Date: Tue, 16 Dec 2003 09:38:11 -0500 (EST)
From: Adrian German <dgerman@cs.indiana.edu>
To: A113 Fall 2003 Distr. List <dgerman@indiana.edu>
Subject: A113 Final Exam


Dear A113 Friends,

Just a reminder that there is an exam today at 12:30 in TH A201.

This is the final exam, open book, open notes, open resources.

The final grades will be turned to the Registrar on Fri. Between
now and then you can come and see me if you have any questions. The
available times are listed here:

http://burrowww.cs.indiana.edu:11400/cgi-bin/fall2003finalappts/schedule

Go ahead and sign up for any time you want if you want to come and
talk about your grades. You will receive an e-mail confirmation with
the time you chose and the place of the appointment. Please let me know
if you experience any problems with the script.

... Adrian

I include some for your evaluation:

• This for a fact: 43.3% of statistics are meaningless.
• Everyone is entitled to my opinion.
• Hard work has a future payoff, laziness pays off now.
• Warning: dates in calendar are closer than they appear.
• Consciousness: that annoying time between naps.
• I took an IQ test and the results were negative.
• Out of my mind. Back in five minutes.
• Energizer bunny arrested. Charged with battery.
• Support your local lawyer: send your kid to medical school.
• My life has a superb cast, but I can't figure out the plot.
• You're just jealous because the voices only talk to me.
• Age is a very high price to pay for maturity.
• I doubt, therefore I might be.
• Procrastination is the art of keeping up with yesterday.
• If you try to fail, and succeed, which have you done?
• I am not a complete idiot, some parts are missing.
• I used to have a handle on life, but it broke.
• No sense being pessimistic. It wouldn't work anyway.
• Ever stop to think, and forget to start again?

Date: Thu, 11 Dec 2003 10:19:48 -0500 (EST)
From: Adrian German <dgerman@cs.indiana.edu>
To: A113 Fall 2003 Distr. List <dgerman@indiana.edu>
Subject: A113 Final Week Update


Dear A113 Friends,

The class is almost over. Today we have a final review for the exam
next week and we will collect course and instructor evaluations. The
second practical exam is tomorrow in lab. Sample topics were posted
yesterday on the class web site. I will also post a set of sample final
exam questions later today and we can discuss those in class. As with
the previous exams the practical and the final are open notes and open
book use whatever you want to use except: another person for the final
and you're not supposed to bring your spreadsheet already completed to
the practical exam on Friday.

Also, drop boxes in OnCourse closed themselves up last night so I made
a change such that they stay open until Dec 16. Let me know if you have
any questions and all the best as you are approaching the finals week.
I am sure you will do very well in all your finals in all your classes.

... Adrian

1. This question is asked on page 411 second paragraph at the top. An answer is provided on the next page (412 at the top). The next few pages (412-413) then emphasize the finding first stated on page 412. This leads naturally into the next question which is about the experiment on p. 415.

2. The reason behind the usefulness is stated on p. 413 at the bottom. There are two possibilities, stated at the top of the next page.

3. Page 417. There's also some discussion (ref. Fosback) on p. 427. But this question is mostly about pp. 419-424.

4. P. 422. Also pp. 423-424.

5. See pp. 424-426. Also p. 427 and pp. 428-439. P. 438 bottom.

6. Pp. 439-440.

7. See pp. 446, 449, 451. Then see pp. 454-458.

Summary of Chapter 10: Analysis of Variance

Analysis of variance (ANOVA) provides a test to determine whether to
accept or reject the hypothesis that all of the group means are equal. 
(Think of the two groups in the salary discrimination problem at the end
of Chapter 8. Think also of the problem at the end of Chapter 4: 23 on 
page 167 "if coockoo eggs differed in length between hosts that would
lend some weight/credibility to that hypothesis". This problem appears 
as 22 on page 406). 

Assumptions for the means model (that we use for ANOVA) are (see p. 369) 
similar to those used for regression analysis. 

First example starts on page 369. We use Hotel.xls to investigate if 
hotel prices in New York are different from those in other cities. (The
second example in this chapter is using Cola.xls while the two problems
that you can choose from for Homework Six, from this chapter, use the 
Heart.xls and Eggs.xls spreadsheets respectively). 

Steps in the first example: 

1. Make a copy of the file (p. 370). 
2. Establish the two hypotheses (p. 371). 
3. Plot the prices by category (pp. 371-373). Why? 
4. Prepare for the ANalysis Of VAriance (using the Analysis ToolPak) by 

unstacking the prices based on the city variable (use StatPlus, p. 373). 
5. Perform the analysis of variance (Anova: Single Factor, pp. 374-375). 
Pp. 376-377 try to explain the rationale of the method (skim it). The 
bottom line (p. 377, lower half) is that the MS column is important (why?) 

and that the F and P-value columns are the deciding factor in our analysis). 
The section ends showing how one can calculate the coefficient of 
determination from the values produced by ANOVA. 

6. What's the conclusion of the analysis so far (p. 377, last but one 

paragraph)? The answer to this question determines the next step taken
(pp. 378-379): create a matrix of paired differences. (Why?) Use StatPlus.  

7. The problem now is adjusting the p-values for multiple comparisons. So
a correction factor is used (which one?) and discussed (when should we use
it?) So pp. 380-381 reach a conclusion late the section. 

8. Knowledge from Chapter 4 is used to provide an additional perspective
(pp. 381-382). Top of section (p. 381) explains the potential benefit of 
this approach (medians, outliers). Overall we are accumulating evidence. 
In final analysis what is the conclusion of your experiment? 

Some exercises revisit this problem. 

9. Remember that we take an engineering, big-picture approach: how can 
we use the theories developed (and implemented in Excel). So we switch to
performing the analysis of variance using regression modeling (pp. 382-283).
The first step is to create indicator variables (use StatPlus). 

10. The second is to fit the effects model to the hotel pricing data and
we are done (p. 384-386). 

The rest of the chapter is not used in the lab assignment. 

However tomorrow in class we will cover the rest and start on the last
two chapters in the book. 

1. See page 297 at the top, tutorial, QuizSite exercises.
2. Starts on page 298. See page 303, page 308 (F-ratio and at the bottom). Ends on page 309.
3. See page 310 at the top. Bottom upper half paragraph. See also pp. 312, 313, 313, and 315.
4. See page 318.
5. See page 319 at the bottom (really starts on next page). Also pp. 322-323.
6. See page 322, middle is a start. Also page 325 (comparison with earlier correlation, stronger).

Here's some help with Lab Assignment Six (Chapter 9) as well:

1. See page 336.
2. See page 340 bottom, page 341 at the top.
3. See page 341, middle paragraph at the top.
4. See page 342 (top half). Pages 341-342.
5. See page 342. Page 349 contains a summary.
6. Page 350 describes the plan. See pp. 351-357, then pp. 358-362.
7. See page 357. Page 362: actual situation (court).

Here are some suggestions for Homework Six:

1. Chapter 8, Problem 20 (p. 331): very neat problem.
2. Chapter 8, Problem 15 (p. 328): on the influence of outliers.
3. Chapter 8, Problem 19 (p. 330): rather involved. Not my first choice.
4. Chapter 9, Problem 14 (p. 366): on validity. Not a first choice either.
5. Chapter 9, Problem 11 (p. 365): neat problem. Also about validity.
6. Chapter 10, Problem 22 (p. 406): seen before.
7. Chapter 10, Problem 19 (p. 405): what is ANOVA?

Personally I find at least one problem in each of the three chapters attractive.

Just for your information here are two lectures from last year:

• Correlation
• Regression

Here also is an example midterm exam from last year.

New QuizSite exercises have been posted.

Practical exams acknowledged (grading forthcoming).

Date: Mon, 17 Nov 2003 17:16:42 -0500 (EST)
From: Adrian German <dgerman@cs.indiana.edu>
To: A113 Fall 2003 Distr. List <dgerman@indiana.edu>
Subject: help session tonight/practical details 


Dear A113 Friends,

If you check the practical one link on the class notes web page
at http://www.cs.indiana.edu/classes/a113/fall2003/classNotes.html
you will see that there will be 12 subjects for the practical exam
on Friday. These are the actual exam problems, and each one of you
will receive a random problem from the set listed there.

There's a help session tonight (7:15pm in Lindley 102). Come if you
need help with any of the listed problems. As a reminder the exam is
open book and you can talk to anyone but only occasionally and without
disturbing the class. You have to build the worksheet from scratch. Any
needed data files will be provided by the AI who will make sure that you
don't access your CFS or your e-mail account and/or the network scratch
folder. When you are done you will turn in the problem to Oncourse and
the AI will check that your submission was correct.

The midterm questions won't be too many, and they will be based on

a) the QuizSite exercises, and

b) the book (big picture questions).

I will try to provide answers to all (or at least the most significant
of) the lab assignment questions by Tuesday night and post them on the
website so you can go into the exam without any uncertainties. Grades
should also be posted by then too.

... Adrian

Here's the message of Sunday:

Date: Sun, 16 Nov 2003 12:55:56 -0500 (EST)
From: Adrian German <dgerman@cs.indiana.edu>
To: A113 Fall 2003 Distr. List <dgerman@indiana.edu>
Subject: A113 Practical One and Midterm Exams


Dear A113 Friends,

I posted some info on the exams: for the midterm expect questions
like those in QuizSite and also big picture questions from the book
(like: where/when/why would you use a pivot table? what's chi square
test/statistic and when/why you might use it?) Just the basic idea.
I will post more info for the midterm soon.

For the practical you will be given one of the projects in the book
(already explained in the book). So if you never read that part you can
read it during the exam and complete the steps and the experiment. There
are about 20 such projects in the book just be familiar with them. You
can't bring the spreadsheet from home, but the exam is open book, so all
you have to do is be able to use the book. I will refine the info posted
already and make it even more precise.

There will be a help session tomorrow at 7:15pm in Lindley 102.

The Practical Exam won't be group work but you can talk to each
other (not much, and not too loud) occasionally. It is important
that those that work by themselves not be disturbed.

The midterm exam is on Thu in class. The practical is on Friday, in labs.

Both exams are open book. On the written exam you can't exchange
information with anyone except the instructor. On the practical exam
you can (occasionally) talk to your neighbours, if you need to. But just
occasionally, and without disturbing the class at all.

Let me know if you have any questions and/or concerns.

There will be reviews in class on Tue and Wed for the two exams.

Hope your weekend is coming along fine.

... Adrian

• solutions to all the exercises in the book
• have been posted on the class/lecture notes web page
• in Word format.

What remains to be done now is

• for us to go through the material in the book in a meaningful manner
• so that you can put together articulate and convicing reports
• for lab and homework assignments.

Information about the exams is forthcoming.

The message of Tuesday:

Date: Tue, 11 Nov 2003 16:23:41 -0500 (EST)
From: Adrian German <dgerman@cs.indiana.edu>

To: A113 Fall 2003 Distr. List <dgerman@indiana.edu>
Subject: a113 min paper (tue nov 11)


Dear A113 Friends,

I looked at the minute papers collected today. The probability
distribution for the sum of two dice is a tent: two straight lines
from (2, 1/36) to (7, 6/36) and from (7, 6/36) to (12, 1/36) where
the first value is the independent variable (the outcome) and the
second is the chance of that outcome. Perhaps you can try this in
lab tomorrow. We'll look at it for sure on Thu in class.

So the distribution is not normal! Don't assume every distribution
will be normal or almost normal. That need not be the case. Here's
a note from last year with a table that gives away the answer:

  http://www.cs.indiana.edu/classes/a113/fall2002/notes/Five.html

                                    *

Also please take a look at problem 23/167. What does the solution
posted have to say about part (c)? The answer is: nothing. What then
are you going to do? First: definitely don't panic. That's the first
step. Second, think instead. Third, carefully also look for sources of
help. Fortunately the question is asked again (this time as question no.
22 on page 406) in the chapter about the analysis of variance. At that
point your qualitative analysis from Ch. 4 (based on simple descriptive
statistics) can be quantified a bit more, and so an answer is provided
with the solution on page 406.

                                    *

For the practical next week you need to be able to just use Excel
and/or StatPlus to calculate and plot things: descriptive statistics,
distributions, check for normality (in two ways, as we did in class
today), and perform simple statistical inference. I might come to
labs tomorrow, but I can't promise. I will post more information
about the exams on the web site in a day or two (under class notes).

(If the practical will test your basic knowledge of Excel/StatPlus
the written midterm will test your basic understanding of your reading
assignments. Hopefully the questions posted as lab assignments help with
preparation for that.)

...Adrian

Date: Mon, 10 Nov 2003 21:57:54 -0500 (EST)
From: Adrian German <dgerman@cs.indiana.edu>
To: A113 Fall 2003 Distr. List <dgerman@indiana.edu>
Subject: A113 Update (Week 3 Fall 2003)


Dear A113 Friends,

The web site for the class is now essentially complete: I have
organized the remaining of this semester around a total of 8 lab
assignments and 7 homework assignments (two of which are behind us
already). All homework assignments will be addressed directly in class,
during lectures. All lab assignments will be addressed directly in labs.
You need to turn in a total of 5 homework assignments (choose 5 of the 7)
the remaining 2 being your bonus, and 7 lab assignments (out of eight).

The suggested due dates are listed on the What's Due page. The main
focus is to go through the materials in the book and build a working
understanding of the type of problems you are supposed to be able to
solve. Solving any of the homework assignments will then be as simple
as summarizing the corresponding lecture or set of lectures (which will
be aimed at providing general solutions for the kinds of problems listed
in the homework assignments) and doing a lab assignmnent will require
that you peruse the book and work through the tutorials.

I spoke to most of you in labs on Fri. At this point you should be clear
on the fact that we work with Excel under Windows, that Excel works with
cells, that are named and used for formulas, that contain absolute or
relative names of cells and/or names of functions (that have parameters).

For now we use Excel to analyze sequences of numbers (data), which we
currently summarize using averages (a measure of central tendency) and
standard deviations (a measure of dispersion). We also should have a
good idea of what a distribution is and we plot them as histograms to
further summarize (graphically) our sequences of numbers (data).

There are discrete distributions and continuous distributions. A certain
distribution (normal) is (a) ubiquitous and (b) very useful and uses two
parameters in its formula (description): the mean (arithmetical average)
and the standard deviation. This week we're working to show both (a) and
(b) after which we move on to Chapter 7.

The lectures will therefore be a journey through a collection of
carefully chosen problems (your homework assignments) and the labs
are meant to give you help (support) in covering the material in the
book. If you come to lecture and labs and understand what is discussed
there should be no problem getting a perfect score in this class.

I'll have more info about the upcoming exams during the week. The
exams will be open book (both) and will simply test that you  have
done your readings and understood them. There will be a help session
next Mon in LH (Lindley Hall) 102 at 7:15pm for an hour---if you need
it. See you tomorrow. Please let know if you have questions and/or any
concerns.

... Adrian

Remaining homework assignments also to be posted by Mon morning.

Second set of questions is posted under Lab Notes Four.

Date: Mon, 3 Nov 2003 14:31:56 -0500 (EST)
From: Adrian German <dgerman@cs.indiana.edu>
To: A113 Fall 2003 Distr. List <dgerman@indiana.edu>
Subject: A113 Week Two 


Dear A113 Friends,

A113 starts in earnest this week.

Everything that you have done for the first week of assignments will
be taken into account only as bonus. That includes of course any work
you may have turned in for last Friday's lab and any of the first two
homework assignments (one was collected on Friday while the other one
was demonstrated in class last Thursday and is due this week).

Plan for this week: tomorrow in class I explain the coins experiment that
you started last Friday (on conditional probability). Then on Wednesday in
lab please make sure you examine the first thre chapters in the book. The
lab notes will offer some guidance. Thursday we cover Chapters 4 and 5, so
on Friday in lab a similar set of notes will try to guide you through the
experiments that you need to have undertaken to be sure you understand the
materials in the chapters as well. Following weeks' plans will be posted
on the What's New page in advance, soon.

Lab assignments will be weekly and in written form. The web site will have
complete details (in lab notes) on what is due and when.

I think the enrollment in this class has finally stabilized so I'd like to
use this opportunity to actually welcome everyone to A113.

Please let me know if you need any help or have questions/concerns.

... Adrian

Here are screen shots that show how the golden coin experiment was done:

1. the start of the experiment
2. the end of the experiment
3. the mechanism (set of formulas)

Notice that the last picture is really big so we can see.

Also notice that for this problem to be really effective one needs to

• provide two fairly short,
• and somewhat persuasive solutions,

that give different answers.

So here they are:

1. Solution One: the probability must be 1/2. The reason is: if the first coin is gold, then you are in either drawer one or drawer two. So there are two possible cases. In one of the two possible cases you get a gold coin and in the other a silver coin. So 1 in 2 is the chance to get a second gold coin if the first one is gold.

2. Solution Two: the probability must be 1/3. The reason is: you basically pick two coins and would like the coins to be both gold. You have three drawers, so the total number of cases is 3. In only one case you get two gold coins, so you have 1 in 3 chances of getting a second gold coin when the first one is gold.

This way one of them must be incorrect (perhaps both?) and a carefully chosen experiment can tell which is incorrect and give a hint as to what the right answer might be.

Remember:

1. the purpose of computing is insight, not numbers.
2. the only way we can deal with infinity is by thinking about it.

Here's a copy of the message sent last night:

Date: Tue, 28 Oct 2003 19:30:29 -0500 (EST)
From: Adrian German 
To: A113 Fall 2003 Distr. List 
Subject: A113 Fall 2003


Greetings,

You are receiveing this message because you are on the roster for A113
(or told me you would add this class so I added you on the list). This
is the e-mail distribution list for A113 for the remaining of the semester
so if you don't want to be on it for some reason, or if this message reached
you in error please let me know.

Labs start tomorrow. Please write to me if you don't know your section or
where to report to. Michael's lab is in Lindley 023 at 9:30am. Nathan's lab
is in Ballantine 308 at 11:15am. And Sriram's is in HPER 154 at 4pm. I have
a list with each student's lab section so if you need the info that refers
to you on that list please let me know.

Finally, I remind you the address of the web site for this class

             http://www.cs.indiana.edu/classes/a113

We'll be using Oncourse for the Practical Exams and maybe for some Homework
Assignments but the web site will be at the address announced above. Please
check the What's New daily and What's Due frequently on this web site.

There's nothing due tomorrow in lab. Try to go over the experiment at the
end of Lecture Notes One. Lab Notes for the week will be posted tomorrow. We
will base everything we do in this class on the text and the web notes. If
you need any help or have any questions whatsoever please let me know.

... Adrian