Business Statistics 41000-01,02,81 - PDF

Document Sample
Business Statistics 41000-01,02,81 - PDF Powered By Docstoc
					                            Business Statistics 41000-01, 02
                                       Chicago Booth
                                       Autumn 2009



Instructor:    C. Alan Bester
               309 Hyde Park Center
               (773) 834-1714

TAs:           Daniel Wilhelm (Head TA)     
               Maria Rios                   

Course Website:
            Please visit this website and bookmark/add it to favorites now.
            Also please take a moment to view the Syllabus and Course FAQ links.
            You can also find a link to this website by searching on Chalk or Google.

Textbook:      Lecture notes, available on course website, are the primary “textbook”.
               Data Analysis for Managers or Data Analysis and Decision Making
               by Albright, Winston, and Zappe (2003 or 2008), (NOT required)

Software:      Microsoft Excel, with StatPro add-in (See below); Minitab (optional)

General Information

         This course is designed with two objectives in mind. First, this is a “ground up”
statistics class. While most Booth students have some prior exposure to statistics, this class
should be accessible to students who, for example, have never seen a sample mean or
median. Second, this course provides sufficient background in statistics to prepare students
for all but a handful of upper level elective courses at Booth. Put together, this means that
while some of the topics covered will be familiar, the pace and emphasis are substantially
different from traditional statistics courses.

        Two areas form the building blocks of statistical analysis techniques. We spend
weeks 1-2 of class on data analysis, which is about detecting, visualizing, and quantifying
patterns in observed data. Weeks 3-5 of class cover probability theory, which allows us to
build mathematical models to describe uncertainty about events. Weeks 6-8 cover statistical
inference, which combines ideas from data analysis and probability to quantify the strength
of evidence in the data and make decisions based upon that evidence. Weeks 9-10 cover
regression, one of the most powerful and widely used tools in statistics. Please note that
this course is highly cumulative – in the past, students who fall behind early in the class
have had extreme difficulty catching up.
Course Materials

        I keep all materials on the course website (link on the previous page). All of the
concepts we cover are in the lecture notes. Though a textbook is listed for the course, it is
very much optional; I will basically never refer to it in class. The lecture notes also contain
links to many of the data sets and examples we discuss in class. They are designed so
students can “work along” while reviewing what we‟ve done.

        We use Microsoft Excel together with a free add-in called StatPro (installation
instructions on the course website). Students may also use Minitab, which is available on the
Booth lab PCs (a tutorial is provided on the course website). Both are currently only
available for Windows PCs.* Students are not responsible for Excel or Minitab commands
on exams, but will be expected to interpret output from both packages as in the lecture notes.


        There are two exams:

               Midterm in Week 6 (11/2 or 11/4)               and
               Final Exam in Week 11 (12/7 or 12/9)

Both exams are in-class and closed book. One 8.5 by 11 inch “cheat sheet” may be used for
the midterm; both sides of the page may be used. Two 8.5 by 11 inch “cheat sheets” (again,
both may be double sided) may be used for the final exam. Please bring a calculator to the
exams. Sharing of calculators is not permitted. Devices capable of transmitting wireless
signals may not be used as calculators.

        Both exams are mandatory. Make up exams are provided only in case of extreme
emergencies – conflicts due to recruiting and/or other classes do not qualify. Students are
expected to take exams with their registered section (e.g., students registered for a Monday
section take their exam on a Monday). If you know in advance you will have a conflict
with either of the dates above, please register for another section.


        Grades are based on the following breakdown:

               10% Homework, 35% Midterm, 55% Final Exam

There is no “drop the midterm” policy. Please be aware that the school enforces a hard 3.33
GPA cap across all sections of a course taught by the same instructor in a given quarter.
Such a cap could be satisfied, for example, by giving the top 35% of students A‟s, the next
55% B‟s, and the bottom 10% C‟s. The actual distribution of grades will differ substantially
from quarter to quarter, and plus or minus grading IS used. Though students are encouraged
to make the instructor aware of issues with grades, such as employers‟ reimbursement
requirements, in the interest of fairness these issues cannot influence final grades.
       Re-grade Policy: Students may request a re-grade on homework assignments or
exams. Requests must be made within 7 days of the assignment being returned and must be
accompanied by a written statement explaining where the student believe the grading error(s)
occurred. Please note the entire exam/assignment will be re-graded – errors in the student‟s
favor may be corrected as well (this may be waived in the case of obvious clerical errors).


        Homework assignments are given weekly. All of the problems and data are available
on the „homework‟ section of the course website. There is no first class assignment.

