pith. sign in

arxiv: 2606.08660 · v1 · pith:XCREFXKDnew · submitted 2026-06-07 · 📊 stat.AP · stat.ME· stat.OT

Active Learning with Bayesian Reasoning: A POGIL-Based Pedagogy in Introductory Statistics

Cheng-Han Yu , Angela Ebeling This is my paper

Pith reviewed 2026-06-27 17:42 UTC · model grok-4.3

classification 📊 stat.AP stat.MEstat.OT
keywords active learningPOGILBayesian reasoningBayes theoremintroductory statisticsconditional probabilityquasi-experimental studybelief updating
0
0 comments X

The pith

A POGIL-style activity teaches Bayesian reasoning in introductory statistics with performance similar to lectures.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces a Process Oriented Guided Inquiry Learning activity that uses two-way tables and hand-computable probabilities to teach conditional probability, Bayes theorem, and belief updating through structured student teams. It evaluates the activity in a quasi-experimental study across four sections of an undergraduate intro stats course, comparing it to lecture-based instruction while adjusting for major, gender, and race via a Bayesian bivariate generalized linear model. The results show similar exam performance on Bayes theorem questions and similar rates of high satisfaction, with no clear evidence of differences. A sympathetic reader would care because the activity offers an active, classroom-ready alternative that avoids simulation or heavy computation, supporting the inclusion of Bayesian ideas in standard intro courses.

Core claim

The POGIL-style activity performed comparably to lecture-based instruction for the Bayes theorem unit, producing similar exam performance and similar probabilities of high satisfaction across instructional styles and demographic groups, with considerable uncertainty and no clear evidence of meaningful differences.

What carries the argument

The POGIL-style activity using structured team roles and two-way tables for hand-computable probabilities to introduce Bayes theorem and belief updating.

If this is right

  • Instructors gain a self-contained, adaptable set of materials for adding Bayesian reasoning to introductory statistics without advanced computation.
  • The activity supports feasible inclusion of Bayes theorem in standard intro courses through active team-based work.
  • The Bayesian bivariate generalized linear model provides a reproducible framework for comparing active learning approaches while accounting for demographics.
  • Similar performance across demographic groups suggests the activity does not introduce new inequities in outcomes for this unit.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same team-role and table-based structure could be adapted to teach other probability topics in intro stats.
  • Future work could examine longer-term retention or transfer of Bayesian reasoning skills beyond the immediate exam questions.
  • The reported uncertainty leaves room for detecting small practical differences in larger samples or different course contexts.
  • This approach may reduce reliance on technology in probability instruction while maintaining active engagement.

Load-bearing premise

The quasi-experimental comparison assumes the Bayesian model adequately controls for all relevant differences between sections beyond the measured factors of major, gender, and race.

What would settle it

A randomized trial across multiple institutions that finds significantly different exam scores or satisfaction rates between the POGIL activity and lecture instruction would falsify the comparability result.

Figures

Figures reproduced from arXiv: 2606.08660 by Angela Ebeling, Cheng-Han Yu.

Figure 1
Figure 1. Figure 1: Changes in Introductory Statistics. Each period shows the main focus and [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Screenshot excerpts from the three models in the POGIL-style activity. [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Flow chart comparing the two pedagogical approaches used in this study. [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Distributions of the response variables by instructional condition. Left: Bayes’ [PITH_FULL_IMAGE:figures/full_fig_p017_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: 95% credible intervals for model-based posterior differences from the retained [PITH_FULL_IMAGE:figures/full_fig_p021_5.png] view at source ↗
read the original abstract

