Presentation for Southwestern Psychological Association meeting

San Antonio, TX

April 9, 2004

Best Practices for Teaching Statistics

David C. Howell

Introduction

More current opportunities

• Sources of alternative explanations

  • Textbooks and explanations
    • No one text will explain everything in a way that everyone will understand it
    • There are a number of ways that students can get quick explanations from a different perspective
      • VassarStats
        • Richard Lowry
        • http://faculty.vassar.edu/lowry/VassarStats.html 
        • Wealth of material in a variety of formats
        • A very useful text companion section called Concepts and Applications of Inferential Statistics
      • Surfstat--an Australian site
        • http://math.uc.edu/statistics/surfstat/surfstat-main.html
        • Very nice explanations of concepts of descriptive and inferential statistics
          • Click on Detailed Contents
        • Glossary of statistical terms
      • William Trochim   at Cornell
        • http://trochim.human.cornell.edu/kb/contents.htm
        • An excellent on-line text for alternative explanations
      • David Lane--Rice
        • http://davidmlane.com/hyperstat/index.html
        • Hyperstat is an excellent source of short explanations and demonstrations
      • David Howell at the University of Vermont
        • http://www.uvm.edu/~dhowell/StatPages/StatHomePage.html
        • This site is not set up as an online text
        • It contains individual pages on a) Questions that I have been asked, b) New material, c) topics that have caused confusion for students
        • It also contains a very complete set of lecture notes (now two years old) for a graduate level course on statistical methods in psychology
      • Google.com
        • This is the best way I know to find an explanation for a statistical concept that you don't understand.
        • Useful for students at any level.
          • e.g. I typed in "confidence intervals" (in quotation marks) an in 0.17 seconds I found approximately 365,000 links.
          • It took 0.21 seconds to find about 3800 links to "median splits."

           

  • Glossaries
    • These are very useful to students who are studying for an exam and don't understand important terms
    • STEPS website
      • http://www.stats.gla.ac.uk/steps/glossary/index.html
      • Very complete
    • Surfstat website
      • sited above
      • http://math.uc.edu/statistics/surfstat/surfstat-main.html
    • Dave Howell's website
      • sited above
      • http://www.uvm.edu/~dhowell/StatPages/StatHomePage.html

    • David Stockburger, Southern Methodist University

      • http://www.psychstat.smsu.edu/introbook/sbk00.htm

      • Not really a glossary, but can be used in that way.

      • Excellent source of brief descriptions of statistical topics. For example, if a student doesn't understand what a sampling distribution is, she can click on a link in the table of contents and go immediately to a definition and a simulation. And the definition is really more of an explanation of what a sampling distribution is and how it differs from a simple sample distribution.

    • Many more sites can be found with a simple Google search
      

• Data sources 

  • Data and Story Library (DSL)
    • http://lib.stat.cmu.edu/DASL/
    • A truly great site for real data on problems that students may find interesting.
  • StatLib
    • http://lib.stat.cmu.edu/datasets/
    • This is a huge collection of data sets--some from texts
  • Robin Lock at St. Lawrence University
    • http://it.stlawu.edu/~rlock/datasurf.html
    • This is a collection of links to all sorts of databases
  • Chance News
    • http://www.dartmouth.edu/~chance/
    • This is one of my favorite sources for examples, though I don't believe that their database is still available.

       

• Java applets for demonstrations

   

  • Gary McClelland --Seeing Statistics (Univ. Colorado Boulder)
    • Available with my Fundamentals book
  • Duke Statistics Department
    •   There is a collection of applets at http://www.stat.duke.edu/sites/java.html that have been compiled by someone, but I don't know who.
    • Within this is a nice set of applets from NPW Associates
      • http://www.statlets.com/
  • David Lane, at Rice, has a very nice set of simulations
    • http://www.ruf.rice.edu/~lane/stat_sim/index.html
    • See, for example, the simulations on confidence intervals.
      • This is a nice way for students to understand the idea that confidence limits cover the parameter a specified percentage of the time.
  • Uses
    • Some of these are very useful for demonstrations in class or as a link to course notes
    • Some are appropriate as assignments
    • Some are appropriate for students to do on their own

• Statistical and probability calculators

  • UCLA Statistics department
    • http://calculators.stat.ucla.edu/
    • Probably the most extensive set of calculators I know
    • I really don't need to give any more sites--this is so complete
      • Instructors may find particular use for http://calculators.stat.ucla.edu/cdf/

         

• Web pages that perform statistical calculations

  • There are a lot of these, but I'll only mention one. 
    • John Pezzullo at Georgetown

    • http://m2.aol.com/johnp71/javastat.html

    • This page has a lot of great stuff, though he is a biostatistician and some of his material is tangential to psychology.

• Data Collection

  • CogLab
    • This is a set of Java applets that allows the student to collect data on many different research designs in psychology
    • Available at http://coglab.wadsworth.com , but requires subscription.
      • Demo version is probably available from Wadsworth
    • Ideal for a course that combines experimental psychology and statistics
    • Data can be exported to SPSS (and SAS and others) and combined across participants or years.

     

David C. Howell 
Last revised: April 5, 2004