ERRATA IN THE EXERCISES FOR
MATHEMATICAL STATISTICS WITH APPLICATIONS (7TH EDITION)
BY WACKERLY, MENDENHALL, AND SCHEAFFER

This is just a list of a few errata that I happen to have noticed.  It
is not intended to be a comprehensive list.


Exercise 2.79:

   P(B) < 1 should be P(B) > 0.


Exercise 2.80:

   Change "P(B) > 0" to "P(B) < 1".  

   Since A is a subset of B, the assumption that P(A) > 0 already
   implies that P(B) > 0, so it is redundant to assume that P(B) > 0.
   If, on the other hand, P(B) = 1, then A and B are necessarily
   independent, and the objective of the exercise is to show that A
   and B must be dependent except in the uninteresting case that one
   or the other of the events has probability zero or one.


Exercise 2.106:

   I would change this to ask "what is the smallest possible value for
   P(A)" or something similar.  "What is P(A)?" gives the impression
   that one should be able to divine the exact value of P(A) from the
   information given, which isn't true.


Exercise 3.38:

   This is the same as the triangular taste test example that I give
   in class, except that here it is not carefully worded.  The
   statement

      "Suppose that the two formulas are equally attractive." 

   should be changed to

      "Suppose that the two formulas are indistinguishable."

   Why?

   If the two formulas are equally attractive but distinguishable, say
   by everyone, then on the average half of all tasters would choose
   the glass containing formula B.  (The other half would choose one
   of the two glasses containing formula A.)  So, if the two formulas
   are "equally attractive" but distinguishable by everyone, then the
   distribution of Y would be binomial on n = 4 trials with success
   probability p = 1/2.

   If the two formulas are "equally attractive" and can be
   distinguished by some but not all tasters, then we would need more
   information to do the problem.

   With the change I suggest above ("indistinguishable" rather than
   "equally attractive" formulas), each taster is effectively picking
   at random among the three glasses and Y is binomial on n = 4 trials
   with success probability p = 1/3. This is what was actually
   intended and is the solution given in the solutions manual.


Exercise 5.85: 

   This exercise should probably be starred, since it refers to
   exercise 5.65, which is starred.

Exercise 5.94:

   The word "uncorrelated" should be deleted from this problem.
   Nothing in the problem changes if Y_1 and Y_2 are correlated.