iSixSigma

Do I Need to Be Proficient in Math to Understand Design of Experiments?

Six Sigma – iSixSigma Forums General Forums Methodology Do I Need to Be Proficient in Math to Understand Design of Experiments?

Viewing 11 posts - 1 through 11 (of 11 total)
  • Author
    Posts
  • #55448

    Gustav
    Participant

    I use SAS JMP to run multivariate testing experiments (MVT, regression), it generates a test plan and performs regression analysis, defining significant effects. The most important thing is to be able to interpret and know the pitfalls.
    Mathematical calculations, as such, are not needed.

    I have read statistics books such as Montgomery Design of experiments and they are usually full of formulas.

    I still cannot understand how are they used in real world.

    Yea I understand that I need to know concepts, different types of designs, basics of statistics as p value, standard deviation, how to calculate proper sample size, etc.

    I speak about design of experiments mostly in web marketing and user experience fields.

    0
    #200035

    Sergey
    Participant

    Montgomery is serious reading! :)
    For first analysis try this

    0
    #200039

    Gustav
    Participant

    @ssobolev
    Thank you for your suggestion. I have already read a book by Ledolter “Experimental Design in Marketing and Service Operations” and I liked it.

    Sorry for offtop, are you from Russia?

    0
    #200040

    Chris Seider
    Participant
    #200042

    Sergey
    Participant

    Hello @gustavjung
    Yes, many greetings from Russia!

    As I see (correct me if I’m wrong) you are already overloaded with theoretical stuff and need more case studies to learn. In this case you may ask the same question at forums or blogs of practitioners of these particular topics. Or you are looking for any roadmaps and decision trees regarding DOE, regression and so on?

    0
    #200043

    SixSigmaGuy
    Participant

    Yes, you absolutely need to understand the mathematics to really leverage the tools. The statistical applications, like Minitab, can only be used correctly if the user understands what is going on behind the scenes. Once you understand the math, you’ll see how simple DOE really is. In fact, there are a lot better, and more accurate, ways to accomplish the same thing without using the statistical application. You may even stop doing DOE altogether, and stick to two variables.

    0
    #200044

    Chris Seider
    Participant

    It depends on what it means to “understand the mathematics”. I don’t think it’s very necessary to how the numbers are calculated–most users truly don’t. It’s important to understand ramifications for the design’s parameters and get a feel for how to get decisions made.

    My two cents

    0
    #200053

    Robert Butler
    Participant

    You need to understand the basic concept of design. If you don’t understand that then you are a sitting duck for any panacea peddler’s offerings which, if followed, will allow you to go wrong with great assurance.

    The basic principle of all designs is not that difficult to comprehend. Verbally it can be stated as follows: An experimental design utilizes all of the experiments in the assessment of the correlation between all of the X variables and the outcome of interest.

    The algebra used to do this is just the simple act of adding and subtracting and averaging the individual experimental results according to the + – pattern of the columns in the experimental design assigned to each of the independent variables of interest.

    The simplest example is the algebra for a two variable, two level factorial design. If you understand how that works then all you really need to know is that the same principle applies to all designs. See the attached document for details.

    0
    #200062

    WIlliam Howell
    Guest

    To be able to execute a DOE you need to be able to do 2 things: control the inputs and record the outputs. Can you control the inputs in the areas of interest you mentioned? If not, it is difficult to run a traditional DOE.

    As for reading, I like Quality Through Experimental Design, by Tom Barker. It is a fairly easy read, especially compared to Montgomery.

    What you really need is a mentor with significant experience in this area. He/she can guide you through the proper steps. In terms of importance, the further you go into the effort, the less critical things become. Mess up the analysis, you just rerun it. Mess up the design choices and you get to rerun the whole thing.

    0
    #200105

    Gustav
    Participant

    Hi @ssobolev!
    I am just looking for improvement of my knowledge and understanding of practical implementation of DOE in marketing / UX and especially its pitfalls.
    I just love and passionate about experimentations)
    And currently I am looking for a job that includes DOE in online marketing or user experience. I need to improve my competence.

    P.S. Currently I am in Moscow so if you would like you can connect with me by LinkedIn
    LinkedIn

    0
    #200117

    Sergey
    Participant

    Hi @gustavjung
    Great to hear, I’ve send a request to you via LinkedIn

    0
Viewing 11 posts - 1 through 11 (of 11 total)

You must be logged in to reply to this topic.