“Every minute, 34.2 million men and women copulate. Only 5.7 percent of all intercourse results in fertilization, but the combined ejaculate, at a volume of forty-five thousand litters a minute, contains 1,990 billion (with deviations in the last decimal place) living spermatozoa. The same number of female eggs could be fertilized sixty times an hour with a minimal ratio of one spermatozoon to one egg, in which impossible case three million children would be conceived per second. But this, too, is only a statistical manipulation.”
In “One Human Minute” by Stanislaw Lem
Lem never fails to disappoint. This is one of those long-forgotten Lem books no one remembers anymore. I read it more than 20 years ago, and it still packs quite a punch. My love with book reviewing started around the time I read this three-essay-volume (“One Human Minute”, “The Upside Down Evolution”, and “The World as Cataclysm”) comprising reviews of non-existent books… As always, when a book is this good my mind goes on a tangent…
Boss: "Will this work?"
Statistician: "Probability of success is 90% so..."
Boss: "Let's do it."
Boss: "It didn't work. You're fired."
Hopefully the statistician will use his period of unemployment to get better at his job. If you're offered a wager where there is a 90% chance that you will win 5 euros and a 10% chance you will be shot dead then you have to be a very poor statistician to think "the probability of success is 90% so I'll take it". The correct response here is "No, there is too much business risk". Translation: the chance that the business might make a bit more money is not worth me getting fired. If it means the business is likely to miss out on an opportunity then it is not the statistician's fault that his employer is managed by trigger-happy clowns.
I suspect that even if the level of statistical information were provided it would be of little benefit to much of the public, since levels of numeracy are frighteningly low even among those who are otherwise highly educated. You only have to read the newspapers after election days to see that quite quickly. Never mind statistical terms like variance, are standard deviation, etc.; the term average and how to apply it barely understood. Even more misunderstood is probability. For example, the gambler's fallacy is widespread - indeed it is a psychological trait that is exploited by casinos, among others. But most people wouldn't know how to interpret a margin of error. It's an enormously technical aspect of statistics that is itself a statistical measurement. It's also a 'guess' based on X rolls of the die. 'If we ran this simulation x times the range of results would be Y, with the most common outcome being Z'. Understanding that requires a level of numerical literacy (rather than just numeracy) that very, very few have. Which isn't a criticism of them.
So margin of error wouldn't be helpful and would often be used to dismiss statistics.
People don't want to know, and wouldn't find it useful, to know, that inflation ranges from X to Y percent. The same way I just want to know 'should I wear my raincoat today?’ not the statistical likelihood it will rain and the way it was worked out. I personally think the bigger problem is that the conclusions reported don't always really follow from the data or the statistics are willfully misrepresented to support a particular point of view. As a result, the conclusions don't fit with people's perceptions and they therefore don't trust the statistics. People routinely fail to distinguish between median and average let alone more complex concepts. However, while this might work out in the long term, how does one engage the problem right now? Take fake news: there is a strong argument to debunk fake news were ever possible but this doesn't solve the more important problem that quite a lot of people want to believe this stuff. Maybe it would be wise to invest in critical thinking from an early age. The problem, though, is that if you report that level of detail, your piece becomes unreadable to a lot of readers. So journalists leave it out, or use tiny footnotes that are only read by people who already had a pretty good idea of the data's limitations.
Most weather forecasts in Portugal are not probabilistic, precisely because there are so many people who are not statistically literate. There is considerable debate in weather forecasting circles as to whether it would be beneficial to explain forecasting techniques in more detail, and to give statistical probabilities as part of the forecast, but at the moment they have erred on the side of tradition and simply forecast 'rain' or 'sunshine' (if the models don't show a clear likelihood for any particular outcome there's always the classic fallback of 'sunny with scattered showers'). Some obviously find this infantilising, but others prefer the clarity of the advice. People get upset either way.
I'm not sure how you resolve that issue. Mandatory statistics awareness courses at school? Add them to the pile, along with nutrition, economics and media awareness.