We introduce a Process Oriented Guided Inquiry Learning (POGIL)-style activity for teaching Bayesian reasoning in introductory statistics through conditional probability, Bayes' theorem, and belief updating. The activity is self-contained, uses hand-computable probabilities organized in two-way tables, and engages students in structured team roles. We evaluated the activity in four sections of an undergraduate introductory statistics course using a quasi-experimental comparison of POGIL-style and lecture-based instruction for a Bayes' theorem unit. Outcomes included student performance on Bayes' theorem final exam questions and satisfaction with instruction. We used a Bayesian bivariate generalized linear model to compare the two approaches while accounting for major type, gender, and race. The results indicated similar exam performance and similar probabilities of high satisfaction across instructional styles and demographic groups, with considerable uncertainty and no clear evidence of meaningful differences. These findings suggest that the POGIL-style activity performed comparably to lecture-based instruction for this unit while offering an active and classroom-ready way to introduce Bayesian reasoning without requiring difficult computation or simulation. We provide adaptable instructional materials and a reproducible Bayesian analytic framework for evaluating active learning innovations in introductory statistics. Our study supports the feasible inclusion of Bayesian reasoning in introductory courses and may help instructors considering active learning.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 2 minor

Summary. The paper introduces a POGIL-style activity for teaching Bayesian reasoning via conditional probability and Bayes' theorem using hand-computable two-way tables and structured team roles. It reports a quasi-experimental evaluation across four sections of an introductory statistics course, comparing POGIL to lecture-based instruction on final-exam Bayes' theorem performance and instructional satisfaction. A Bayesian bivariate generalized linear model is used to compare outcomes while conditioning on major type, gender, and race; results indicate similar performance and satisfaction probabilities across styles with considerable uncertainty and no clear differences. The authors conclude the activity is a viable, classroom-ready alternative and provide adaptable materials plus reproducible analysis code.

Significance. If the equivalence result holds after addressing design limitations, the work supplies a concrete, low-computation active-learning module for introducing Bayesian updating in service courses, together with open materials and a reproducible Bayesian evaluation framework. These elements directly support the feasible inclusion of Bayesian reasoning in introductory statistics and offer a template for assessing other pedagogical interventions.

major comments (1)
  1. [Methods (statistical analysis)] Methods (statistical model description): The Bayesian bivariate GLM conditions only on major type, gender, and race. In a quasi-experimental section comparison this leaves open confounding paths from unmeasured section-level factors (instructor fixed effects, time-of-day, differential attrition, or prior probability exposure) that could also affect both exam scores and satisfaction; the central claim of comparable performance therefore rests on an untested no-confounding assumption that is not load-bearing only if the design were randomized.
minor comments (2)
  1. [Abstract] Abstract: reports 'similar performance' and 'considerable uncertainty' but omits sample sizes per section, exact link functions and priors in the bivariate GLM, and any effect-size or posterior-probability summaries, making it difficult to gauge the precision of the equivalence conclusion.
  2. [Results] Results: the manuscript should report the effective sample size after any data exclusion, the full posterior summaries (means, 95% credible intervals) for the instructional-style coefficients on both outcomes, and a sensitivity analysis that adds plausible unmeasured confounders.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their constructive comments on our manuscript. The major comment identifies a key limitation of the quasi-experimental design, which we address below by expanding the discussion of assumptions and potential confounders while preserving the reported findings as preliminary evidence.

read point-by-point responses

  1. Referee: The Bayesian bivariate GLM conditions only on major type, gender, and race. In a quasi-experimental section comparison this leaves open confounding paths from unmeasured section-level factors (instructor fixed effects, time-of-day, differential attrition, or prior probability exposure) that could also affect both exam scores and satisfaction; the central claim of comparable performance therefore rests on an untested no-confounding assumption that is not load-bearing only if the design were randomized.

    Authors: We agree that the study is quasi-experimental and that unmeasured section-level factors (instructor, time-of-day, attrition, prior exposure) could confound results; the no-confounding assumption is not guaranteed by randomization. With only four sections it is not statistically feasible to add instructor fixed effects without severe overfitting or loss of identifiability. We have revised the Methods and Discussion sections to state this assumption explicitly, to enumerate the unmeasured factors, and to frame conclusions as absence of clear evidence of difference rather than proven equivalence. The Bayesian model already reports substantial posterior uncertainty, which we now highlight as reflecting both sample size and design limitations. These changes constitute a partial revision; the core data and model remain unchanged because the design cannot be altered retrospectively. revision: partial

Circularity Check

0 steps flagged

No circularity: empirical comparison rests on observed data and standard statistical adjustment.

