STAT 478-01 Regression Analysis
Fall 2019
|
TR 9:00-10:20
|
VH 1216
|
Office Hours: MWF 10:30-11:20, 1:00-2:30; TTh 10:30-11:20; or by appointment
VH2236, x7236 (no voice mail), clthatch@truman.edu
-
- Course Description: Simple and multiple linear regression theory and applications, including matrix formulation, estimation and inference, validation of assumptions, model building, and time series. Additional topics may include weighted least squares, logistic regression, Poisson regression, generalized linear models, nonlinear regression, or mixed effects models.
- Pre-requisites: Completion of the Essential Skills in Statistics, MATH 192 Essentials of Calculus or MATH 198 Analytic Geometry and Calculus I, MATH 285 Matrix Algebra or MATH 357 Linear Algebra, and STAT 250 Statistical Computing.
- Course Objectives:
- Recognize the overall importance of linear regression for modeling and understanding relationships between variables
- Be able to use simple linear regression as both a descriptive technique and for inference, including scatterplots, matrix formulation, least squares and maximum likelihood estimation, prediction, interpretation of slope and y-intercept, confidence intervals and hypothesis tests regarding coefficients, and prediction intervals and confidence intervals for the mean response
- Be able to build multiple linear regression models with numerical covariates, curve fitting, indicator variables, and interaction effects
- Be able to compare multiple regression models through hypothesis testing and measures of model fit and variable selection (e.g. AIC, BIC)
- Understand the limitations of regression, including correlation vs causation, extrapolation, multicollinearity, and outliers and influential points
- Understand the assumptions of regression models, including how to validate these assumptions, the consequences of their violations, and possible corrections to the model when the assumptions are violated
- Understand and interpret output from a statistical package
- Understand time series models such as linear trends, exponential models, and autocorrelation and lagged models in the linear regression context.
- Credit Hour Justification: This is a three-credit hour course that meets for two 80-minute sessions of classroom instruction each week for the full semester. The “average” student should expect to spend at least six hours each week on out of class work (reading the textbook, doing homework problems and class project, and studying). However, this is an average time per week for an average student and may have weekly variations.
- Required Text: Introduction to Regression Analysis, 5th ed, Montgomery, Peck, and Vining
- Web Materials: I will post the syllabus, homework assignments, handouts, and other important information on Blackboard. A copy of this syllabus and the course schedule is also available on the Web at https://clthatch.sites.truman.edu/.
- Calculators and Software: You will need a calculator capable of square roots and exponents. A “statistical” calculator is not required for this course, but you may find it useful. The mathematics department maintains a collection of graphing calculators that may be checked out for the semester on a first-come, first-served basis. While you may use a calculator application on your phone for classwork, cell phones will not be allowed during exams and you will need an actual calculator for exams.
You may use R, Minitab, SPSS, Excel, or Open/LibreOffice (all available on university computers) for any homework assignment; a few assignments may require you to use specifically R or Minitab, but you may still check your answers using the other software. We will primarily discuss the use of R and Minitab in class. R may be freely downloaded from www.r-project.org , and R Studio may be freely downloaded from www.rstudio.com for use on a personal computer. A student version of Minitab can be rented ($30) for the semester for use on a home computer from www.onthehub.com/minitab/. Additionally, Open/LibreOffice can be freely downloaded from www.libreoffice.org for use on a personal computer. If you wish to use another software package, check with me first. - Attendance: Regular attendance in this course is expected; however, attendance is not strictly required. That is, I will not penalize your grade simply for missing class, but you should be aware that frequent absences are associated with less understanding of the material and lower grades. I strongly encourage you to come to every class period. Please stay home if you have had a fever, vomiting or diarrhea in the last 24 hours; you do not need to provide me with medical documentation.
Note: Students are still responsible for completing all assigned work by the given deadlines unless other arrangements have been made with the instructor prior to the absence. If the absence is unexpected, the student should contact the instructor as soon as possible to arrange to make up the missed work.
The university attendance policy is available in the Course Catalog: Attendance Policy - Preparation for Class: Of course, simply attending class is not sufficient to master the material. Completing the reading and problem assignments on time is very crucial. If you do not understand a topic being studied, get help as soon as possible. Before each class you should review the previous class’s material and be prepared with any questions. You should bring a calculator to class for active participation. On occasion you may also use a laptop/tablet or other networked device for in class demonstrations and activities. You do not need to bring your textbook to every class; however, there will be occasions where tables from the text will be useful in class.
