Are perfect-reduplication and quasi-reduplication related?

[MarcoP]

I now get CORREL to 0.8535.
Calculating the associated p-value (because I can) we get the p-Value of .30621 (the result is significant at p < .05)

...

   

Hi David,
thank you again for this discussion, which is extremely helpful to me! I don't know much about statistics and it's great to have a chance to look into new things.
I am not familiar with the concept of P-value, but I guess it is similar  to You are not allowed to view links. Register or Login to view. mentioned by Rene at the end of You are not allowed to view links. Register or Login to view.. Visual inspection of the curve suggests that for 6 samples a 0.85 correlation should be significant, but I am not sure I am getting this right. 
Also, if I enter r=0.8535 and N=6 in You are not allowed to view links. Register or Login to view., I get a P-value of .030621, which is exactly 10 times smaller than what you found, and just smaller than the 0.05 threshold. So maybe this correlation measure is significant after all?

Thank you also for sharing a plot of %Red vs %Quasi. I attach a similar plot with sample labels added. As you say, evidence "suggests that reduplication and quasi reduplication are linked", but I am not sure we are on solid ground yet (your P-value argument looks like a great way to look at these results).
The plot points out two problems:

1. the obvious one that, working at such a coarse granularity as whole sections, we only have 6 samples - this implies that only a very high correlation is significant and a single outlier can make the results unreliable;
   
2. the diagram-intensive AstroCosmoZodiac pages behave differently from the other more paragraph-oriented pages.
   

I think I will try a folio by folio analysis, using the normalization system described by Nablator You are not allowed to view links. Register or Login to view., so that we can look at things at a more detailed level.