Basic science and statistical understanding

[MrsTerryPratchett]

I'm not a scientist. I do have a BSc. but nothing more impressive. I do wince when I hear some assumptions that people make who don't have basic science (and statistical) knowledge. What basic scientific or statistical principles do you wish the general public (and members thereof) knew about?

I'll start. Causation and correlation. A correlation doesn't prove causation. Particularly when that correlation is a correlation of one. "I had the flu vaccine and got the flu really badly" isn't causation.

Risk. Relative risk particularly.

@Smallwhiterat

Risk. Relative risk particularly.

Good one. I did a module on risk and how humans are very poor at assessing it!

%
many people don't get it

Risk, yes, although I think a lot of the issue with that is not really a maths problem - for example, someone could fully understand the maths that explains how relatively safe planes are, yet still be terrified of flying.

What exponential actually means....

Error bars.

Or just statistical variation. Cases doubling could mean from 2->4 (not significant, could easily be random variation) or it could mean from 200->400 which is clearly more significant.

Data without blips is a sign of some dodgy manipulation.

I wish people could read the small print. Including journalists.
All this 'ONS data shows teachers aren't at more risk than other professions' stuff in the news and on here. If you read the actual data it compares covid cases in school staff to 'other essential workers'. If you then look at the little asterisk next to 'other essential workers' and read the small print it actually isn't comparing to ALL other essential workers, it's only healthcare, care and emergency services.
Which puts a whole different angle because what it actually shows is teachers are at as high a risk as healthcare and care staff, which, i think we'll all agree is pretty high.
But so many people seem to not be able to see this even when it is pointed out. Admittedly, the government have actively presented it in a misleading way, but this is exactly why we should always read the small print!

Not strictly statistics, but risk assessment as a combination of risk AND consequence.

This will sound petty, but the meaning of asymptotic. Lots of posters here are using it interchangeably with asymptomatic, which I'm sure is simply autocorrect or auto-suggest.

That reliability of estimates eg of the proportion of people believing some statement to be true - depends on the sample size and not on what proportion of the population you asked. So estimates derived from a sample of 10,000 are perfectly accurate even when it's eg an estimate of the number of people in the UK who would describe themselves as vegan. It doesn't matter that less than 1% of the population has been sampled.

What does matter is that the sample is an unbiased and random sample, which is much more difficult to achieve.

Journalists want a headline in a lot of cases and bloody know they are misrepresenting for clicks.

On the journalism point, I wish people understood that a percentage increase in risk is only as important as the original risk. So for example if my chance of contracting Ploopy is 0.000000001% I don't actually care if that risk doubles if I eat cake. If my Ploopy risk is one in two to start with, I really care if cake doubles that. Depending on how serious Ploopy is to my health.

The fact that population data doesn't predict outcome for individuals would be a good place to start.

Populations and samples, then uncertainty.
Definitely risk.
The idea of group and individual impacts.

Evaluation of reliability of sources.
What you've seen circulating on FB is more likely to be untrue/inaccurate than something published in a peer reviewed reputable article. Or even a serious news program / serious news website.

@Magpiecomplex

This will sound petty, but the meaning of asymptotic. Lots of posters here are using it interchangeably with asymptomatic, which I'm sure is simply autocorrect or auto-suggest.

It's my thread. Pettiness is entirely allowed.

Oh, (struggling to see a Covid use but it bugs me) - now everyone has worked out that you can't give the bottom 20% a pay rise without increasing the mean salary - if only people could understand that the same is not true of the median salary. Mathematically it is perfectly possible to increase the minimum wage to 80% or even 99% of the median without affecting the median.

Even more basically; mean, median and mode can be wildly different. You can pick the 'average' accordingly based on your own interests.

And that is a COVID thing because people keep saying, "the average age of people who die from COVID is over the average age people generally die" like that's an indication that we shouldn't care. That very much depends on how those numbers are calculated.

My issue is that no matter how much people understand statistics, they can always be represented to prove a point by choosing what you show and how. A section of one of my modules for statistics was specifically on how the media presents data and it's definitely made me more sceptical of the "proof" in situations like this.

There is a strong tendency to disregard numbers and statistics when faced with detailed information about one very rare event. No matter how rare an event is a single occurrence can be seen as "proof" that the statistics are wrong.

On the subject of headlines: reading beyond them to find what the actual numbers are.

X introducing new measures to combat virus surge doesn’t mean that they’ve also lost control of the virus if a surge is 100 new cases a day in a population roughly the size of the uk.

All of these!
Also, a particular bugbear of mine is the use of the term 'the science', usually preceding a garbled 'explanation' involving a clear lack of understanding of correlation v causation, risk, percentages, sample sizes, individual v population level effects etc. Oooh that was quite cathartic. This thread speaks to my inner nerd.

Oooh that was quite cathartic. This thread speaks to my inner nerd.

(Can I ask what asymptotic means??) 😳

Asymptomatic is where you never have any symptoms.

Presymptomatic is the period before you have symptoms but are infectious.

I don’t think confusion about this was helped at all by a lot of papers describing their results as ‘asymptomatic at the point of testing’ and people missing the crucial ‘at the point of testing’ off the end.

@Augustbreeze

(Can I ask what asymptotic means??) 😳

I believe it's a line approaching another line on a graph without actually meeting. In philosophy it's why you can't shoot a turtle.

Nerd joke.