       Students may work in groups (suggested size 3-6 members) and may turn in
homework as a group (all names must be listed on the first page). We highly encourage
students to work all problem sets on their own, then meeting as a group to discuss answers.
Working problems is the only way to learn statistics.

        Homework assignments are due at the beginning of the following week‟s lecture
(exception: the first two weeks‟ worth of homework is collected in week 3). Please hand in
hard copies. In the event of absence, homework may be submitted by email with the
instructor‟s permission.

        Homework is expected to be written up neatly (preferably typed, but handwritten is
acceptable) in “engineering format”, that is with work shown and answers boxed. Please pay
careful attention to instructions on what Excel/Minitab output to turn in.

      Late homework is not accepted. Assignments missed due to emergencies are dropped
when grades are computed, provided the instructor is notified before the due date.

Communicating with the Professor and TAs

       Please get in the habit of checking the “updates” link for your section on the course
website. It is essentially a weekly blog of what we‟ve done in class, future assignments, and
what you should be focusing on at various points in the quarter.

        The best way to contact us is via email at the addresses listed on the previous
page. Please CC both TAs on all emails, including those to the instructor. We make
every effort to respond promptly. Please keep in mind we have roughly 300 students this
quarter; please check the website before asking “when is homework due”. We also make
frequent use of the class email list – please check your Booth email regularly. If you don‟t
use your Booth account, please set up forwarding to another address.

        The instructor‟s office hours are by appointment and may be arranged by phone for
off campus students. The TAs will hold weekly reviews at times to be announced (email &
web). The first TA review will help familiarize students with StatPro and Minitab – please
bring a laptop if you plan to use one.

This is not a traditional statistics class. My Ph.D. is in economics, not statistics, so I tend to think
about and teach statistical techniques more in terms of their costs and benefits than how they are
derived mathematically. In particular, if you had an engineering or mathematical statistics class as an
undergraduate, you will probably find the presentation of material here very different. This is NOT a
class about memorizing a set of formulas and then plugging in numbers to solve problems.
Rather, the goal is to develop a strong conceptual and practical understanding of the statistical tools
we use to make business decisions. I believe doing so not only enables you to use these tools more
effectively, but also allows you to understand which tool is appropriate in a given practical setting.


         This is the question I get asked most in the first week of class. The typical preface is, “I‟ve
had statistics before, but it‟s been awhile and/or I don‟t feel like I really understood it”.
Unfortunately I can‟t tell you the answer because it‟s different for every student. The best thing I can
do is tell you what we are trying to do in this class and how I think the applied regression is different.

          I plan this course with two broad objectives in mind. First, this is a “ground up” statistics
class. While I recognize most students at Booth have seen statistics before, I have a responsibility
(that I take very seriously) to make this class accessible to students who have literally never seen a
sample mean or median. I also believe that the best way to understand in statistics is to develop a
solid foundation in data analysis and probability. As a result, the class starts slowly and more
advanced concepts like multiple regression receive limited treatment (we spend about 2 to 2 ½ weeks
on regression) at the end of the course.

         That said, my second objective is to give students sufficient background in statistics to take
all but a handful of upper-level elective courses here at Booth. I believe that if you have a firm grasp
of the fundamental concepts in statistics, you can understand what‟s going on in these classes, you‟ll
just have to invest a little more of your time to master the terminology and technical details that we
won‟t cover. Be aware, however, that if you plan to take data-driven marketing or empirical finance
or economics classes, you will likely encounter statistical techniques and terminology that are not
covered specifically in this class. We cover a lot of conceptual ground in this course, going from
basic data analysis from multiple regression in 10 weeks. The cost of this is that we don‟t have time
to develop the broad array of practical examples that you would see in the regression course.

        So in my opinion, most students who choose to skip this class and go right to regression
should do so for one of two reasons. Students who already have a solid foundation in statistics might
wish to take regression to avoid repetition. Though there‟s no one universal test, if you can give a
good answer and explanation to the question, “For a large number of i.i.d. observations, what is the
sampling distribution of the sample mean?”, you should almost certainly take regression. Second,
students who plan to take a lot of upper level electives where regression is used extensively (e.g.,
empirical asset pricing, data-driven marketing) might wish to do the extra work of getting themselves
ready to take regression off the bat, rather than reviewing the techniques needed for each course later
on. Of course, you always have the option of taking business statistics and regression, and in my
opinion 41000 and 41100 are most valuable as a two-course sequence. But with so many excellent
courses available here at Booth, many students don‟t have room in their schedules for both!

  Software such as Parallels Desktop, VMware Fusion, and CrossOver allows Windows apps to be run on a Mac; please note
the TAs and I do not provide support for such packages. You may wish to contact Computing Services for more options.