EASY! Which analysis?

#1
Dear all kind people on Talk Stats,

In a nut shell: I have two sets of data - 1 set measured by myself and another corresponding set measure by a lab. I want to use these two sets of data to get a calibration / correction factor for my method (as the lab uses a different method)

I have completed a scatter graph to get the equation of a line and have a r value of 0.83. I have also averaged the difference between my results and the lab results and got a +2.2 on my results compared to the labs.

I am right to assume that this value (the averaged difference between the data sets) is the most robust way of calculating a correction / calibration factor?

Any help would be greatly appreciated.

Kind regards,
 

ondansetron

TS Contributor
#2
That assumes a linear relationship between the values. What if the relationship is better represented by some nonlinear function (maybe extremely high and low values differ by different amounts than one another and differently from middle values, for example)?

Which way are you working? Are you trying to convert your values to the lab's or the lab's values to yours, or is it both?

What does a scatter plot look like?
 
#3
I am trying to convert my values to the labs values so in the future when I make a measurement I can use a conversion factor so my result is in line with the labs.

I have attached the scatter plot.

Many thanks
 

Attachments