1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Homework Help: Sums of Independent Random Variables

  1. Sep 3, 2009 #1
    1. The problem statement, all variables and given/known data

    Vicki owns two separtment stores. Delinquent charge accounts at store #1 show a normal distribution, with mean $90 and std. deviation $30, whereas at store #2, they show a normal distribution with mean $100 and std. deviation $50. If 10 delinquent accounts are selected randomly at store #1 and 15 at store #2, what is the probability that the average of the accounts selected at store #1 exceeds the average of those accounts at store #2?

    2. Relevant equations
    Let X1=N(90,30); where n=10, XBar1=N(90,30/10)=N(90/3). Let X2=N(100,50); where n=15, XBar2=N(100,50/15)=N(100,3.33)

    3. The attempt at a solution

    We need P(XBar1-XBar2>=1)=P((XBar1-XBar2+10)/Sqrt(-.33))>=(1+10)/Sqrt(-.33)
    =11/(Sqrt(-.333))???? This doesn't make any sense to me!!

    Any assistance would be very much appreciated!
  2. jcsd
  3. Sep 3, 2009 #2


    User Avatar
    Homework Helper

    hmmm... haven't done much probability for a while

    but based on your sampling I would try & come up with a probabilty distribution for the average of the measurement...

    so to start, say we pick X n times, then let Y be the average of the n samples, so the random variable Y is given by:
    Y = (X +...+ X)/n (sum of n times)

    What is the expected value & variance of Y, assuming each picked X is independent? Shouldn't be too hard to calculate... (though independence is an important assumtion....)

    If you have done this for both cases, then consider independent the random variables Y1 & Y2, how do you find the probabilty of Z = Y1-Y2 >0.

    Maybe something similar to first process, to find expectation & variance, then sums of guassian distributions are also gaussian...
  4. Sep 4, 2009 #3
    I see what you're saying, that the sample means should reflect the multiple of the sample size. However, don't I still end up with a denominator equal to the square root of minus 1/3?
  5. Sep 4, 2009 #4
    After several more attempts (and re-reading your analysis), I think I finally solved the problem. Thanks, Lanedance!
  6. Sep 5, 2009 #5


    User Avatar
    Homework Helper

    good one, glad it worked
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook