STA 5126 - Statistical Methods for the Social Sciences Dr. Mohr, Fall 2011 Instructor: Dr. Donna Mohr, Building 14/2702, Phone: 620-2884 (do NOT use voice mail) FAX: 620-2818 (put my name on cover sheet) email: dmohr@unf.edu Online materials: once you are registered, go to http://blackboard.unf.edu, prior to the course beginning, some materials at http://www.unf.edu/~dmohr/sta5126 Office Hours: Mondays and Wednesdays 10:30-11:30 am, 4:30 – 5:30 pm Tuesdays and Thursdays 1:30 – 3:30 pm Other times, just call for an appointment Course Goals, Content and Requirements: This course is intended for graduate students in psychology, sociology, education or the other social sciences. It is assumed that you have already had a course in elementary statistics, and now need to apply more advanced statistical methods. (See ‘A Note on Prerequisite Material’, below.) No mathematical theory will be developed, but intensive computer use is required. By the end of this course, you should be able to choose appropriate statistical techniques for the most common experimental situations, use SPSS to perform the appropriate analyses, and interpret the results to a reader. We will briefly review prerequisite material on one-sample, paired and two-sample t-tests. Among the statistical techniques we will consider in more detail are analysis of contingency tables, the one-way analysis of variance including multiple comparisons, the two-way and higher order analysis of variance, simple and multiple regression. If time allows, we will add a brief introduction to repeated measures experiments. In addition to homework problems and exams, you will be required to undertake several statistical analyses of real data sets on the computer, describing the methods and results in a formal report. Text: I will have copies of my course notes posted to the course's Web page. These notes contain extensive explanations of the heuristics behind the formulas, homework problems, and numerous computer examples. In addition, I will provide handouts on the use of SPSS. No other text will be required. However, if at all possible, you should use the money you save by not purchasing a text to buy a copy of SPSS for your home computer. See the document SPSSinfo.pdf on the Blackboard site. While the course notes on the Web page are available for your use, they are my intellectual property and should not be cited without appropriate reference. Calculators: You must have a decent calculator and you must bring it to class. Your cell phone will not be enough. It must be a real scientific calculator, able to do sample means and standard deviations, but it does not have to be a graphing calculator. The TI-83/84 series are excellent but pricey (~$100). The TI-30 series are quite cheap and will do a lot of good stuff (under $25). Look for symbols like STAT, x or on the keyboard. Those are good signs that the calculator will do simple statistics. Optional Reference Texts. You should certainly have a copy of an elementary statistics text for basic materials. The one you used in your undergraduate elementary course will certainly suffice. If you got rid of that book, I may have some old ones in my office. First come, first serve. Some people say that they would like to have a real reference text to read or use later. In that case, I suggest you go on line and find a recent version (4th edition or later) of David Howell’s Statistical Methods for Psychology, which is quite complete. Course Format: The course will meet twice a week for lecture and discussion. An important part of every class will be the presentation of examples illustrating statistical techniques, and comparison of those techniques to others so far discussed. Attendance is important. Grading: Grade will be based on 1) Two mid-term exams, each worth 20% of the semester grade 2) Final exam, worth 20% of the semester grade 3) Reports - 25% of the semester grade. There will be 4 written reports in which you analyze data sets using techniques studied and give the results in a formal report. 