Webinars

This webinar presents the development of a new instrument designed to assess the practices and beliefs of teachers of introductory statistics courses. The Statistics Teaching Inventory (STI) was developed to be used as a national survey to assess changes in teaching over time as well as for use in evaluating professional development activities. We will describe the instrument and the validation process, and invite comments and suggestions about its content and potential use in research and evaluation studies.

Ellen and Nancy use Calibrated Peer Review, an online writing and peer evaluation program available from UCLA, to introduce statistical literacy to Nancy's freshman biology students and to bring a real-world context to statistical concepts for Ellen's introductory statistics classes in an NSF-funded project. CPR allows instructors in large classes to give their students frequent writing assignments without a heavy grading burden. Ellen and Nancy have their students read research journal articles on interesting subjects and use guiding questions to evaluate these articles for statistical content, experimental design features, and ethical concerns.

What's the best way to introduce students of mathematics to statistics? Tradition offers two main choices: a variant of the standard "Stat 101" course, or some version of the two-semester sequence in probability and mathematical statistics. I hope to convince participants to think seriously about a third option: the theory and applications of linear models as a first statistics course for sophomore math majors. Rather than subject you to a half-hour polemic, however, I plan to talk concretely about multiple regression models and methodological challenges that arise in connection with AAUP data relating faculty salaries to the percentage of women faculty, and to present also a short geometric proof of the Gauss-Markov Theorem.

Permutation tests and randomization tests were introduced almost a century ago, well before inexpensive, high-speed computing made them feasible to use. Fisher and Pitman showed the two-sample t-test could approximate the permutation test in a two independent groups experiment. Today many statistics educators are returning to the permutation test as a more intuitive way to teach hypothesis testing. In this presentation, I will show an interesting teaching example about primate behavior that illustrates how simple permutation tests are to use, even with a messier data set that admits of no obvious and easy-to-compute approximation.

Our research shows that half of the students entering a statistics course use the word random colloquially to mean, "haphazard" or "out of the ordinary." Another large subset of students define random as, "selecting without prior knowledge or criteria." At the end of the semester, only 8% of students we studied gave a correct statistical definition for the word random and most students still define random as, "selecting without order or reason." In this session we will present a classroom approach to help students better understand what statisticians mean by random or randomness as well as preliminary results of the affect of this approach.

In introductory statistics courses, web-based applets are often used to visually conduct large simulation studies illustrating statistical concepts. However, it is difficult to determine what (if anything) students learn from repeatedly pressing a button when using applets. More advanced options such as writing/running computer code are typically considered to be much too advanced for most introductory courses. The web-based software package, StatCrunch, now offers simulation capabilities that strike a middle ground between these two extremes. The instructor/student needs only to perform a small number of steps using the menu driven interface with each step being key to understanding the underlying data structure. This talk will cover the steps required to study concepts such as the central limit theorem, confidence intervals, hypothesis testing and regression using StatCrunch.

This webinar will introduce several sources of data and tools that could be useful in both general and social science-specific statistics instruction. The Social Science Data Analysis Network (SSDAN) and the Inter-university Consortium for Political and Social Research (ICPSR), both a part of the University of Michigan's Institute for Social Research, are collaborating on two NSF-funded projects to support quantitative literacy in the social sciences. Resources from each organization and TeachingWithData.org, a result of the partnership, will be highlighted. Materials range from small extracts of data from the Census and American Community Surveys used with specific teaching modules to full datasets with accompanying online analysis tools.

This webinar will present data, tools, materials and the pedagogical approach of the Statistics Online Computational Resource (SOCR) for technology-enhanced probability and statistics education. Following a review of the different types of SOCR online resources, we will go over two specific classroom utilization examples. The first one provides a hands-on demonstration of a statistical concept (CLT) using interactive virtual experiments and simulations. The second example will showcase the use of SOCR resources to address interesting social, health, environmental, scientific, and engineering challenges. In this case, we'll focus on the Ozone pollution in California, formulate health-related hypotheses, identify appropriate data and employ web-based exploratory and statistical data analysis tools. What is www.SOCR.ucla.edu? The Statistics Online Computational Resource provides portable online aids for probability and statistics education, technology based instruction and statistical computing. SOCR tools and resources include a repository of interactive applets, computational and graphing tools, instructional and course materials. SOCR aims to develop new Java applets, design diverse extensible SOCR learning activities, develop XML/HTML navigation/search tools for interactive materials, and validate and assess technology-enhances pedagogical techniques. SOCR Products Tools/Applets: Distributions, Experiments, Analyses, Games, Modeler & Graphs. Multilingual instructional resources: EBooks, continuing statistics education workshops/seminars Learning activities: interactive, data-driven and technology-enhanced learning activities Examples: Central Limit Theorem Hands-on California Ozone Data Activity Data: Diverse publicly accessible datasets for copy-paste/download utilization Example: Latin Letters Frequency Distribution Dissemination: papers, conferences, workshops, etc. SOCR Evaluation and Efficacy We have conducted several control-based studies of the efficacy of technology-enhanced statistics education. Using IRB-approved studies, quantitative and qualitative measures of student performance were recorded in classes using traditional (control) instruction (R or Stata based) and classes using SOCR resources and tools. Non-parametric analyses of the data showed very statistically significant (SOCR) treatment effects (p < 10-4) on student performance and perception of the material. The practical significance of these treatment effects were more modulated. More details about these studies are available here. Summary Main SOCR server, applets Data, activities and EBooks Feedback and Forum Graphical SOCR Navigator

This year, Jeanne and Bill assumed co-editorship of the Chance News Wiki, which as of March 15 will be moving to CAUSEweb. The Wiki provides reviews of current news stories that are relevant to teaching statistics and probability, along with links to original articles and related resources. This webinar will describe the various ways that Chance project materials have been used, in areas ranging from traditional introductory statistics to statistical literacy courses to first-year seminars. We will also discuss the mechanics of posting to the Wiki, and hope to inspire some new contributors.

This is a CAUSE Special Presentation for USCOTS Research Cluster members.