iSixSigma

Table of Percentiles

Six Sigma – iSixSigma Forums Old Forums General Table of Percentiles

Viewing 2 posts - 1 through 2 (of 2 total)
  • Author
    Posts
  • #38492

    Chang
    Participant

    Under Parametric Dist Analysis-Right Cent in Reliability/Survival
    ———————————-
    Table of Percentiles
                 Standard                         95.0% Normal CIPercent  Percentile      Error          Lower        Upper1            180.6306      117.3091     156.6731     1114.718050          1252.1065    127.7221   1210.2345    1302.164599          1789.7833   1173.933   1549.5698    2234.4704
    ——————————————
    Should I conclude that there is 1 (one) percent of the product will fail at 180.6306 hours or should it be actually at 156.6731 hours (worst case)?
    Is there a problem that the variation of the Percentile 99/Percentile 1 is very large (in this case = Raito of 9.9 to 1)?

    0
    #115224

    Bob J
    Participant

    Joshua,
    Based on the data and your analysis model the best estimate is what is listed under the “Standard Percentile” column…  Since this is an estimate based on a finite sample the CI gives you the probable bounds on the estimate… 
    Bottom line is to use the estimate…
    The change in the CI (as well as the Error) for the 99 percentile value could be associated with a poor fit between the model and the data…   You might want to check a few different distributions to see if you can get a better fit…
    Hope this helps….
    Best Regards,
    Bob J  

    0
Viewing 2 posts - 1 through 2 (of 2 total)

The forum ‘General’ is closed to new topics and replies.