Re: green bubble syndrome

On 2024-10-11 09:51:58 +0000, Wilf said:

On 10/10/2024 at 17:28, Jolly Roger wrote:

 Apple sold 2.5 BILLION iPhones (as of 2023, so not counting 2024), and
you are trying to tell us that a survey of 1000 people is significant?
Quick question: How many times do you think 1000 goes into 2.5 billion?

 If the sample is chosen properly (and that's the critical part), results from a small but representative sample of the whole population can be statistically significant.  So just because someone has no background in statistics is a not a reason to necessarily doubt the premise.

As a mathematics and statistics graduate, I can tell you for a fact that the "premise" is utter crap. Statistics are massively misused, especially when it comes to idiotic surveys.
Surveying a small proportion of the total does not and cannot give you any meaningful results. At best it can give you a very rough approximation, but in reality can be way out. Such results *MUST* be reported as being for only the surveyed people. It's moronic and misleading to survey only 1000 people and then claim the results are true for the entire planet - the results are only ever true for those 1000 people!
Try measuring 1000 randomly cut pieces of string and then claiming every piece of string in the world is Xcm long or even has an average of Ycm long. It's utter nonsense.
Plus, it gets even worse with ridiculous things like "health studies", where they not only survery a tiny number of people, but they also usually completely ignore any other possible factors except the one or two they're interested in.