Binary Logistic Regression; " Table$setRow(): value 'r2t' is not atomic "
Posted: Mon Mar 04, 2024 11:03 am
Hi,
I am trying to run a series of binary logistic regressions to compute participant scores on a cognitive task where; 1) the IV is the point differences between"Self" and "Comp" options presented to each participant on 44 discrete trials and 2) the DV is whether the participant chose "Self" or "Comp" in response to each trial. I also run a box-tidwell test to check IV-DV linearity by adding the IV multiplied by the natural log of the IV (IV*lnIV) to the model before each regression. This works perfectly for 57 analyses/ participants. However, for 3, I am met with a red banner message " Table$setRow(): value 'r2t' is not atomic " at the box-tidwell stage. Also, when I remove the multiplied variable my slope estimates are exceptionally large; circa -21.55. I have attached a sample of my data including the 3 problematic cases and 3 functional cases. May I please ask for your insight surrounding this issue ?
Kindest wishes,
Lorna.
I am trying to run a series of binary logistic regressions to compute participant scores on a cognitive task where; 1) the IV is the point differences between"Self" and "Comp" options presented to each participant on 44 discrete trials and 2) the DV is whether the participant chose "Self" or "Comp" in response to each trial. I also run a box-tidwell test to check IV-DV linearity by adding the IV multiplied by the natural log of the IV (IV*lnIV) to the model before each regression. This works perfectly for 57 analyses/ participants. However, for 3, I am met with a red banner message " Table$setRow(): value 'r2t' is not atomic " at the box-tidwell stage. Also, when I remove the multiplied variable my slope estimates are exceptionally large; circa -21.55. I have attached a sample of my data including the 3 problematic cases and 3 functional cases. May I please ask for your insight surrounding this issue ?
Kindest wishes,
Lorna.