- Academic Integrity: You will be expected to follow all university policies on academic integrity1. In particular, you should only use authorized (and acknowledged) sources on graded work. Serious violations will be reported to the Chair of the Department of Statistics and the Dean of the School of Science and Mathematics. You will, however, find that my list of authorized sources is fairly generous; refer to the sections on Homework and Tests below.
- Grading: The final course grade will be calculated as 0.30*(homework percentage) + 0.40*(test percentage) + 0.10*(project percentage) + 0.20*(final exam percentage). Letter grades will be assigned on a standard 10-point scale (e.g. 90% and above is an A).
- Homework: The purpose of homework is to allow you frequent and prompt feedback on your progress. Homework assignments will be given approximately once per week. Point values for individual problems will be identified when the assignment is posted. You may work together on your homework, and you may turn in one paper for the group. You should always show your work on homework problems; failure to do so can result in a loss of half of the points for that problem— even if the answer is correct. Papers that are too messy to read, are not stapled or paperclipped together, or do not have the problems in the right order may lose 10% of the assignment’s possible points.
Homework assignments are due by 3pm in my office or my mailbox. You may resubmit up to one (1) homework assignment to replace a missing or unsatisfactory score. Any assignment submitted after the due date will be treated as a resubmitted assignment and will count for the 1 resubmission limit; no other late homework will be accepted (except as arranged with the instructor per the attendance policy). You may appeal to me directly if you feel you have extenuating circumstances (death in the family, etc.). Resubmissions have implicitly waived the right to prompt feedback and will have the lowest priority in being graded.
Your homework percentage— (points earned)/(points possible)— will count for 30% of the course grade. - Project: We will have a semester project that involves collecting and analyzing data and presenting the findings in a written report. More information about the project will be provided later. Your project percentage will count for 10% of the course grade.
- In-class Tests: There will be three (3) in-class tests. In the nature of mathematics, these tests will be partly comprehensive. These tests are closed-book, closed-notes; however, you may bring in one standard (8.5 x 11) sheet of formulas, definitions, etc. Make-up tests will be given only with prior notice and only when merited. Your test percentage— (points earned)/(points possible)— will count for 40% of the course grade. The dates and tentative topics for these tests are:
Test 1: Thursday, Sept. 19 – Ch. 1-2 Basic Statistics Review and SLRTest 2: Tuesday, Oct. 29 – Ch. 3-5, 9 MLRTest 3: Tuesday, Dec. 3 – Ch. 6-8, 10 Additional Topics
- Final Exam: The final exam will be Thursday, Dec. 12 at 7:30-9:20. The final exam will be comprehensive, including material taught after Test #3. Your final exam percentage will count for 20% of the course grade.
The complete final exam schedule for the university may be found at Final Exam Schedule
- Homework: The purpose of homework is to allow you frequent and prompt feedback on your progress. Homework assignments will be given approximately once per week. Point values for individual problems will be identified when the assignment is posted. You may work together on your homework, and you may turn in one paper for the group. You should always show your work on homework problems; failure to do so can result in a loss of half of the points for that problem— even if the answer is correct. Papers that are too messy to read, are not stapled or paperclipped together, or do not have the problems in the right order may lose 10% of the assignment’s possible points.
- How to Learn Statistics 2
- Make it personal: Learning is easiest when you are interested in the material. You have to connect what you are learning to what is important to you. If you are taking statistics only as a required course, think about why your field would want you to know statistics. Although statistics uses a lot of mathematical notation, as a field it primarily arose from social scientists and natural scientists who needed to use it. If you are in those fields, you’ll probably need to use it too.
- Learn the language: Statistics may seem like a foreign language with its own terminology and notation. Most of it is pretty logical, however, and you have probably used many of the terms yourself already. This can be a good thing or a bad thing. For example, many lay people use the term correlation and you probably have some notion what it means; however, the statistical definition is rather precise and does not apply to many situations where people use the term. So, make sure that you know the statistical definition of terminology that we use.
