Transcript Materials
‘The’ Second Course in Statistics Robin Lock, Burry Professor of Statistics St. Lawrence University, [email protected] Dick De Veaux, Williams College [email protected] Breakout Session at USCOTS07 Positioning the “Second” Course (in the “good old days”) Calculus I Calculus II Calculus III Intro Stat Linear Alg. Probability Math Stat Current Consensus Intro Stat ????? AP Stat GAISE: http://www.amstat.org/education/gaise/ AP: http://apcentral.collegeboard.com What (could be, should be, might be, is) the (or a) second course in statistics? Mathematics Proofs One Variable Calculus MultiVariable Calculus Linear Alg. Discrete Diff. Eq’s Economics Financial Micro Intro Economics Industrial Macro Labor Public Chemistry General Organic Physical Psychology Behavioral I N T R O Developmental Social Physiological Abnormal Borrowing from Other Disciplines Math: In the second course, repeat the topics in the first course in a multivariable setting Stat: Multiple regression, multi-factor ANOVA, … Economics: Have two main second courses that build on the ideas of the first course Stat: Experimental Design & Applied Regression Borrowing from Other Disciplines Chemistry: Develop a single sequence Stat: Might require redesign of intro Psychology: Have lots of “second” courses that expand on different aspects of the first course. Stat: Experimental Design, Sampling, Applied Regression, Nonparametric Methods, Categorical Unique to Statistics Modeling approach: Data Response Variable = = Model + Error F(Predictors & Explanatory factors) + “Unexplained” Variability Categorical vs. Quantitative Approach Predictor(s) Quantitative Categorical Quantitative Response Categorical Multiple Regression Multifactor ANOVA Logistic Regression Loglinear Models Data Production Approach Experimental Design Expand on designs from intro Randomized, factorial, block, … Analyzing designed data Sampling Beyond the SRS Cluster, stratified, … Estimates based on sample method What if the Standard Method Doesn’t Work? Approach Data not normal? Nonparametric methods Bootstrap CI’s and Permutation Tests Transformations Errors not independent? Time series analysis Questions? • • • • • • How do we attract students to take Stat II? What should we assume from Stat I? What about software/technology? Can we still use activities and explorations? Do we have to assume more math background? Is there a good textbook? • What will students do after Stat II? Create Your Own Stat II Divide into groups - perhaps by similar institutions or ideas for a second course Develop a syllabus for a second course – by consensus or with minority reports Assume A GAISE compliant Stat I Prepare to report back to the rest of the group.