Quote:

Originally Posted by **helgymatt** No, your values for CO2 should work just fine for analysis. Your transformation from pH to CO2 is simply a unit conversion, and both the variances and distribution of these data are normal. |

I think you need to talk to your statistics professor on this one.

A logarithmic manipulation is a non-linear transformation of the data, which does not preserve the distribution characteristic of the data. You can verify this yourself with a simple test.

Start with this set of 100 points:

S1 = {1.0, 1.1. 1.2, ….9.8, 9.9 10.0}

If you plot the distribution of this data it obviously has a uniform distribution.

Now take the log of each data point and get a new set:

S2 = {0.0, 0.041, 0,079, … 0.9912, 0.9956, 1}

If you plot this set you will see that it is no longer uniform.

Actually I can take criticism very well. Unfortunately, your understanding of this problem is not up to speed. This is actually a very complicated statistics problem and does not resolve with the simplistic approach you are using.

This stuff is way beyond what anyone else on this board is interested in. I am more than willing to work with you. PM me and I will share my data with you and discuss the issues. Otherwise, Have a nice day.