a new approach to introductory statistics nathan tintle hope college
TRANSCRIPT
A new approach A new approach to introductory to introductory
statisticsstatisticsNathan TintleNathan Tintle
Hope CollegeHope College
OutlineOutline
Case study: Hope College the past Case study: Hope College the past five yearsfive years
A completely randomization-based A completely randomization-based curriculumcurriculum
The bigger pictureThe bigger picture
Case study: Hope CollegeCase study: Hope College
Five years agoFive years ago 2 courses: algebra-based and calculus-based intro 2 courses: algebra-based and calculus-based intro
statsstats 3 hours of lecture with graphing calculator use; 1 3 hours of lecture with graphing calculator use; 1
hour of computer lab work (algorithmic type labs)hour of computer lab work (algorithmic type labs) Process for changeProcess for change
Curricular changeCurricular change Pedagogical changePedagogical change Infrastructure changeInfrastructure change Client discipline buy-inClient discipline buy-in Math department buy-inMath department buy-in
Case study: Hope CollegeCase study: Hope College Where we are now: Where we are now:
Three coursesThree courses Algebra-based intro statsAlgebra-based intro stats Accelerated intro stats (for AP Stats students and Accelerated intro stats (for AP Stats students and
others)others) Second course in stats (multivariable topics)Second course in stats (multivariable topics) Note: NO Calculus pre-requisite’sNote: NO Calculus pre-requisite’s
New dedicated 30-seat computer lab for New dedicated 30-seat computer lab for statistics (HHMI funded)statistics (HHMI funded)
Buy-in of relevant partiesBuy-in of relevant parties Revolutionary new curriculumRevolutionary new curriculum
Embrace the GAISE pedagogy: active learning, Embrace the GAISE pedagogy: active learning, concept based, real dataconcept based, real data
Changes in contentChanges in content
Content changesContent changes
George Cobb, USCOTS 2005George Cobb, USCOTS 2005 A challengeA challenge
Rossman and Chance 2007 NSF-Rossman and Chance 2007 NSF-CCLI grantCCLI grant ModulesModules
Hope College 2009Hope College 2009 Entire curriculumEntire curriculum
Traditional curriculumTraditional curriculum
Unit 1. Descriptive statistics and Unit 1. Descriptive statistics and sample designsample design
Unit 2. Probability and sampling Unit 2. Probability and sampling distributionsdistributions
Unit 3. Statistical inferenceUnit 3. Statistical inferenceNo multivariable topics;
No second course in statistics without calculus
Curriculum outlineCurriculum outline
Unit 1. (1Unit 1. (1stst course) course) Introduction to inferential statistics Introduction to inferential statistics
using randomization techniquesusing randomization techniques Unit 2. (1Unit 2. (1stst course) course)
Revisiting statistical inference using Revisiting statistical inference using asymptotic approaches, confidence asymptotic approaches, confidence intervals and powerintervals and power
Unit 3. (2Unit 3. (2ndnd course) course) Multivariable statistical inference: Multivariable statistical inference:
Controlling undesired variabilityControlling undesired variabilityRandomization techniques=Resampling techniques=permutation tests
Unit 1.Unit 1.
Ch 1. Introduction to Statistical Ch 1. Introduction to Statistical Inference: One proportionInference: One proportion
Ch 2. Comparing two proportions: Ch 2. Comparing two proportions: Randomization MethodRandomization Method
Ch 3. Comparing two means: Ch 3. Comparing two means: Randomization MethodRandomization Method
Ch 4. Correlation and regression: Ch 4. Correlation and regression: Randomization MethodRandomization Method
Unit 2.Unit 2.
Ch 5. Correlation and regression: revisited*Ch 5. Correlation and regression: revisited* Ch 6. Comparing means: revisited*Ch 6. Comparing means: revisited* Ch 7. Comparing proportions: revisited*Ch 7. Comparing proportions: revisited* Ch 8. Tests of a single mean and proportionCh 8. Tests of a single mean and proportion
*Connecting asymptotic tests with the *Connecting asymptotic tests with the randomization approach, confidence randomization approach, confidence intervals and powerintervals and power
Unit 3.Unit 3.
