39767_en-US

liliconcom

working with covariates/time drift

Hi,

 

I have 3 DOE experiments which are exactly the same, but show different magnitudes of response due to the machine that takes the measurements varies over time. The relative changes in the measurments are the same but are 10x and 100x of experiment one

 

How can I best capture this factor in my model?

Re: working with covariates/time drift

I would stack the data from three experiments and add a Time as a covariate (maybe its interactions with other factors as well if necessary) in my regression analysis. Hope this covariate factor would account for the systematic change in response.