- Practice, practice, practice: To learn statistics (or any other material), you generally need to be exposed to the content three or four times. Hence, just hearing the material once in class may not be sufficient. You should read the book, do the homework, review your class notes between classes, and review your notes and rework problems before an exam. Another thought, although it is often said that “practice makes perfect,” keep in mind that practicing something in the wrong way will only make it worse. Check your answers, and review what you have missed on homeworks or exams.
- Keep up: Like many subjects, statistics will build on concepts throughout the semesters. If you start to fall behind, get help early. If you really understand why you are doing what you are doing, you will find that most of the material follows logically from a few major concepts and techniques. On the other hand, even if you don’t know why you are doing it, keep doing it; sometimes understanding follows. If you don’t know how to do a problem, think about possible partial answers. For example, if I ask you to perform a two-sample t-test to compare the means of two groups, even if you have no idea what a t-test is, you can probably calculate the means of each group and make a guess of whether they are the same or different.
- Do well: You will enjoy the subject more if you actually succeed in it. This may mean putting in extra effort (see practice, practice, practice). If you find the material difficult, how much prouder will you be when you have mastered it?
If you have ever watched a baby learn to walk, you know that (s)he will fall down several times and get up and try again and again and again. If you put that much effort into learning statistics, you will learn it. Remember, we have all learned to walk; so you can learn statistics. - Don’t stop now: One semester of statistics does not make you a statistician. There is plenty of material that we do not cover. While I encourage you to take other upper-level statistics courses, you don’t have to take a class to learn more about statistics. The library has many texts that you may find interesting; try browsing the QA section of the general collection.
- Additional Sources of Assistance: I encourage you to come to Office Hours if you have any questions. Additionally, the Center for Academic Excellence (112 Kirk Bldg) and MAC (Adair Bldg) offer tutoring.
- Disability Services: The university is committed to making every possible effort to comply with the Americans with Disabilities Act (ADA). If you have a disability for which you are or may be requesting an accommodation, you are encouraged to contact both me and the Office of Student Access and Disability Services (x4478) as soon as possible. For more info, see http://disabilityservices.truman.edu/
- Title IX: As an instructor, one of my responsibilities is to help create a safe learning environment on our campus. Additionally, faculty members are mandatory reporters; we are required to share information regarding sexual misconduct or information about a crime that may have occurred on Truman’s campus with the University. Students may speak to someone confidentially by contacting University Counseling Services at 660-785-4014 (660-665-5621 for after-hours crisis counseling). For more information regarding Truman’s policies and procedures relating to any form of gender discrimination, please see http://eoaa.truman.edu/university-non-discrimination-policy/ and http://eoaa.truman.edu/complaint-reporting-resolution-procedure/.
- Emergency Procedures: In each classroom on campus, there is a poster of emergency procedures explaining best practices in the event of an active shooter/hostile intruder, fire, severe weather, bomb threat, power outage, and medical emergency. This poster is also available as a PDF at this link: http://police.truman.edu/files/2015/12/Emergency-Procedures.pdf . Students should be aware of the classroom environment and note the exits for the room and building. For more detailed information about emergency procedures, please consult the Emergency Guide for Academic Buildings: http://police.truman.edu/emergency-procedures/academic-buildings/
This six-minute video provides some basic information on how to react in the event there is an active shooter in your location: http://police.truman.edu/emergency-procedures/active-shooter/active-shooter-preparedness-video/
Truman students, faculty, and staff can sign up for the TruAlert emergency text messaging service via TruView. TruAlert sends a text message to all enrolled cell phones in the event of an emergency at the University. To register, sign in to TruView and click on the “Truman” tab. Click on the registration link in the lower right of the page under the “Update and View My Personal Information” channel on the “Emergency Text Messaging” or “Update Emergency Text Messaging Information” link. During a campus emergency, information will also be posted on the TruAlert website http://trualert.truman.edu/.
1See for example: Truman State University General/Graduate Catalog. Truman State University: Kirksville, MO, 2019. http://catalog.truman.edu/content.php?catoid=10&navoid=451#Academic_Dishonesty
2Portions adapted from Hutchinson, Paul. 1997. (http://www.angelfire.com/biz/rumsby/ASTUDY.html)