Chapter 9: Introduction to multiple Chapter 9: Introduction to multiple regression (ANCOVA/GLM)regression (ANCOVA/GLM)
Chapter 10: Multiple logistic Chapter 10: Multiple logistic regressionregression
Chapter 11: Multi-factor Chapter 11: Multi-factor experimental designexperimental design
Key ChangesKey Changes
Descriptive statisticsDescriptive statistics Only select topics are taught (e.g. Only select topics are taught (e.g.
boxplots); other topics are reviewed boxplots); other topics are reviewed (based on assessment data; CAOS)(based on assessment data; CAOS)
Study designStudy design Discussed from the beginning and Discussed from the beginning and
emphasized throughout in the context emphasized throughout in the context of its impact on inferenceof its impact on inference
Key ChangesKey Changes
InferenceInference Starts on day 1; in front of the students Starts on day 1; in front of the students
throughout the entire semesterthroughout the entire semester
Probability and Sampling Probability and Sampling distributionsdistributions More intuitive approach; de-emphasized More intuitive approach; de-emphasized
dramaticallydramatically
Key other changesKey other changes CyclingCycling
ProjectsProjects
Case studiesCase studies
Research ArticlesResearch Articles
PowerPower
Example from the Example from the curriculumcurriculum
Chapter 2Chapter 2 (pdf is available at (pdf is available at
http://math.hope.edu/aasihttp://math.hope.edu/aasi) )
AssessmentAssessment
CAOSCAOS Better learning on inferenceBetter learning on inference Mixed results on descriptive statisticsMixed results on descriptive statistics Increased retention (4-month follow-up)Increased retention (4-month follow-up)
Big pictureBig picture
ModularityModularity Advantages: broader impact; flexibilityAdvantages: broader impact; flexibility Disadvantages: can’t fully realize the Disadvantages: can’t fully realize the
potential of a randomization-based potential of a randomization-based curriculumcurriculum Efficiency of approach allows for cycling Efficiency of approach allows for cycling
over core concepts, quicker coverage of over core concepts, quicker coverage of other topics and additional topics are other topics and additional topics are possiblepossible
Big pictureBig picture
Resampling methods in generalResampling methods in general Permutation tests: Not only a valuable Permutation tests: Not only a valuable
technique practically, but a motivation for technique practically, but a motivation for inferenceinference
Bootstrapping?Bootstrapping?
Keeping the main thing the main thingKeeping the main thing the main thing Core logic of statistical inference (Cobb Core logic of statistical inference (Cobb
2007)2007)
Big PictureBig Picture
Motivating concepts with practical, Motivating concepts with practical, interesting, relevant examplesinteresting, relevant examples Capitalizing on students intuition and interestCapitalizing on students intuition and interest Real, faculty and/or student-driven, research Real, faculty and/or student-driven, research
projectsprojects Danny’s example translated to the traditional Danny’s example translated to the traditional
Statistics curriculumStatistics curriculum One sample Z TestOne sample Z Test Calculating probabilities based on the central limit Calculating probabilities based on the central limit
theoremtheorem Art and science of learning from data (Agresti Art and science of learning from data (Agresti
and Franklin 2009)and Franklin 2009)
Big PictureBig Picture
Confidence intervalsConfidence intervals Ranges of plausible values under the Ranges of plausible values under the
null hypothesisnull hypothesis ““Invert” the test to get the confidence Invert” the test to get the confidence
intervalinterval
PowerPower Reinforcing logic of inferenceReinforcing logic of inference Practical toolPractical tool
Big PictureBig Picture
The second courseThe second course Projects can be student driven or Projects can be student driven or
involve students working with faculty in involve students working with faculty in other disciplinesother disciplines
Other effortsOther efforts CATALSTCATALST West and WoodardWest and Woodard Rossman and ChanceRossman and Chance OthersOthers
Textbook websiteTextbook website
http://math.hope.edu/aahttp://math.hope.edu/aasisi
-First two chapters-First two chapters-Email me for copies of -Email me for copies of
other chaptersother chapters-If interested in pilot -If interested in pilot
testing, please talk to metesting, please talk to me-Draft of paper in revision -Draft of paper in revision
at the Journal of Statistics at the Journal of Statistics Education is available Education is available (assessment results)(assessment results)
AcknowledgementsAcknowledgements
FundingFunding Howard Hughes Medical Institute Undergraduate Howard Hughes Medical Institute Undergraduate
Science Education Program (Computer lab, pilot Science Education Program (Computer lab, pilot testing and initial curriculum development)testing and initial curriculum development)
Great Lakes College Association (Assessment and Great Lakes College Association (Assessment and first revision)first revision)
Teagle Foundation (second revision this summer)Teagle Foundation (second revision this summer) Co-authors: Todd Swanson and Jill Co-authors: Todd Swanson and Jill
VanderStoepVanderStoep Others: Allan Rossman, Beth Chance, George Others: Allan Rossman, Beth Chance, George
Cobb, John Holcomb, Bob delMasCobb, John Holcomb, Bob delMas