4 Variable names and recommendations

Once data is imputed and corrected it is stored in the Sepages data base, for this reasons variable names are standardised accross datasets and recommendations are the same. Figure 0.1 recaps the different steps of the data pre-processing.

4.1 Variable names

  • val_crude is the raw data received from the lab
  • val_cat is the categorised data for those that have more than 30% censored data5
  • val_i is the crude data after imputation for data below LOD and between LOD and LOQ
  • val_i_cor is the imputed data after correction for protocol variables

4.2 Analysis recommendations

All analyses should be performed on the corrected variables, ie with suffix _i_cor. And sensitivity analyses should be peformed on the imputed values, ie with suffix _i.

4.3 Concluding remarks

For any further information feel free to contact M. Rolland at .

  1. 1 = below LOD; 2 = between LOD and LOQ; 3 = greater than LOQ↩