full rationale

The paper reports a quasi-experimental evaluation of POGIL vs. lecture using a Bayesian bivariate GLM fitted to exam and survey outcomes while conditioning on observed covariates (major, gender, race). No derivation, prediction, or uniqueness claim reduces to its own inputs by construction; the central result is a direct model-based comparison of real student data. No self-citations are invoked as load-bearing mathematical facts. The analysis is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review provides no explicit free parameters, axioms, or invented entities; the Bayesian model is mentioned but not detailed.

pith-pipeline@v0.9.1-grok · 5752 in / 917 out tokens · 16888 ms · 2026-06-27T17:42:22.905104+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

77 extracted references · 6 canonical work pages · 1 internal anchor

  1. [1]

    Journal of Statistics Education , volume=

    Conditional Probability and Education Reform: Are They Compatible? , author=. Journal of Statistics Education , volume=. 1995 , doi=

    1995

  2. [2]

    Journal of Statistics Education , volume=

    Student-Centered Instruction in a Theoretical Statistics Course , author=. Journal of Statistics Education , volume=. 2009 , doi=

    2009

  3. [3]

    Journal of Statistics and Data Science Education , volume=

    A Systematic Literature Review of Undergraduate Data Science Education Research , author=. Journal of Statistics and Data Science Education , volume=. 2025 , doi=

    2025

  4. [4]

    , journal=

    Dogucu, Mine and Kazak, Sibel and Rosenberg, Joshua M. , journal=. The Design and Implementation of a. 2025 , doi=

    2025

  5. [5]

    Journal of Statistics Education , volume=

    What Might a Twenty-Year-Old Conference Tell Us About the Future of Our Profession? , author=. Journal of Statistics Education , volume=. 2013 , doi=

    2013

  6. [6]

    Introducing

    Eadie, Gwendolyn and Huppenkothen, Daniela and Springford, Aaron and McCormick, Tyler , journal=. Introducing. 2019 , doi=

    2019

  7. [7]

    Scandinavian Journal of Statistics , volume=

    A Historical Overview of Textbook Presentations of Statistical Science , author=. Scandinavian Journal of Statistics , volume=. 2023 , doi=

    2023

  8. [8]

    Technology Innovations in Statistics Education , volume=

    The Introductory Statistics Course: A Ptolemaic Curriculum , author=. Technology Innovations in Statistics Education , volume=. 2007 , url=

    2007

  9. [9]

    2016 , institution=

    Guidelines for Assessment and Instruction in Statistics Education College Report 2016 , author=. 2016 , institution=

    2016

  10. [10]

    2005 , address =

    Guidelines for Assessment and Instruction in Statistics Education College Report 2005 , institution =. 2005 , address =

    2005

  11. [11]

    Computational Statistics & Data Analysis , volume=

    A general class of zero-or-one inflated beta regression models , author=. Computational Statistics & Data Analysis , volume=. 2012 , doi=

    2012

  12. [12]

    Psychological Methods , volume=

    Beta regression for modeling proportions , author=. Psychological Methods , volume=. 2006 , doi=

    2006

  13. [13]

    Journal of Applied Statistics , year =

    Beta Regression for Modelling Rates and Proportions , author =. Journal of Applied Statistics , year =

  14. [14]

    Journal of the Royal Statistical Society: Series B , volume=

    Regression models for ordinal data , author=. Journal of the Royal Statistical Society: Series B , volume=. 1980 , publisher=

    1980

  15. [15]

    2010 , publisher=

    Analysis of Ordinal Categorical Data , author=. 2010 , publisher=

    2010

  16. [16]

    Practical

    Vehtari, Aki and Gelman, Andrew and Gabry, Jonah , journal=. Practical. 2017 , doi=

    2017

  17. [17]

    loo: Efficient Leave-One-Out Cross-Validation and WAIC for

    Vehtari, Aki and Gelman, Andrew and Simpson, Daniel and Carpenter, Bob and Bürkner, Paul-Christian , year =. loo: Efficient Leave-One-Out Cross-Validation and WAIC for

  18. [18]

    brms: An R Package for

    B. brms: An R Package for. Journal of Statistical Software , volume=. 2017 , doi=

    2017

  19. [19]

    The brms Book: Applied

    B. The brms Book: Applied. 2025 , note=

    2025

  20. [20]

