MATH 251, Probability and Statistics I, Fall 2007 Mon. Oct. 29,  Day 28 After class

HW Day 28  Read  5.2. For 5.2 homework, memorize the contents of the boxes, pp. 361-62. Please  skim through Weibull dist. (pp.367-8). Confidence intervals (6.1) tend to be hard to grasp conceptually; I urge you to read at least pp. 382-391  in preparation for the next class.
Hand in:  Prep for 6.1: You took 4 Numbers (random sample)  from the Birkenstock box in class and wrote them down. (Box is outside my door if you missed this)  For HW, Find mean xbar.  Find xbar + .841.  This is your interval estimate of the unknown mean of the box's population.  ("margin of error" is .841) (Return your numbers afterward.)
Next class, you will add your values to the circulating list, and graph your interval on the transparency circulating.
     If xbar = 8.0       7.159|_____________8.0_____________|8.841

Sec. 5.2 

5.29 axle diameter (mean, s.d.)
5.31 abc axle diameter (cf. X and Xbar)
5.35  mean number of moths in a trap
5.41  airplane weight (They suggest changing it into a "mean" problem, but you can do it as a "sum" problem using the algebra of means and variances and noticing that if Xbar is normal, so is the sum of the  Xi's)
5.37, 5.39 Sheila's gestational diabetes
5.40 carpet flaws 
[5.39 and 5.40 both involve the computations needed for "significance testing".  The remaining step is this:  If our results are very unlikely, they call our assumption (the value of the mean) into question.]
Postpone the rest:
Linear combinations of independent normal r.v.'s (as p.365-6) 
5.51 X+Y couples.
5.49 sum of 4 blocks
5.44 hole and shaft (Y-X)
5.64 screw-on caps break
5.46 difference of means--trustworthiness
5.47 difference of means--general

Read, discuss 
  Sec.5.2
5.33 unbiased

5.65 p. 377 dye pH (Read for intro to "control limits", compare with 5.29 and 5.31.   This is "Quality Control")




Postpone :
5.45 diff. of means

Optional 
Sec.5.2 (more practice) 
5.32 lightning strikes.  (Use algebra of means and variances for part a)

5.61 p. 377 car passengers.

Chaper 17, on your disk or downloadable, is all about issues like # 5.29, 31, 65--"Quality Control")


5.48 market gain: if results build on one another, usual mean may not be appropriate.

Homework questions?  Day 27
Prep. for 6.1:
Take 4 Numbers (random sample) from the Birkenstock box:  Write them down.  For HW, Find mean xbar.  Find xbar + .841.  This is your interval estimate of the unknown mean of the box's population.  ("margin of error" is .841) (Return your numbers afterward.)
Next class, you will add your values to the circulating list, and graph your interval on the transparency circulating.
     If xbar = 8.0       7.159|_____________8.0_____________|8.841

Mean and variance quizzes back: Mostly quite good.  Main problems:
Pesky added constant.  Another way to remember what  the variance of a +bX is.
    Let Y = a, with probability 1 (no matter what happens, a is the result).
  y |  a          Mean of Y = a.   Variance of Y = 0
  p |  1          Then sigma2(a + bX) = sigma2(Y + bX) = sigma2(Y) + sigma2(bX) = 0 +  sigma2(bX) = 0+ b2sigma2(X)
Variance of (aX +b Y
) = a2Variance X + b2Variance Y
Variance of X - Y = Variance of (X +(-1)Y) = Variance of (X + Y), because (-1)2 = 1.
When do you need independence?  Only for variance of sum of 2 random variables!  NOT for mean of sum, NOT for variance of a+bX.

Not a problem but you should notice:
     Mean and s.d. of a distribution (population)
are actually completely parallel to our computations for data, viewed correctly:
          X    1   2   4
          p   .2  .2  .6
# out of 10    2   2   6 Consider a population of 10, with these proportions.
 Mean:  (1+1+2+2+4+4+4+4+4+4)/10 = (1·2 + 2·2 + 4·6)/10 = 1·.2 + 2·.2 + 4·.6 = 3
 Population variance (divide by n)     [Sample variance (divide by n-1)--the one we learned in Chapter 1]
 = [(1-3)2+(1-3)2+ (2-3)2+(2-3)2 +(4-3)2+(4-3)2+(4-3)2+(4-3)2+(4-3)2+(4-3)2+(4-3)2]/10
= [(1-3)2·2 + (2-3)2·2 + (4-3)2·6]/10 = [(1-3)2·.2 + (2-3)2·.2 + (4-3)2·.6]
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
Sec. 5.2 (Central Limit Th. and friends) Day 27

Got through most but not all of Central Lim. Th.stuff on Day 27. Pick up there Wed.

"If X is normal then Xbar is normal"  is a special case of
linear combination of independent normal random variables is normal, (p.365-6) with the mean and s.d. you can calculate using the algebra of means and variances.
If X and Y are independent normal random variables,  then aX + bY is normal also.  Use "algebra" to find the mean and standard deviation.

Example:  Ann runs the mile in X minutes, where X is N(8, 2)
                Betty runs the mile in Y minutes, where Y is N(9, 3)
    Ann and Betty compete against each other a lot.  What proportion of the time does Betty beat Ann?  (Assume independence.)
       Want Prob. that Ann's time is longer than Betty's.  P(X > Y) = P(X - Y > 0)
                   Need distribution of X-Y.  It's Normal.  Mean =  µX - µY = 8 - 9 = -1
                              Variance = sigma2X  + (-1)2sigma2Y  = 22 + 32 = 4+9 = 13
                               S.d. = sqrt( 13) = 3.6   
   Back to P (X - Y > 0).  Standardize by subtracting the mean (-1) and dividing by the s.d. (3.6)
   
P(X - Y > 0) = P(Z  > [(0-(-1))/3.6 ] ) = P(Z  > .28 ) = 1 -  .6103 = .3897, almost 40% of the time.

   Suppose each records her time for 4 days.  What's the probability that Betty's 4-day average  is faster than Ann's ?
     Want P( Xbar >Ybar ) = P( Xbar - Ybar > 0).   Need distribution of Xbar -Ybar.
                         The s.d. of Xbar is 2/2 =1  (since sqrt(4) = 2).  So Xbar is N(8, 1)
                          The s.d. of Ybar is 3/2 =1.5  (since sqrt(4) = 2).  So Ybar is N(9, 1.5)
                    Xbar -Ybar is Normal. Mean = 8 - 9 = -1.  Variance = 12 + (-1)2(1.5)2 = 1+2.25 = 3.25
                              S.d = sqrt(3.25) = 1.8
      P( Xbar - Ybar > 0) = P(Z  > [(0-(-1))/1.8 ] ) = P(Z  > .56 ) = 1 -  .7123 = .2877,  almost 30% of the time.

Statistics application:  Experiment: Two treatments; compare the mean results--what difference do we see?  If there really is an underlying difference (Ann is faster than Betty on average), are we sure of seeing it?  If there's no underlying difference, what's the chance of seeing something that looks like a difference?


Sievers home  Math251-Fall07/Day2s28.htm  1:30pm  10/29/07
This page belongs to Sally Sievers who is solely responsible for its content. Please see our statement of responsibility.