Is the power analysis for my full factorial is too low?

[#1]

Hello everyone,

I did a screening design and the two resulting main factors are temperature (continuous) and solvent (categorical).

I would like to do a full factorial with these two factors at 3 levels.

I decided to opt for a full factorial since the total number of experiments is very manageable (9), but when I check the power

analysis, I feel that it might be too low? I was reading that you want the power to be at least 0.8.

I tried adding some center points but it doesn't get me close to 0.8 --- and despite replicates reaching that threshold, 18 experiments is too many.

Advice is appreciated!

Thank you!

[#2 · Accepted solution]

The power analysis option is a tool that you can use to calculate expected power.  The power is based on the signal to noise ratio which is based on the ratio of anticipated coefficient to anticipated RMSE.  By default the ratio is 1:1 which is very low, hence low power.  Put in larger values for anticipated coefficients to see the effect of a larger S/N ratio.

Here is a detailed explanation of how to use the tool:

https://www.jmp.com/support/help/en/16.1/#page/jmp/power-analysis.shtml?os=win&source=application#ww... 

There have been a number of excellent discussions on this topic - I suggest you do a search for power analysis.  Here is one such discussion:

https://community.jmp.com/t5/Discussions/Should-I-consider-power-analysis-in-DOE/m-p/501063