Stat 1000: Assignment 8 Tips (Distance/Online Sections)
Published: Wed, 03/06/13
My tips for Assignment 8 are coming below, but first a couple of announcements.
Please note that my second two-day review seminar for
Stat 1000 will be on Saturday, Mar. 9 and Sunday, Mar. 10, in room 100 St. Paul's College,
from 9 am to 6 pm each day. This seminar will cover the lessons in Volume 2 of my book.
For more info about the seminar, and to register if you have not done so already, click this link:
Make sure you do: Tips on How to Do Well in Stat 1000
Did you read my Tips on what kind of calculator you should get?
Did you miss my Tips for Assignment 7?
If you are taking the course by Classroom Lecture (Sections A01, A02, etc.), there is no Assignment 8.
Tips for Assignment 8 (Distance/Online Sections D01, D02, D03, etc.)
Don't have my book? You can download a free sample containing Lesson 1 at my website here:
Question 1:
If
you are ever asked to decide if a particular situation is binomial or
not, remember, to be binomial, four conditions must be satisfied:
(i) There must be a fixed number of trials, n.(ii) Each trial must be independent.(iii) Each trial can have only two possible outcomes, success or failure, and the probability of success on each trial must have a constant value, p.(iv) X, the number of successes, is a discrete random variable whereX = 0, 1, 2, ... n.
Obviously, if there is not a clear value for n and/or p, it
cannot be binomial, but be careful. For example, if I randomly select
100 residents of Winnipeg and ask them, "Were you born in Canada?" that
is binomial. I am speaking to n=100 people. There is a percentage of
Winnipeggers who are born in Canada, p. I just don't know what that
percentage is at the moment. I would just say p=percentage of
Winnipeggers born in Canada. I have every right to say each trial is
independent (since we are selecting the people randomly), and,
obviously, there are only two outcomes (Yes, born in Canada; or No,
not).
However, if there is clearly not a fixed number of trials, n (say, if they decide to do something until a certain result happens), then there is no way it can be binomial. If the probability of success each trial keeps changing, it is not binomial. For example, if I am selecting cards from a deck without replacement then the probability I select a king changes every time depending on whether I have already selected a king before and on the fact that there are less and less cards left with each selection. The trials are not independent, the chance of success, p, is not fixed. Not binomial.
Question 2:
If you are solving a binomial problem,
and they ask you to compute a mean and/or standard deviation, read
carefully. Do they want the mean of X? or do they want the mean of
p-hat, the sample proportion? Be sure to study the sections about the
Distribution of X and the Distribution of p-hat in my Binomial
Distribution lesson (Lesson 6 in my new edition, Lesson 7 in older
editions). Take a look, especially, at question 10 of that lesson as a
good run through of these concepts. If your book does not have a
question 10 (why are you using such an old edition? upgrade!), you may
find that question 1 in my last lesson in the book may be helpful.
Questions that give you μ and σ
are undoubtedly dealing with bell curves. Make sure you have studied
my lesson on the Distribution of the Sample Mean (Lesson 7 in the new
edition, Lesson 6 in older editions). Always be very careful to note,
are they asking you for the probability of one individual value (X)?, or
are they asking you for the probability of the average or mean of n
values (x-bar, the sample mean)? If you are dealing with X, use the X
standardizing formula. If you are dealing with x-bar, use the x-bar
standardizing formula.
Also,
note that you can only do probabilities for X in these cases if you are
specifically told that X is normally distributed. Otherwise, there is
no X-bell curve, and the probability is unknown. However, thanks to the
Central Limit Theorem, we can always assume there is an x-bar bell
curve (the sample mean is normally distributed), as long as n is large.
Note, in Question 3, parts (c) and (d),
it appears you do not have enough info to solve the problem because
they have not given you a value for μ, the population mean. That is
because the actual value is irrelevant. If you ever come across a
problem like this, pretend that μ = 0, and then be careful to draw a
picture and shade the region they describe. Proceed to compute your
z-score, etc. Try it again, but this time using μ = 10 and making the
appropriate adjustments to your bell curve and shade the region. You
will discover you get the exact same z-score. This works no matter what
you use for μ. That's why the easiest choice is μ = 0.
Question 4:
Again, be sure to look at my question 10 in my Binomial
Distribution lesson to understand how to approach this question. By the
normal approximation, they mean use the p-hat bell curve. You
will need the standardizing formula for p-hat bell curves that I use in
question 10(c).
Question 5 is just more of the same stuff
from Lessons 6 and 7 in my book (different people get different
questions). Some of you get a question where they give you the mean and
variance and ask you what n and p are. Just use trial and error. It
is a multiple choice question. Use the given n and p from each choice
and sub it into the formulas to see if they produce the required mean
and variance. There is no need for algebra.
Question 6 uses the Sampling Distribution simulator at Rice Virtual Lab in Statistics.
Read Carefully. They first have you play around with it using a Normal
Distribution. But, the actual questions they ask want you to use a "Skewed Distribution".
Note that, in the top left corner, you are given the mean and standard
deviation of the population (μ and σ).
The various questions you are
being asked are about the distribution of the sample mean. Compare the
answers for the mean and standard deviation of x-bar that the applet is
giving you (next to the third histogram) with the theoretical values you
expect if you computed the mean and standard deviation of x-bar using
the formulas I give you in "The Distribution of the Sample Mean" lesson of my book. This applet is
illustrating the same concepts I discuss in Figures 1 and 2 in Lesson 7.
The
fact is, you can answer all the questions they ask without using the
applet. Just use your formulas for the mean and standard deviation for
x-bar (Mean(x-bar) = μ and Standard Deviation (x-bar) = σ divided by the square root of n) . All you need is the μ
and σ given
at the top left corner for the Skewed Distribution in the applet. You
should notice that μ = 8.08 and σ = 6.22 for the Skewed Distribution.
Then, just note the value for n in each part of the problem.