iSixSigma

Standard deviation

Six Sigma – iSixSigma Forums General Forums Methodology Standard deviation

Viewing 8 posts - 1 through 8 (of 8 total)
  • Author
    Posts
  • #53386

    Sridhar
    Member

    Hi,

    Can anyone let me know the practical significance of std. deviation in an industry?

    Thanks
    Sridhar

    0
    #189917

    Hody
    Member

    Since the tails of a normal distribution run to infinity in both positive and negative directions, you have to have a way decribe the width (and shape) of the population. “Std Dev” is the distance from the mean to the point of inflection along the curve (i.e. changing from curving down to curving up). It is easily calculated on a caculator and in Excel. It is also the sq root of “variance” which is useful itself since total variance is also the arithmetic sum of variances from all factors.

    Other ways to decribe the width of a population: “Range” is highly dependent on sample size. “Inter-Quartile Range” is less dependent on sample size, but most calculators and Excel are a little clumsy to calculate it. Neither of these are as useful as std dev unless you have non-normal distributions.

    Was that helpful?

    0
    #189920

    Sowmyan
    Member

    Practical use of standard deviation:
    To measure variation
    To assess best process capability that can be achieved by controlling mean shift.
    To assess extent of improvement required in controlling variation of ‘as is’ process to meet customer requirement.

    0
    #190153

    eduard o
    Participant

    would not be easier to use averaging and add some sensible tolerance into it

    0
    #190154

    Cone
    Participant

    Take a stats class and come back with a reasonable question.

    0
    #190232

    MBBinWI
    Participant

    what would you consider “sensible” and why? If you can answer that question without statistics that gives you a way to ensure acceptable quality with least cost, I’ll worship at your alter.

    Otherwise, you’re just another ignorant hack.

    0
    #190254

    purrhapz
    Participant

    Standard Deviation is important because customers rarely feel averages, they feel the variation. For example, it is possible for a service to take 30 days “on average”. If I had one person who received their service within 1 day and another who received it within 59 days, the average time is 30 days but these customers each had quite a different experience. Typically, there are quite a few occurrences in which the actual results for an individual data point are not equal to the average. The standard deviation explains basically how far away (on average) the individual data points are from that average number. This is important to businesses because you need to understand “how far off” we are from that average number to be able to identify the best solutions to address issues. Customers want predictability. They want to know that something will take on average 30 days plus or minus a day or so… not 30 days plus or minus 29 days.
    Hope that helps.

    0
    #190266

    MBBinWI
    Participant

    “customers rarely feel averages, they feel the variation” is becoming trite. Yes, variation is problematic, but so can the average be problematic. I was once a product manager on a product where a customer observable attribute we surveyed was perfectly acceptable on one version of the product, so we didn’t change it. The next version of the product it was totally unacceptable. Why, because the customers expected improvement, even though they didn’t feel there was anything unacceptable in the previous product. Thus, the mean was no longer acceptable.

    0
Viewing 8 posts - 1 through 8 (of 8 total)

You must be logged in to reply to this topic.