Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie

Normal Distribution question

Options
  • 12-08-2015 4:13pm
    #1
    Pablo_
    Registered Users Posts: 188 ✭✭


    This is simple enough question but just wondering about data I am putting in to a norm dist curve. The answer might help me in the future with real applications:

    These are process times. I am looking at trends of 80 entries.
    Mean time = 35 mins
    Std Dev = 14.5 mins
    Minimum time = 15 mins...i.e. min possible time

    So the curve says (z score) I have 40% chance of getting 15 minutes process time...

    Do I then just simply negate all values under this? i.e. They are not realistic. Is everything above this figure a realistic probability?

    Does the fact that the Std Dev is so large, compared to minimum and Mean, make this curve unreliable?

    Thanks in advance


Comments

  • Options
    #2
    MathsManiac
    Registered Users Posts: 1,595 ✭✭✭MathsManiac


    What makes you believe that the data would be normally distributed in the first place? (It seems unlikely given the information supplied.)

    Have you done a normality test on it? Or even graphed it to see what it looks like?

    Do you want to post the data up so we can have a look?


  • Options
    #3
    Pablo_
    Registered Users Posts: 188 ✭✭Pablo_


    Thanks, Here is the graph. This one is slightly diiferent, minimum possible time is 18mins. Mean is 44 and Std dev is 16.7.

    Can I incorporate the minimum into this graph. Therefore I could have a tool for predicting, or simply evaluating, how often we reach a certain process time?

    Then my further question is can I use z scores for the above said. i.e I have 34% chance of 27.3 minutes (the mean minus 1 std dev) ?? Or is this far too simplistic or just plain misuse of this tool?


  • Options
    #4
    MathsManiac
    Registered Users Posts: 1,595 ✭✭✭MathsManiac


    But that is clearly NOT the graph of the data.

    That is a graph of a theoretical normal distribution with mean 44 and standard deviation 16.7. It is not the graph of an actual data set.

    Without looking at actual real data generated by whatever process is involved, you cannot hope to determine whether the it produces normally distributed data or not.


  • Options
    #5
    MathsManiac
    Registered Users Posts: 1,595 ✭✭✭MathsManiac


    I should also add (going back to your original problem) that it is not possible for normally distributed data to have a mean of 35 mins, a standard deviation of 14.5 mins, and a minimum possible value of 15 mins.

    That is why I (or someone else here) would need to see the actual data in order to be able to suggest a suitable alternative model for what you are trying to achieve.


Advertisement