That's very kind of you, but it's still mostly gobbledy gook to me. ( I did do further maths A level about a million years ago, but with mechanics and not statistics. Mostly it comes in useful doing equations to find out the right sized round tin for a cake recipe giving a rectangular tin size! or vice versa.) But something about your posts over the last few years makes me trust you and your results.
Remarkably few researchers, 'experts' and scientists have a good grasp of stats. At least that has been my experience over the last 40 years. We've also gone from being pragmatic experimentalists to theoretical modellers, which is why a great deal of science has descended into expertism, fashion and celebrity, imho.
so standard deviation over a period of time is key. surely if you check the standard deviation over first hundred years 1824-1924 and then separately for 1924-2024 you can immediately tell if there is a significant difference between the two periods? a significant... change in variance?
Yes, you can formalise what our eyeballs already tell us by partitioning the data and running either a classic ANOVA or Kruskal-Wallis. I’ve been toying with the idea of doing this so may well take us through the process in the next article.
"I shall be doing this in best bib and tucker ... I shall also use my (expensive) professional stats..."
For a split second I thought you said you would be wearing expensive SPATS! You know, the things gentlemen used to wear to protect their shoes.
I'd love to try spats and carry a cane!
Bertie Wooster!!!
That's very kind of you, but it's still mostly gobbledy gook to me. ( I did do further maths A level about a million years ago, but with mechanics and not statistics. Mostly it comes in useful doing equations to find out the right sized round tin for a cake recipe giving a rectangular tin size! or vice versa.) But something about your posts over the last few years makes me trust you and your results.
Remarkably few researchers, 'experts' and scientists have a good grasp of stats. At least that has been my experience over the last 40 years. We've also gone from being pragmatic experimentalists to theoretical modellers, which is why a great deal of science has descended into expertism, fashion and celebrity, imho.
stats breaking my brain.
so standard deviation over a period of time is key. surely if you check the standard deviation over first hundred years 1824-1924 and then separately for 1924-2024 you can immediately tell if there is a significant difference between the two periods? a significant... change in variance?
Yes, you can formalise what our eyeballs already tell us by partitioning the data and running either a classic ANOVA or Kruskal-Wallis. I’ve been toying with the idea of doing this so may well take us through the process in the next article.