POLITICAL SCIENCE 516 M 7-9:40 p.m. R. Hofstetter AH-4116 594-6244 e-mail: richard.hofstetter@gmail.com (22498) MCN-105 MT6-7 p.m. STATISTICS FOR POLITICAL SCIENCE GOALS: Statistics flow from common sense and are tools for assessing the degree of confidence one may have in assertions about empirical reality, nothing more. It is impossible to be a rational actor in the absence of understanding at least the rudiments of statistics for that reason. This course introduces students to basic concepts, theories, and methods by way of which qualitatively and quantitatively oriented political and other scientists arrive at decisions about observations using statistics, calculators, and computers. The course is designed to serve as an introduction to the use of systematic statistical methods in political science suitable for political science graduate students at SDSU. The goal is to convey an understanding of statistical decision making involving the use of simple probability to evaluate empirical claims, and to familiarize students with a common statistical package (SPSS) used to analyze data involving common statistical models. Announcement: I do not plan to miss any classes except on national holidays (9/2 and 11/11 this year). PREREQUISITE: Lose whatever sense of entitlement that you have, I expect students to work hard in this class but hard work does not guarantee a passing grade. Willingness to work hard, learn about serious research, share knowledge with others, have open mind facilitates success in this class and elsewhere. MATERIALS: Required: Alan Agresti and Barbara Finlay. (2009). Statistical Methods for the Social Sciences. New York: Prentice Hall, 4th ed. May be available for rent and/or in electronic form or used. If you plan to work in a research capacity or any kind, I do suggest you add this book to your professional library. Optional: Any inexpensive introductory guide to SPSS. I did order one for the course. But you may wish to check utube/spss for SPSS mini- lessons. If you type “utube/spss” in a browser, this will supply you with various SPSS demonstrations and is cheaper than the book. An inexpensive USB drive for backup of data, setups, and work (1-2 gigabytes adequate). Frys’ specials are probably least expensive but get there early. You will also need an inexpensive calculator with a square root key. I recommend one that is solar powered since neon lights will run them. Avoid hand computers and cell phones, since neither will be allowed to be used in this class. POLICIES AND PROCEDURES: This course is quite different than other political science classes. It does not involve a large amount of reading, just 367 pages during 15 weeks, but the reading is technical and requires a great deal of very careful, repetitive study. Casual reading of statistics the day before a class simply does not lead to learning. The class also involves continuous work, including problems and computer exercises due every week in addition to quizzes or POLS 516 [JUNE, 30, 2013] PAGE 1 midterms. Everyone is required to do all of the assigned work and turn it in on time each week. It is best for students to explore other options if they are not prepared to put in the required effort. A good rule of thumb is to plan to spend at least 9-10 hours outside of class each week although some weeks may require more effort. Preparation of detailed reading notes also helps students to learn. Study groups that meet at least weekly composed of 3-5 students and that discuss homework problems are extremely useful for developing collective understanding, and I urge students to form such groups. Please do not bring laptop computers to class, and if you bring them, do not turn them on. Unless you are expecting an emergency call (in which case tell me before class begins), do not bring cell phones to class and if you bring them keep them out of sight and turned off. Cell phones and computers may not be used as calculators in this class so do not ask. In the case of certain emergency situations discussed with me in advance, cell phones may be left on. Students should also open e-mail accounts. My e-mail address is available on the syllabus and on blackboard to enable students to communicate with me outside my scheduled office hours for students' convenience. I anticipate that students will make use of this opportunity, since I do not check telephone voice mail at SDSU. I only occasionally receive e-mail M and T mornings, but do monitor e-mail most other days before evening. I do not receive e-mail evenings on Sunday afternoons or evenings, and occasionally do not receive e-mail at other times. It is not useful to send e-mail at the last minute to raise questions since I may not see your message in time to respond. Students should complete reading and computer problems in a manner that maintains pace with the topics being discussed and the lectures. Students should study the material for the week prior to coming to class. Students are required to attend all lectures and discussions, turn in all homework problems and computer problems at the beginning of the class on the date due (or leave them in my departmental mailbox prior to the time due), and take examinations and quizzes in class. Completing