55023_en-US

liliconcom

five factor design matrix in yates standared order

Hi,

is it possible to make a five factor design matrix in yates stadanred order? I read on the disscusion that the run order right to left ordering is the yates standard order. I would like jmp to make a table that looks the the one below, only for 5 factors instead of 4.

Re: five factor design matrix in yates standared order

is it possible to make a five factor design matrix in yates stadanred order?

yes

I read on the disscusion that the run order right to left ordering is the yates standard order.

Randomizing the order is better, I don't care what Yates said.

I would like jmp to make a table that looks the the one below, only for 5 factors instead of 4

DOE/Classical Design/Full Factorial. Add 5 continuous 2 level factors, then use the sort option.

 

...But you wouldn't really do this experiment, right?

 

Re: five factor design matrix in yates standared order

I find this really irritating about JMP. For beginners it's really confusing when JMP decides to display the design in random (non-standard order). I've been looking for 10 minutes and still can't find the option that allows me to express the design in standard order. This feels like MS Excel or Word where Microsoft assumes it knows what you want

Re: five factor design matrix in yates standared order

@markntaylor I acknowledge your dismay.

 

Having said this though, the design and development of JMP tries to make it easy and efficient to practice, emphasis on the word 'practice', sound statistical methods in sometimes subtle as well as not so subtle ways. So since it is sound statistical practice to NOT actually run designed experiments in Yates order, JMP makes it more difficult for the experimenter to arrange the design matrix in that order.

 

I've been using JMP and counseling others since version 1. I can't tell you how many times I heard the novice JMP or statistical practitioner say something like, "Well I don't know which option to pick...much less what some of these mean...so I'll just go with the default figuring the designers put that there for a good reason." So making visualizing the design in Yates order a default could lead the uninformed to run the experiment in that order...and worse yet, the experimenter wouldn't know they are opening themselves up to all sorts of failure modes by doing so.

 

I hope this helps at least a bit?