Register here - as a registered user you get more features and fewer ads.
I would like to see a wider range of numbers on the horizontal axis so that we could see if this is just a random effect, and to test against the possibility of selection of data.
One would expect the cumulative error (plotted on the vertical axis) to have a peak at 2.72/2. Presumably, if the diameters of the circles were completely random, the scatter on the graph would also be random, so the significance of the graph depends on the ratio of the maximum to the minimum, and the distribution or scatter of values between them; we are not shown this.
A point which always troubles me about this kind of data is the question "how do you measure the diameters of stone circles, and what errors are there in the data?". Are they in fact perfectly circular, or rather, with what degree of precision do they actually fit a circle ? What is the meaning of quoting the size of a megalithic yard to a precision of 0.7% (i.e. 2 parts in 272) when the original data has unspecified (and probably much lower) precision ?
Something is not right. This message is just to keep things from messing up down the road