topic Re: Classical Screening design vs Definitive Screening Design in Discussions

Classical Screening design vs Definitive Screening Design

albiruni81 — Tue, 27 Mar 2018 02:40:13 GMT

Hi All,

Anyone can give me a summary of difference between a classical screening design and DSD? What are the advantages and disadvantages between this two methods

Rgrds

Irfan

Re: Classical Screening design vs Definitive Screening Design

txnelson — Tue, 27 Mar 2018 03:07:51 GMT

Here are a couple of references:

Jones, B. and Nachtsheim, C. (2011a). Efficient Designs With Minimal Aliasing. Technometrics, 53, 1, 62 – 71.

Jones, B. and Nachtsheim, C. (2011b). A Class of Three-Level Designs for Definitive Screening in the Presence of Second-Order Effects. Journal of Quality Technology, 43, 1, 1 – 15.

Also, here is a link to Brad Jones' 2015 Discovery Presentation

A New Analytical Approach for Definitive Screening Designs

Re: Classical Screening design vs Definitive Screening Design

cwillden — Tue, 27 Mar 2018 15:14:11 GMT

DSDs are really somewhere between a screening and an RSM. The general idea is you can screen a large number of factors and still fit a decent response surface to the active effects in one very efficient design. The drawback is that they are far less trivial to analyze. Over the last several years, a lot of work has been done on how to best analyze DSDs, and JMP has a dedicated analysis platform for DSDs now. It would be worthwhile to study up on that methodology using the links @txnelson provided, as well as Brad Jones' talk at the most recent JMP Discovery.

Classical screening focuses on designs with 2-level factors to determine active main effects, usually 2-factor interactions, and sometimes higher order interactions. You can add center points to detect curvature, but you cannot model the resulting curvature because the source of the curvature is ambiguous with multiple continuous factors. The benefit is they are very easy to analyze.

Re: Classical Screening design vs Definitive Screening Design

albiruni81 — Fri, 30 Mar 2018 02:36:24 GMT

Hello Cameron,

Thank you for you reply

When yu are saying that JMP has a dedicated platform for DSD what do you mean by that?Where can i get this option from?

Rgrds

Irfan

Re: Classical Screening design vs Definitive Screening Design

albiruni81 — Fri, 30 Mar 2018 02:36:56 GMT

Thanks Jim

Re: Classical Screening design vs Definitive Screening Design

txnelson — Fri, 30 Mar 2018 02:42:05 GMT

DSD is available from the pull down menus

DOE==>Definative Screening==>Definitive Screening Designs

BTW, Brad Jones, the co inventor of DSD, is a long time employee with JMP

Re: Classical Screening design vs Definitive Screening Design

albiruni81 — Fri, 30 Mar 2018 02:59:27 GMT

Hello Jim,

Is this how we analyze the DSD table?

Rgrds

Irfan

Re: Classical Screening design vs Definitive Screening Design

txnelson — Fri, 30 Mar 2018 03:06:51 GMT

If you already have a defined DSD, and you are just looking to analyze the data, the answer is yes. If you need to setup a DSD, then you would select Definitive Screening Design.

Full documentation on this is covered in the DOE Guide. See:

Help==>Books==>Design of Experiments Guide

Re: Classical Screening design vs Definitive Screening Design

albiruni81 — Fri, 30 Mar 2018 04:19:08 GMT

Thanks jim, does it mean that DSD is the way to go as opposed to the classical screening design?

Re: Classical Screening design vs Definitive Screening Design

txnelson — Fri, 30 Mar 2018 04:32:58 GMT

yes........

Re: Classical Screening design vs Definitive Screening Design

albiruni81 — Fri, 30 Mar 2018 09:09:58 GMT

Hi Jim,

When i do this power analysis, any reason that the model comparison is showing that the fractional factorial is superior than the DSD

Rgrds

Irfan

Re: Classical Screening design vs Definitive Screening Design

Mark_Bailey — Fri, 30 Mar 2018 11:44:09 GMT

You lose power and estimation efficiency due to the center levels for each factor.They have no leverage in estimating the parameters for the main effects. You gain the ability in a true screening situation to (possibly) avoid additional runs (augmentation) in order to model higher order terms.

Search the JMP Community Blogs for "definitive screening designs" for more details and background information. (Easier for you than 'back and forth' with a lot of questions.)