To translate what you say above to "statistic-speak," you and others claim that there are apparent associations between patterns of stars and geographic locations. Such associations might yield a positive correlation of somewhere between 0.5 and 0.8. A positive correlation of 0.5 would account for 25% of the variance; a positive correlation of 0.8 would account for 64%. 0.9 would account for 81% and 1.0 wuld equal 100%. You wouldn't expect much higher than 0.8 because you are aware of inaccuracies both with angles and distance as well as changing patterns in the sky over time.
I think that one could break down the 7 sided figure you describe into angles and distances (convert the angles and distances into percentages and treat each angle/distance as part of one's sample). One then could carry out a Pearson product moment correlation of such a sample to get an actual correlation coefficient, r. One could also do a regression analysis with variations in star patterns over time--and test the regression for significance.
It wouldn't be saying much beyond what you've already been saying, but I think that many would find it interesting.
|Orion's Global Legacy||1102||gadavid||26-Apr-18 21:40|
|Re: correlation||127||gadavid||01-May-18 17:43|
|Re: correlation||104||drrayeye||02-May-18 00:25|
|Re: correlation||181||gadavid||02-May-18 18:37|