Hello Rachel,

Thanks for the nice words. And you wrote i.a.

Quote Originally Posted by thefreeprisoner View Post

Might you perhaps be interested in helping _kim_ with the wonderful survey that she's building? We definitely need somebody who knows what they are doing to analyse the results to help us figure out which conclusions (if any) we might be able to draw with some degree of mathematical confidence (if that makes any sense).

Rachel xx

O yes, that makes sense: Statistics is about it i.e. about saying things about classes of things with some mathematical confidence and probabilistic justification, based on sampling from these classes. As to statistical surveys and inferring results from these: My most insistent advice is probably: KISS!

Ahhh... here I am not doing innuendo nor "Nudge, nudge, know what I mean?!" but indulging in acronyms programmers like: "KISS!" = "Keep It Simple, Stupid!" (also with more polite variants: see the net). That is, in the present case: averages, standard deviations and correlation coefficients often bring one as far as one can rationally get with complex data (that are not very extensive). These have the additional advantage that they are fairly easy to get, calculate and understand. (The problem with much statistics of the more complicated kind is that these often involve assumptions of distributions that may not be justified, and assumptions of independencies that often are not justified.)

And now that I have you and since I failed doing it so far: Thank you very much for all the hard work in transcribing Dr. Bell's lecture. It's a good one and it deserves wide reading.

Best regards,


P.S. I just posted a reply to Gerwyn in another thread that you may be interested in, i.a. as regards to writing to the media: