning is clear: every single job that was lost in December was a job held by a woman. That would be an extraordinary fact and, if the tweets that were flying around about the article were any indication, that is how most of the people who read the article understood it.
Of course, it’s simply untrue. As the article slyly mentions about halfway down “These are net numbers, which can mask some of the underlying churn in the labor market.” The 140,000 figure is an aggregate. What happened in reality was some jobs were lost and some more were added. Individual men and women lost jobs and individual men and women gained jobs but, when you take the aggregate, there were 140,000 fewer jobs overall and 156,000 fewer women were employed. The propaganda effect of the headline is to pretend that the 140,000 figure represents individual jobs losses and not aggregate ones. This is a small but very important difference between an aggregate figure and the individual cases that make it up. The aggregate figure is still notable and the article goes on to explain the quite logical reasons for it which is that the pandemic lockdowns have disproportionately hit industries dominated by women. (In relation to school closures, it’s also notable that some of the school closures were driven by education unions which refused to re-open the schools so presumably these job losses were mostly supported by the employees themselves).
It’s always wise to try and imagine what aggregate numbers mean at the level of the individual as that is the everyday effect of whatever is being measured and that is what most of us really care about. To return to the corona cases, if you have 10,000 cases, that sounds bad. But if 50% of those cases never had any symptoms then it’s less bad. If only 5% of the cases developed into serious illness and only 2% ended up in hospital that means 500 people were really sick and 200 of those had to go to hospital. Suddenly the whole thing takes on a very different perspective. You can then factor in the age and medical status of the people to get a more fine grained understanding of what is going on at the ground level.
The use of statistics ties in very closely with the use of expertise and science in propaganda. Statistics are almost always compiled by experts but one of the main things you are taught in science is to be incredibly precise in your language. This is why maths is the language of science because it allows for precision. An ‘error’ such as the one made by CNN in its headline would earn you a failing grade on a year 8 maths exam, but in propaganda it’s all par for the course. Just a way to nudge you in the direction of the preferred interpretation. In this way, most reporting on science in the media is biased from the outset and actually serves to tarnish the reputation of science and the experts. Rather than deal with that problem at its source, the media has simply decided to label anybody who questions the statistics or expert testimony as a conspiracy theorist.
As educated consumers of propaganda, we should always take every statistic, graph and chart shown in the media with the highest scepticism. If you haven’t got the time to find out exactly what the numbers mean, it’s best to assume they are simply being used to push a particular angle on the story.
All posts in this series:
- Propaganda School: Introduction
- Propaganda School Part 1: Guilt by Association
- Propaganda School Part 2: The Passive Voice
- Propaganda School Part 3: Editorialising the News
- Propaganda School Part 4: Headlines and Taglines
- Propaganda School Part 5: Anchoring
- Propaganda School Part 6: Metaphor
- Propaganda School Part 7: Predicting-the-Future
- Propaganda School Part 8: Appeal to Authority
- Propaganda School Part 9: Buzzwords
- Propaganda School Part 10: Lies, damned lies and statistics
- Propaganda School Part 11: Revenge on the Nerds