Hi community,
Suppose I conducted a factorial 2^2 experiment with categorical factors and a continuous response. It produced three independent observations per run, so with 3 replicates I had 36 observations of response time (in seconds). The observations range from just a few seconds to as many as 800 seconds. However, there are 7 'observations' where the participants ran out of time. So, these observations are right-censored (although I don't have a precise value at which they were cut off). I don't want to exclude them, since they contain information. I tried assigning an arbitrary value of 1000 seconds, which exceeded any observed value. But for obvious reasons, this may be introducing bias.
After producing a table in the DOE platform in JMP Pro 15, the Model Specification dialog defaults to the Standard Least Squares personality with Effect Screening emphasis. There's no option in that dialog to identify a censor code. I'm not doing a survival analysis per se, but should I be taking advantage of JMP's ability to handle censored data? If so, what is the appropriate platform that I should be using to code those data as censored?