4) Homework problems - 15% of grade. Problems will be assigned from each chapter of the notes, along with several review assignments. Problems more than a few hours late will normally not be accepted. Exam Dates Exam 1 Thursday September 29 Exam 2 Thursday November 3 Final exam Thursday December 9 from 9am – 10:50 am CONTENTS: Chapter 1: PREREQUISITE - ONLY BRIEF REVIEW IN CLASS. Review of elementary terminology: populations and samples, parameters and statistics, quantitative and qualitative variables, sampling variability, ‘provability’ of an effect, simple t-test and p-values. Chapter 2: PREREQUISTE – ONLY BRIEF REVIEW IN CLASS. Some truly useful basic tests for quantitative variables: random samples, independent and dependent variables, paired t-tests, independent sample t-tests, F-test and Levene test to compare variances, normality assumption. Chapter 3: Some truly useful basic tests for qualitative variables: Z-test for proportion in one group, z-test comparing proportions in two groups, Fisher’s exact test, Chi-squared test comparing multiple groups, independence. Chapter 4: Comparing means in several groups, the One-Way Analysis of Variance: hypotheses in one-way ANOVA, graphical displays, notation and data layout, sums of squares, mean squared error, the F-test statistic, normal probability plots. Chapter 5: Advanced topics in the One-Way ANOVA: Multiple comparisons, family/experimentwise significance, Bonferroni Inequalities, Tukey’s Honestly Significant Differences, Scheffe’s contrasts, posthoc versus apriori comparisons. Chapter 6: (brief coverage) When the assumptions for the one-way ANOVA look shaky: Levene’s test for homogeneity of variances, Kolmogorov-Smirnov statistic, transformations of variables, nonparametric alternatives to the one-way ANOVA. Chapter 7: The Two-Way Analysis of variance: graphical summaries, data layout and notation, mathematical model, main effects and interactions, development of the F-test. Chapter 8: Advanced Topics in the two-way ANOVA: Types of sums of squares in computer packages, multiple comparisons with and without interactions, R-square. Chapter 9: (brief coverage) Examples of higher-order ANOVA: Interpreting interactions, fitting less than the largest possible model. Chapter 10: Simple linear regression: algebraic fundamentals, scatterplots, fitting the best straight line, testing hypotheses about the parameters. Chapter 11: Multiple regression: the multiple linear regression model, interpreting regression coefficients, F-tests and the overall F-test, t-tests for individual coefficients, R-square, indicator variables for qualitative variables, interactions. Chapter 12: Logistic regression: a version of regression when the dependent variable is ‘success/failure’. This option recommended for students in criminal justice, sociology, or other non-experimental fields. Chapter 13: Repeated measures ANOVA: within-subjects experiments. This option recommended for students in psychology. A NOTE ON PREREQUISITE MATERIAL This course assumes that you have had an elementary course in statistics. At UNF, the course might be STA 2014 – Elementary Statistics for the Health and Social Sciences, or STA 2023 – Elementary Statistics for Business. The precise course does not matter. Pretty much any college level introductory statistics course will suffice – so long as you remember what was taught you. In order to cover the kinds of statistics graduate students have to know, we will review certain prerequisite topics very quickly. I will lecture on the most important ideas, but only briefly. You are primarily responsible for re-teaching yourself this material. TO HELP YOU IN YOUR REVIEW, I HAVE PROVIDED the following material, at http://www.unf.edu/~dmohr/sta5126/. It is also on our Blackboard site. Chapter 1 of my course notes, covering elementary terminology, basic hypothesis testing, and basic confidence intervals for means. The un-starred problems at the end of this chapter will be due Tuesday 8/30/2011, second week of class. Chapter 2 of my course notes, covering two-sample and paired t-tests. The un-starred problems at the end of this chapter will be due Tuesday 9/6/2011, third week of class. A table of the t-distribution in EXCEL format. TO HELP YOU IN YOUR REVIEW, YOU SHOULD OBTAIN the following material: An elementary statistics text. Any old introductory text will do, it does not have to be very recent. If you have your old text, great. If you can find one second-hand, OK. I have a few in a stack in my office, you are welcome to one, first come first served. I will not assume you have ever used SPSS. We will cover a basic tutorial in class. PREREQUISITE TOPICS The distinction between populations and samples The distinction between parameters and statistics The interpretation of the mean and standard deviation The calculation of the sample mean and sample standard deviation The empirical rule Boxplots Null and alternative hypothesis Significance levels and p-values The hypothesis test for the mean of a population: the t-test A confidence interval for the mean of a population: the t-interval Comparing the means in two dependent samples: the paired t-test Comparing the means in two independent samples: the two-sample t-test Assumptions of t-tests Comparing variances in two independent samples