I wouldn't put yourself out too much as I'm not really interested in your analysis and after over 40 yrs in the NHS I'm well aware of how statistics can be twisted to show whatever result you require, if you so desire.
Others may be interested, just not me. Good night.
All right, I guess that settles it.
I may have some experience in statistical inference professionally, and may have spent a lifetime in the area.
I emphasise may have.
Since you have no interest, I wonder why you posted the links. Note, I didn't dismiss it outright, and kept my impressions to myself.
Could it be possible that my hypothetical example of dualling was tad hard to trivially reject.
Not for you but others on statistics in social sciences. And this without questioning who sponsored what reasearch. A little nugget from "How to lie with statistics"
In the 1950s at the height of racial tension, a questionnaire was handed out to people in a study in the US. One of the questions asked of people was whether African Americans had sane job and social mobility opportunities as the predominantly white general population.
A large percentage of the respondents replied in the affirmative. Based on this sample one could reach the conclusion that society was largely egalitarian with racial background not a primary obstacle, if at all.
But the questionnaire was trickier.
The same questionnaire asked lots of other questions about people's opinions and preferences. If you analysed the answers to these questions you would come to the opposite conclusion. That equality was not the norm. That many white people held very prejudicial views about African Americans with consequent discrimination.
And here's the rub.
There was a statistically very high correlation, way beyond the highest confidence intervals, that showed that those stated that African Americans were not discriminated against, and in fact, had equal opportunities to white segments of the population were the ones most likely to hold racist and discriminatory attitudes towards African Americans and other races.
Conversely, those who felt that there was discrimination, and obstacles, were least likely to hold racist or discriminatory attitudes.
Now I do not make this up. Anyone curious enough is welcome to check the source I have referred.
So this was one example where if the study had not been carried out carefully, an opposite conclusion may have been reached.
That the US was largely equal opportunity with few or bo obstacles in terms of either attitude or opportunity for African Americans.
Now, as Welsh Gas has done the usual ("I am done, and will not discuss with you."), may I respond:
I shall discuss with you to the end Sir, and will yield only to the truth. I am not interested in anything else. Accuracy is the only thing that matters.
I shall take it that unless you can back yourself up with some substance you could not last the tough course that is truth.