assignments punctually will require prior planning, since SDSU computers may be difficult to access during certain times in a semester and computers/printers sometimes malfunction. In the past, some students have followed the unwise policy of waiting until the last day before assignments are due to complete computer work. It is prudent to familiarize themselves with all campus labs that run SPSS. Instruction and introductory comments on how to use SPSS and computers will be provided during special sessions in weeks two and three. An inexpensive suggested paperback has also been ordered for the class and may be helpful. Beyond this, if you do not know how to perform the computer work required for this class, it is your responsibility to find out (including by asking me) and to complete the work. I may give tests that require use of computers. I am available in AH 4116 during scheduled office hours to provide assistance. Other computer software may be used to complete assignments but I may not be able to provide instruction on how to use it. Quizzes or a midterm examination will be given at each class after the third meeting, but no final examination will be given. Quizzes and examinations are closed book, although three pages of handwritten notes (no zeroxes) and printed statistical tables (zerox OK) from the text may be used (of course students need to remember to bring tables to class). Please make two copies of other graded work (one to turn in for a grade and one to keep as a backup in the unlikely event that the submitted work is misplaced). It is the student’s responsibility to bring calculator, tables, handwritten notes, assignments, paper and pencils/pens to class. POLS 516 [JUNE, 30, 2013] PAGE 2 Learning in this course depends on mastering material cumulatively (week by week) to a much greater extent than other political science courses. I expect students who are having difficulties in this course to contact me during office hours at the beginning of the difficulties so that grades do not suffer. Graded exercises are assigned for each week. Students would be wise to complete further exercises at the end of each chapter for practice. Additional practice problems should help with understanding statistical techniques and I will be happy to discuss them, but extra work may not be graded. Although students may consult others in the class or with me in figuring out how to use SPSS to compute statistics, the purpose of the exercises is for each person to learn how to use SPSS with data to perform computations that lead to conclusions. Thus, it is imperative that each person learn how to use SPSS with or without consultation. I feel free to test students on their individual ability to use SPSS to complete problems, and this policy includes midterm examinations. Work with your group members but make sure that you know how to use SPSS to do what is required. Academic dishonesty will not be tolerated. All written work must be your own original work (i.e., not previously submitted for credit in any other course, either at SDSU or at any other academic institution) and also not copied from anyone else or any other source. Purchasing/downloading papers which you turn in as your own is de facto cheating as is copying the work of other students in this class. Plagiarism is using the work, e.g., ideas, writings, concepts, arguments, data, analyses, etc., of others without attribution. To clarify the preceding, discussing how to do problems in a study group is permissible, indeed desirable, if you are having problems with how to solve a problem or understand a concept, but simply copying other’s work is not. Each person will be tested on his/her knowledge of how to do all of the work in the class, so students need to learn how to do all the problems. When in doubt, use a footnote to other person’s work. Please familiarize yourself with the University Policy regarding Cheating and Plagiarism at: http://wwwrohan.sdsu.edu/dept/ senate/policy/ pfacademics.html Also be aware of the Student Grievances procedure, available on-line at: http://www.sa.sdsu.edu/srr/statement/sectionVII.html Plagiarism not only includes a student representing the works, ideas, and writings of others as his or her own. It also includes a student using a previous work product (published or unpublished) that the student was involved in (e.g., as sole writer or co-writer) for any academic-related activity (e.g., class assignments and tests; theses; and the non-thesis options, etc.), representing it as original work that was performed for the new academic-related activity. Unpublished works include, but are not limited to: unpublished manuscripts, grant proposals, and web-based material. POLS 516 [JUNE, 30, 2013] PAGE 3 NOTICE: Do not come into this class expecting any particular grade. I do not give grades; students earn grades by their performance in class participation and on assignments and tests. Except in the case of documented illness requiring medical attention or a death in the immediate family, late assignments will be graded 0 points. Students who fail to turn in all assigned work by the last class meeting (team computer exercises and individual survey project steps) will fail this course with a grade of F