But there is a correlation, if the datapoints represent the paired occurrence of each over a sequence of time periods.
Likewise, autism and mobile phones; autism and climate alarmist articles; autism and petrol consumption; autism onset and grandparents still living. Probably not autism and fridge ownership; nor autism and TV ownership (well, in the first word at least, as the market in those good is saturated). However, on second fridges and third TVs: who knows, there probably is correlation.
[Note. As with everything that is increasing with time, it is positively correlated with everything else that is increasing with time. There are also possible problems with the consistency of what is being measured on autism: it could be diagnostic skill/effectiveness and/or changing the threshold for reporting occurrence.]
It’s the causation evidence that is lacking. That’s a scientific problem for scientists and a statistical problem for statisticians.
What we currently have, with just correlations, is science and statistics for and from journalists and politicians.
Best regards
JamesV
Now we have such great computers, it would surely be possible to come up with a perfectly accurate mathematical description of every single non-existent correlation you could care to investigate. Perhaps autism and vaccine uptake are related by something like sqrt (v/1.228)*sin (3a/9.1v)+a-12.1v^4.3=0?
Marksany
You gotta love the graph nerds in the comments there now
diogenes
remind me never to employ nigel…..
JAGMAN
In the words of that great philosopher Homer J Simpson “pffftt statistics, you can prove anything with statistics”
But there is a correlation, if the datapoints represent the paired occurrence of each over a sequence of time periods.
Likewise, autism and mobile phones; autism and climate alarmist articles; autism and petrol consumption; autism onset and grandparents still living. Probably not autism and fridge ownership; nor autism and TV ownership (well, in the first word at least, as the market in those good is saturated). However, on second fridges and third TVs: who knows, there probably is correlation.
[Note. As with everything that is increasing with time, it is positively correlated with everything else that is increasing with time. There are also possible problems with the consistency of what is being measured on autism: it could be diagnostic skill/effectiveness and/or changing the threshold for reporting occurrence.]
It’s the causation evidence that is lacking. That’s a scientific problem for scientists and a statistical problem for statisticians.
What we currently have, with just correlations, is science and statistics for and from journalists and politicians.
Best regards
Now we have such great computers, it would surely be possible to come up with a perfectly accurate mathematical description of every single non-existent correlation you could care to investigate. Perhaps autism and vaccine uptake are related by something like sqrt (v/1.228)*sin (3a/9.1v)+a-12.1v^4.3=0?
You gotta love the graph nerds in the comments there now
remind me never to employ nigel…..
In the words of that great philosopher Homer J Simpson “pffftt statistics, you can prove anything with statistics”
@Jagman: but only to the likes of Homer.
Best regards
Nigel – which seems plenty good enough to satisfy most politicians. 🙂