|
Post by Kage2020 on Sept 9, 2004 14:41:38 GMT -5
Okay, maybe not 'saddistics' but that is what I've always referred to it as. I don't suppose anyone on the board is up on their statistical interpretations and what the beggers actually mean? Trying to figure out appropriate non-parametric tests (I think that means something! ) for two or more variables. Mann-Whitney works for consideration of a variable in two groups, apparently, but not sure about appropriateness when working with, say, carbon and nitrogen isotope results within two groups (i.e. between two different populations, or variation between sexes within a population). And, once the appropriate test is pointed out... what do all the buggerations of numbers mean!? I've had one person tell me about "asymptotic significance" and that it means one thing, and another tell me the complete opposite. Bar-stewards! This is kind of putting a dampener on my results and discussion... <sniff> I knew that I should have avoided any form of saddistics at all costs. Darn myself for feeling that I should test for significance of variation... <whimper> Oh yes... HELP!!!!
|
|
|
Post by zholud on Sept 10, 2004 3:06:01 GMT -5
I’ve not done non-parametrics for ages, and even then it was only one or two lessons so unlike others, seems I’m of no help here… I can advise you to go and search for some mail-lists on statistics. You may try it in econometrics list, but unsure whether you’ll get feedback - ECONOMETRIC-RESEARCH@JISCMAIL.AC.UK … go on their site and join the group first… maybe there are some closer to your statistic archaeology (if there is such a thing) groups as well.
|
|
|
Post by Kage2020 on Sept 11, 2004 5:09:13 GMT -5
Well, was hoping for a quick fix... Dealing with tests of significance, appropriate tests, etc. Darned carbon/nitrogen and testing whether quantitative difference translates to a qualitative/statistical data. Damnations.
Ah well... hand it in and try and find more useful tests. Heck, even the none useful ones turned out to be partially useful. Standard boxplot has a lovely way of ignoring those that statistically outlie the main group...
Right, back to the slog. Methinks that strontium will receive attention today. And maybe oxygen as well!
|
|