regardless of other performance. All work must be turned in by 7 pm, December 9, 2013, in order to avoid an F. GRADING: Examinations Quizzes Homework problems Computer problems Total 30% 30 20 20 100% Computation of Grades. Please read this paragraph carefully. Grades will be based on a straight percentages with 90%-100% being an A, 80%-89% a B, 70%-79% a C, 60%-69% a D, and below 60% an F. In order to pass this course with a grade above an F, students must complete all assignments and turn them in by the last class day, December 9, 2013. Late papers will be accepted until the end of the course, but 10% will be subtracted for each week an assignment is late. Final course grades will be based on the distributions reported above. Numerical scores will be entered in blackboard to communicate scores to students, but final letter grades will not be entered in blackboard. If mistakes occur in grading, then it is imperative that students bring the original graded papers for exams, quizzes, or homework to me within two weeks of the graded materials being returned so that I can make corrections. No changes will be made in the absence of the original graded work. I will do my best to grade all assignments, quizzes, and midterms and post the scores on blackboard within two weeks of the exercise being completed. NOTE: Make ups for quizzes and exams will be given only when permitted by me and with a written documented reason for an absence. Excused absences will be given only for extreme medical or similar emergencies with written documentation. Airplane reservations, vacations, athletic events, etc., do not qualify. Assignments: The course is divided into a series of topics, described below. Approximately one week of lecture and discussion is allocated to each topic. Assigned computer problems are at end of the syllabus. One set of two computer problems may be submitted May 5 to replace the lowest prior exercise score. Week 1 Aug 26. What do statistics, science, and computers have in common? Why is each relevant for the other? Introduction to computers, science, statistics. Read Agresti & Finlay, Chapter 1, "Introduction," pp. 1-9. Exercises 1.2, 1.9, 1.14, 1.17 due Sept. 9. Week 2 Sept. 2 Labor Day Holiday Week 3 Sept. 9 How is SPSS used? How to do it? SPSS Workshop, AL-106. Read Agresti & Finlay, Chapter 1, "Appendix A," pp. 545-555. Last day to drop is today, September 5, last add for this class is September PAGE 4 POLS 516 [JUNE, 30, 2013] 9. Exercises 1.2, 1.9, 1.14, 1.17 due today. Week 4 Sept. 16. SPSS Workshop, AL-106. How to use the software (SPSS) for this course. a. Review notes, read Lee A. Kirkpatrick and Brooke C. Feeney. (2006), or check an introductory session on UTUBE/SPSS. A Simple Guide to SPSS for Windows. Belmont, CA: Thomson Wadsworth, 9th ed. How is SPSS used? How to do it? b. SPSS Workshop, AL-106. Week 5 Sept. 23 What is a sample? Why use one? How is measurement done? What is the big fuss about qualitative and quantitative? Read Agresti & Finlay, Chapter 2. “Sampling and Measurement,” pp. 11-29. Exercises 2.1, 2.7, 2.15, 2.26, 2.28 due September 30. Week 6 Sept. 30 What are descriptive statistics? Which are used the most and why? Agresti & Finlay, Chapter 3, “Descriptive Statistics,” pp. 31-71. Exercises 2.1, 2.7, 2.15, 2.26, 2.28 due today. Exercises 3.1, 3.9, 3.16, 3.62, 3.70 due October 7. Week 7 October 7. What is probability? Why is it useful? Is probability the universal truth? Agresti & Finlay, Chapter 4, “Probability Distributions,” pp. 73-106. See also Appendix 1 at the end of this syllabus for explanation of using binomial distribution for computing probabilities of certain kinds of data. Exercises 3.1, 3.9, 3.16, 3.62, 3.70 due today. Week 8 October 14. What is estimation? How is it conducted? Is it useful? Agresti & Finlay, Chapter 5. “Statistical Inferences: Estimation,” pp. 107-141. Exercises 4.1, 4.10, 4.24, 4.53 due today. Computer problems 4.49 due today. Exercises 5.1, 5.26, 5.56, 5.61 due October 21. Computer problems 5.49 due October 21. First Midterm Examination. Week 9 October 21. Is inference significant? Is significance inferential? Why should we care? Agresti & Finlay, Chapter 6, “Statistical Inference: Significance Tests,” pp. 143-182. [Be sure to use the seven-step paradigm for hypothesis testing that I describe for all tests of hypotheses in this course.] Exercises 5.1, 5.26, 5.56, 5.61 due today. Computer problems 5.49 due today. Exercises 6.3, 6.16, 6.48, 6.49 due October 28. Computer problems 6.37 (download data from blackboard) due October 28. Week 10 Oct 28. How do we know Nazis and Communists are different? Are they? Agresti & Finlay, Chapter 7. “Comparison of Two Groups,” pp. 183-219. Exercises 6.3, 6.16, 6.32, 6.48, 6.49 due today. POLS 516 [JUNE, 30, 2013] PAGE 5 Computer problems 6.37 due today. Exercises 7.2, 7.14, 7.51 due November 4. Computer problems 7.45 due November 4. Week 11 Nov. 4. Is love associated with war? How do we find out? Agresti & Finlay, Chapter 8. “Analyzing Association Between Categorical Variables,” pp. 221-253. Exercises 7.2, 7.14, 7.51 due today. Computer problems 7.45 due today. Exercises 8.1, 8.8, 8.17, 8.35 due November 18. Computer problem 8.29, 8.32 due November 18. Second Midterm Examination. Week 