Statistical analysis on two variables

Good day

I would like to ask some advice.
I am working in a processing plant and I have to evaluate the outcome of two trials that we are planning. Both trials have the same dependent variable (I'll just call it process output) but I have to evaluate them separately. The challenge is that the two trials will overlap and I am not sure if I will be able to distinguish between the effect of each individual.
A bit of detail about the trials:
Trial 1 will test if replacement of one of the chemicals in the plant with a new chemical will increase the process output. The new chemical will be continuously used for 3 months, so the analysis will be a comparison of the process output data before the trial (old chemical) with the process output data of the 3 month trial period (new chemical). Trial 1 will commence about 3 weeks before trial 2.
Trial 2 will test if a change in the process will increase the process output. This will also be tested for 3 months but on an "on-off" basis - the process change will be switched on for 2 days and then switched off for 2 days for the duration of the 3 months. The "on" vs "off" periods can then be evaluated using paired t-tests.

What will be the best statistical analysis method for trial 1? When I do the data analysis for trial 1, should I exclude the "on" days of trial 2 from the data in order not to attribute the potential increase achieved by the process change to the chemical evaluation?

I know the ideal would be to separate the trials but unfortunately due to several constraints that won't be possible.


TS Contributor
I recommend that you consider using an individuals control chart. Baseline the control chart on the current process. If the process goes "out of control" when you make the changes described, you will know that the change had a significant effect on the process output.