FRS 001E - Sec. 001 - (1 unit) - CRN 75077 - M 4:40 – 6:00 pm - 3 Wellman
How to Lie with Statistics – In the Social and Behavioral Sciences
Instructor: Keith Widaman, Department of Psychology, College of Letters and Sciences
Description: The course content centers on the interpretation of data, particularly data presented in the media (newspapers, television). A great deal of public policy is debated over the airwaves or in newspapers, and informed citizens should develop skills for interpreting such data. Thus, the content of the course is the presentation of data, the presentation of information about the study (which is crucial to any interpretation of data), and the conclusions drawn from data. The primary goals for student learning include goals in several domains. With regard to knowledge, the course should increase the student's knowledge of principals of proper presentation of data and the ways in which presenters can bias perceptions of readers or viewers by using various tricks. With regard to attitudes, students should develop a critical attitude toward presentations of data, leading to an outlook that one cannot trust everything that is written in papers or presented on TV. But, special care will be taken to develop optimal levels of criticism; every study ever conducted has flaws, and the course will attempt to instill in students the ability to distinguish fatal flaws from among the flaws existing in a study. Such instruction should lead to the awareness by students that crucial information for interpretation of data is often missing, so that there are risks in relying on typical forms of data presentation for important societal decisions. Finally, the course should hone the skills of students in understanding how to present data and, importantly, how to interpret the conclusions offered by researchers and journalists.
Format: Class will meet for one 1.5 hour session each week for the first 8 weeks of the quarter. In addition to this time commitment, students will spend 2 hours per week reading major newspapers (not the campus newspaper) to (a) find articles reporting data and (b) write a single critique of the way in which data were presented (see below). The text used for the course will be the classic text "How to Lie with Statistics," written by Gerald Huff. This book, originally published in 1954, is considered a classic on the topic and its illustrations are still very current in content and importance. Students will read one chapter of the Huff book each week, and each student will co-lead the discussion during one of the class meetings. With a single course meeting per week, each class meeting will have 2 or 3 students selected to co-lead discussion (with me) of the topics covered in the chapter of the text. Grading: Course grades will be assigned on the following basis: 30% general participation in class, 30% on the oral presentation of material (when students co-lead discussion), and 40% on term project. The term project will consist of a notebook including (a) Xerox copies of at least 10 articles from major newspapers that report data, and (b) a 5-+page paper in which the student will select one of these articles and provide a critique of the strengths and weaknesses of the ways in which data were presented and then interpreted.
About the Instructor:
Professor Keith Widaman is a member of the faculty in the Department of Psychology. His research interests include the development of human mental abilities, life-span development, and statistical techniques used to represent longitudinal data. Current projects include a review of methods used for testing correlations and modeling mental abilities of children of mothers with PKU. He currently teaches a course that provides an introduction to statistical techniques used in psychology and the behavioral sciences.