12 Nov. 11. Veterans’ Day Holiday Week 13 Nov. 18. Is the world predictable? Why should we care? How do we do predictions? Agresti & Finlay, Chapter 9. “Linear Regression and Correlation,” pp. 255300. Chapter 10. “Introduction to Multivariate Relationships,” pp. 301-320. Exercises 8.1, 8.8, 8.17, 8.35 due today. Computer problem 8.29, 8.32 due today. Exercises 9.1, 9.10, 9.57, 10.1, 10.11 due November 25. Computer problems 9.25, 10.27 due November 25. The world is complicated and how do we describe it? Agresti & Finlay, Week 14 Nov. 25. How can predictions be based on many variables? How can we know the unique effect of a variable on what we are interested in? Agresti & Finlay, Chapter 15. Multiple Regression and Correlation,” pp. 483-518. Exercises 9.1, 9.10, 9.57, 10.1, 10.11 due today. Computer problem 9.25, 10.27 due today. Exercises 11.2 due December 2. Computer problems 11.35 due December 2. Week 15 Dec. 2. Chapter 15. Multiple Regression and Correlation,” pp. 483-518 continued. Exercises 11.2 due today. Computer problems 11.35 due today. Exercises 11.41, 11.42 due December 9. Computer problems 11.36 due December 9. Week 15 Dec. 9. Third Midterm Examination. Exercises 11.41, 11.42 due today. Computer problems 11.36 due today. Note Concerning Computer Problems POLS 516 [JUNE, 30, 2013] PAGE 6 Instructions: Go to this course on blackboard where you will find designated command files (embedded in AGRESTI.SPS) to generate SPSS data files for each assignment. Students should copy the SPSS command and data files onto their own storage devices so that they can use any computer with SPSS installed on it to complete the computer exercises and thereby enhance flexibility. While I encourage students to work together on problems and thereby help each other and themselves, I also expect each student to master SPSS and to write his/her own responses to the assigned problems. I may give examinations that require each student to complete computations using SPSS during the examination time and that work will require fairly thorough knowledge of procedures for completing these problems. Appendix 1 Probabilities (The Binomial Distribution) When one has an a priori probability that can be assigned to the outcome of an event and is evaluating repeated independent trials of that event, such as coin flips, the binomial distribution can be used to evaluate the probability of discrete events. Let us imagine that we are flipping a coin three times and wish to evaluate the probability of two heads in 3 flips. If we have an honest coin, the probability of a head in any single flip is .5 and the probability of 3 heads in 3 flips is computed using the multiplication rule: P(3H)=P(H) X P(H) X P(H), or .5 X .5 X .5 = .125. There is only one way that a person can flip 3 heads in 3 trials. However, there are several ways one can flip 2 heads in 3 trials. Let’s start with a single outcome: P(2H)=P(H) X P(H) X P(T) is one of the ways of doing this, but only one of the ways. The probability of this one outcome is P(2H) = .5 X .5 X .5 = .125. The problem is that there are other ways of reaching the desired outcome and we have to include those in the assessment of the problem. The binomial expansion is used to accomplish this. The binomial involves two terms: 1) One states the number of ways that a single event can occur in a specific number of trials; and 2) the other is the probability of any one of the events occurring. The latter is given as P(2H), or .125 as shown above. The number of ways that two heads can occur in three flips is given by: (PxQ1-x) n! Where Px is the probability of a success and Q 1-x is 1- the r!(n-r)! probability of a success, n is the total number of events and r is the number of events that we are interested in. x 1-x In this case, P =.5, Q =1-.5=.5, r = 2, the number of H type outcomes, and n = 3, the total number of flips. n! (n factorial) = 1 X 2 X 3 = 6, and r! = 1 X 2 = 2. n – r = 3 – 2 =1, so the number of ways we can obtain two heads in three flips is 6 / 2 = 3. The probability of one of these occurring is the number of ways it can happen times the probability of any single way, or 3 X .125 = .375. The the probability of 2H in 3 flips = .375. Numerous laboratories on campus include SPSS among other programs. Data from blackboard and the command files that you write to conduct analyses should be written on a USB drive. 2 gigabyte or larger USBs are available inexpensively and are more than adequate for this course. Fry’s usually has excellent sales if you show up when the store first opens (off Aero Drive on I15 north, or in San Marcos just south of the freeway). POLS 516 [JUNE, 30, 2013] PAGE 7 A student version of SPSS can be purchased from the bookstore if it is not possible to use campus computers which are free. If you decide to purchase SPSS be sure that you understand what the version you buy does (needs to do multiple regression and other common statistics but not more sophisticated statistics, crosstabulations, t-tests, means, standard deviations, etc.), how long the license permits you to use the program (some licenses may be for 1 year or shorter—after that the program can no longer be used without paying for it again). POLS 516 [JUNE, 30, 2013] PAGE 8