Download Math 108 Introduction to Statistics

Math 108 Introduction to Statistics
Spring 2017
Instructor:
Dr. Jeff Graham
Office:
Fisher 312
Office Hours: MWF 10 – 11 or by appointment
Email:
[email protected]
Phone:
372-4467
Text:
online, see below
About the course
This is a flipped class. This means no lectures by me during class time. You will need to go to
oli.cmu.edu to create an account and sign up for the class for the lectures. You should have
received an email about this already, but in case you missed it, you will need the course id
which is MATH108-S17 and the password: Graham. There is a $25 to join the class, but this is a
smashing bargain compared to buying a text for the course. You need to do this ASAP if you
haven’t done this already. There are about 23 online quizzes for you to get through this
semester. The dates that they are available are on the course syllabus on oli.cmu.edu. If you
miss the dates, too bad for you. Keep up with the work!
Grading
There will be several components to your grade:
Online quizzes:
In class activities:
Exams:
Final Exam:
Bonus:
Attendance:
A--- 900 – 1000 points
B--- 800 – 890 points
C--- 700 – 790 points
D--- 600 – 690 points
F--- below 600 points
23 worth 10 points apiece for a total of 230 possible points.
Weekly homework problems totaling 230 possible points.
There will be three exams each counting 100 points.
The final exam is worth 200 points and is comprehensive.
There is a 20 point bonus for completing all the online quizzes.
40 pts. This (mostly) includes keeping up with the online material.
General Policies
The university policies on attendance and cheating as described in the student handbook will be
followed. A quick summary: Come to class and don’t cheat.
Learning Goals
These sound a little intimidating, but they boil down to learning some math and how to use it.
1. Abstract a problem into a symbolic or mathematical model or framework.
Statistics is a mathematical framework for collecting, organizing, and interpreting data. There
are three main topics in our introductory statistics class; descriptive statistics, probability, and
inferential statistics. A statistical study involves identifying a question, designing a method of
collecting meaningful data, analyzing, and drawing sound conclusions from this analysis. A
typical statistics class would spend at least some time on each of these steps.
2. Interpret such a model in terms of a real world construct.
Statistics texts are very big on this topic. Every text has numerous examples drawn from the
natural and social sciences, engineering, medicine and many other fields of study. Students in a
statistic class will learn to take the information from the problem, perform appropriate
calculations, draw conclusions from the calculations and interpret the results in the original
context.
3. Reason from precisely stated principles using deductive methods and draw valid
conclusions.
Probability theory starts with axioms, and so this section of the course involves reasoning from
those axioms. For example, given information about a population, we ask students to determine
the probability a random sample will turn out a certain way. Hypothesis testing and related
topics (Type I, II errors and the like) are also examples of deductive reasoning. A highly simplified
version of the typical hypothesis test is as follows: given a hypothesis to test, calculate a test
statistic from a data sample, and based on the value of this test statistic the student decides
whether or not to reject the hypothesis. This glosses over the reasoning needed to choose the
appropriate test statistic, but it is a prime example of deductive reasoning in this course.
4. Recognize, manipulate and reason from or about abstract patterns.
Mathematics is the science of patterns and so every course we teach is really all about
recognizing, manipulating, and reasoning from patterns. In statistics, descriptive statistics and
regression are all about identifying patterns. For example, the main goal of descriptive statistics
is to attempt to identify information about the distribution of a set of data. Is it skewed? Is it
symmetric? Are there outliers? Does it look normally distributed? Linear regression tries to
determine the strength of a linear relationship between two variables. For both of these topics,
one generally starts with a plot of the data to help identify any obvious features.