DOE-5: Fractional Factorial Designs, Confounding and Resolution Codes

[@DavidMonkII]

my goodness, that hand trick might help me remember on my IASSC test! thank you!

[@instituteofqualityandrelia7902]

Thanks! Yes, the hands trick is a very easy way to remember the resolution codes and extent of confounding!🙏

[@DavidMonkII]

Hey I passed my IASSC test. This video surely helped explain confounding and a simple way to remember!! thanks again! @@instituteofqualityandrelia7902

[@clipeus87]

Finally a clear video on fractional factorial design, thanks a lot!

[@instituteofqualityandrelia7902]

Glad it was helpful!

[@User123rag]

Excellent explanation! One gem of a collection!! Thanks a lot for your efforts.:)

[@uhemant1]

Thank you so much Ragha Laxmi!

[@ankitverma8820]

Wonderful explanation of confounding. Highly recommended for all.

[@instituteofqualityandrelia7902]

Glad it was helpful!

[@GioPalencia]

Very clear and direct to the point. Thank you so much!!!

[@instituteofqualityandrelia7902]

Thank you!🙏

[@parelectricengineeringllc1816]

Fantastic video. Great insight and excellent presentation

[@instituteofqualityandrelia7902]

Glad you enjoyed it!

[@philipaston8956]

Thank you for this clear explanation of confounding and aliasing!

[@instituteofqualityandrelia7902]

Glad it was helpful!

[@mdmahmudulhasanmiddya9632]

Unbelievable explanation.I am enjoying your vedio sir

[@instituteofqualityandrelia7902]

Thank you!

[@GyanVisharad]

Very information sir..with simple example..Thanks

[@instituteofqualityandrelia7902]

Glad to know!

[@calsimeth1588]

It makes so much more sense to me now! Thank you!

[@instituteofqualityandrelia7902]

I am glad to know! Thanks!

[@Yun-Sura]

Bless you for this explanation!

[@instituteofqualityandrelia7902]

Thank you! Appreciate your feedback!

[@stephenwright7274]

Thanks for making these videos. Clear and to the point!

[@instituteofqualityandrelia7902]

I am glad you are finding these useful!

[@item6931]

This is actually excellent.

[@instituteofqualityandrelia7902]

Thank you! I am glad you found it useful!

[@rvijayaragunathan]

superb.... well explanation sir ...

[@instituteofqualityandrelia7902]

Thanks and welcome

[@rohanmore1774]

Short and succinct !!! Thumbs up !!!

[@instituteofqualityandrelia7902]

Thanks a lot Rohan!

[@AhmadAli-rm3pi]

very nice lecture i solute you thank you sir

[@instituteofqualityandrelia7902]

So nice of you

[@francoisl7663]

Excellent, thank you.

[@instituteofqualityandrelia7902]

Welcome! Glad to know that you liked it!

[@mariaelenacerecedoarroyo5826]

great video!! Can you recommend some books in order to study the resolution designs?

[@instituteofqualityandrelia7902]

Thanks for your interest! There are many good books. Design of Experiments by Montgomery is excellent reference. George Box's book is very good. My book 'Six Sigma for Business Excellence' is also provides good information on DOE as well as other Six Sigma Tools. If you outside India, you can still buy e-book on Amazon. Good luck!

[@mdmahmudulhasanmiddya9632]

Very good explain sir

[@instituteofqualityandrelia7902]

Thanks and welcome

[@ljay3280]

How about a situation where you have 2 factors ,one with 7 levels and the other with 6 levels. Full factorial gives 42 runs. How do you run these with fractional factorial with less runs?

[@instituteofqualityandrelia7902]

Standard fractional factorial designs are available for two-level designs. In ur case, A has 7 and B has 6 levels. Fractional factorial design is not possible in this case as there are only two factors. If you have more factors with >2 levels, you need computer generated designs. For this you will need clear understanding of the degrees of freedom and your requirements.

[@andresalbertocorimayhuasil142]

Excellent video, it was really helpful. I got a question, what does a 3*2*2 factorial design means? I understood it refers to three factors and each of them with a specific level, i. e., the first factor has three levels, the second one has two levels, and the last one two factors as well. I thought it was a fractional factorial design. However, I tried to do this on STATISTICA and I could not find a design with 12 runs (according to 3*2*2). You would be so kind to help me to understand this. Thank you.

[@instituteofqualityandrelia7902]

This is not a usual convention. Experiments with different factor levels are called General Full Factorial Designs. You need to create a General Full Factorial Design in software as this is not a standard design.

[@vedantigandhi8209]

You have explained the concept of confounding very well. Thank you for the nice explanation. I just have 1 doubt regarding fractional factorial design. Why did we choose ABC=+1 and not ABC=-1? Thanks a lot! :)

[@instituteofqualityandrelia7902]

Thank you Vedanti. ABC=+1 is the principal fraction. In this principal fraction confounding becomes additive, in the sense, effect of A is actually A+BC, for example. ABC=-1 can also be used if desired. In this fraction effect of A is A-BC, for example. One should be aware of this while selecting the fraction.

[@marcouscangaolea1264]

thanks a lot. Pretty useful this video is

[@instituteofqualityandrelia7902]

I am glad you found it useful! 😊

[@TechAger]

Very nicely explained Sir... Thanks

[@instituteofqualityandrelia7902]

I am glad you found it useful! Appreciate your feedback!

[@mdmahmudulhasanmiddya9632]

Sir what is run?

[@instituteofqualityandrelia7902]

Run is a trial in an experiment

[@mdmahmudulhasanmiddya9632]

@@instituteofqualityandrelia7902 thank you sir

[@KNR85]

"Experimenters have found that higher order interactions tend to be small and can be ignored often" Is there any reference book or something that we can use to support this statement for the research projects.

[@instituteofqualityandrelia7902]

Hello Nirushan kathiresu, I am repeating the same answer that I mentioned in your other similar question. Basically, the fractional factorisl designs are based on one of the principles known as Sparsity of Effects. According to the sparsity-of-effects principle, it is unlikely that complex, higher-order effects exist, and that the most important effects are the lower-order effects. Thus, assign the experimental units so that main (1st-order) effects and the 2nd-order interaction effects can be investigated. You can get many references for this.

[@beshr1993]

that actually helped, thank u :)

[@instituteofqualityandrelia7902]

Thanks! Good to know!

[@asfarnasir]

What is the resolutuon of the discussed example?

[@MrJewZie]

3 or III

[@instituteofqualityandrelia7902]

@@MrJewZie Thanks for directly answering! Yes, the resolution of the design discussed in the vide is 3 or III as main effects are alliased with two-factor interactions!

[@instituteofqualityandrelia7902]

The resolution of the design discussed in the vide is 3 or III as main effects are alliased with two-factor interactions!

[@paolo7206]

Good video

[@instituteofqualityandrelia7902]

Thank you!😊

[@mdmahmudulhasanmiddya9632]

Thank you

[@instituteofqualityandrelia7902]

You're welcome

[@dr.akankshasom2309]

"Experimenters have found that higher order interactions of three and more factor tend to be small and can be ignored often"How did we reach this conclusion.

[@instituteofqualityandrelia7902]

Thanks for your question! Actually this statement is more of assumption based on 'The sparsity of effects principle'. This is more of an assumption that should be based on technical knowledge and judhement and needs to be validated based on analysis of experimental results. If the assumption is not valid, we should find low R-sq values, R-sq-pred values. The experimener is also taking some risk to save time and resources while reducing number of runs in fractional factorials. With best wishes..Hemant

[@paolo7206]

Nice video

[@instituteofqualityandrelia7902]

Thanks!