I run a linear mixed model; nested it for the species ID. Hereby I have transformed my response variable. This is what my output is:

enter image description here

Now I read that I the parameter estimate tells me how much the squareroot Wheel_prop of the outcome changes with every increase of one unit in the factor (Context) 2 for example and to know about the untransformed variable I have to translate the estimate with the anit-squareroot. (so ^2)

Now my question is, if there could be any error-interpretations because of the changing algebraic signs? For example translate the estimate of the factor (context)2 it would be

## [1] 19.45727388

So from getting less it would change into getting more?

  • $\begingroup$ First consider that $2+3=5$ but $2^2+3^2=13\neq 5^2$. The duplicate question concerns taking the 4th root of the response, but the answers apply to square roots too. (Also please type out code rather than take screenshots, for ease of reading.) $\endgroup$ Jan 7 at 20:01
  • 1
    $\begingroup$ Just to record that I looked at this question and gave up because for me the output is far too difficult to read. Also, the more you assume that people are familiar with the software you use, and the more you expect that people will read through a lot of output to work out what you are doing, the less likely it is that good answers will arrive quickly. $\endgroup$
    – Nick Cox
    Jan 8 at 13:39