    Understanding predictive information criteria for

    Gelman, Andrew and Hwang, Jessica and Vehtari, Aki , journal=. Understanding predictive information criteria for. 2014 , doi=

    2014

  21. [21]

    and Stern, Hal S

    Gelman, Andrew and Carlin, John B. and Stern, Hal S. and Dunson, David and Vehtari, Aki and Rubin, Donald B. , edition=. 2013 , publisher=

    2013

  22. [22]

    2006 , doi=

    Prior distributions for variance parameters in hierarchical models , author=. 2006 , doi=

    2006

  23. [23]

    Advanced

    B. Advanced. The R Journal , volume=. 2018 , doi=

    2018

  24. [24]

    Journal of Machine Learning Research , volume=

    The No-U-Turn Sampler: Adaptively Setting Path Lengths in Hamiltonian Monte Carlo , author=. Journal of Machine Learning Research , volume=

  25. [25]

    2024 , note=

    Stan Modeling Language Users Guide and Reference Manual , author=. 2024 , note=

    2024

  26. [26]

    Statistical Science , volume=

    Inference from iterative simulation using multiple sequences , author=. Statistical Science , volume=. 1992 , doi=

    1992

  27. [27]

    2021 , doi=

    Rank-normalization, folding, and localization: An improved R-hat for assessing convergence of MCMC , author=. 2021 , doi=

    2021

  28. [28]

    Statistica Sinica , volume=

    Posterior predictive assessment of model fitness via realized discrepancies , author=. Statistica Sinica , volume=

  29. [29]

    Treatment

    Andrew Gelman and Aki Vehtari , title =. Journal of the American Statistical Association , volume =. 2021 , publisher =. doi:10.1080/01621459.2021.1938081 , URL =

    work page doi:10.1080/01621459.2021.1938081 2021

  30. [30]

    Four transformations on the Catalan triangle

    Karl W. Broman and Kara H. Woo , title =. The American Statistician , volume =. 2018 , publisher =. doi:10.1080/00031305.2017.1375989 , URL =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1080/00031305.2017.1375989 2018

  31. [31]

    Bishop, Yvonne M. M. and Fienberg, Stephen E. and Holland, Paul W. , title =. 1975 , pages =

    1975

  32. [32]

    and Ott, Miles Q

    Johnson, Alicia A. and Ott, Miles Q. and Dogucu, Mine , year=

  33. [33]

    Journal for Research in Mathematics Education , year =

    Garfield, Joan and Ahlgren, Andrew , title =. Journal for Research in Mathematics Education , year =

  34. [34]

    Developing a Statistically Literate Society

    Kvatinsky, Tova and Even, Ruhama , title =. Developing a Statistically Literate Society. Proceedings of the Sixth International Conference on Teaching Statistics (ICOTS 6) , editor =. 2002 , pages =

    2002

  35. [35]

    Empirical Research in Understanding Probability , booktitle =

    Borovcnik, Manfred and Bentz, Hans-J. Empirical Research in Understanding Probability , booktitle =. 1991 , pages =

    1991

  36. [36]

    Vision , year =

    Cui, Lucy and Lo, Stephanie and Liu, Zili , title =. Vision , year =

  37. [37]

    Journal of Statistics Education , year =

    Keeler, Carolyn and Steinhorst, Kirk , title =. Journal of Statistics Education , year =

  38. [38]

    and Ben-Zvi, Dani , title =

    Garfield, Joan B. and Ben-Zvi, Dani , title =. 2008 , address =. doi:10.1007/978-1-4020-8383-9 , isbn =

    work page doi:10.1007/978-1-4020-8383-9 2008

  39. [39]

    Nature , year =

    Ghahramani, Zoubin , title =. Nature , year =

  40. [40]

    , title =

    Murphy, Kevin P. , title =. 2022 , address =

    2022

  41. [41]

    and Lock, Patti Frazer and Lock Morgan, Kari and Lock, Eric and Lock, Dennis F

    Lock, Robin H. and Lock, Patti Frazer and Lock Morgan, Kari and Lock, Eric and Lock, Dennis F. , title =. 2021 , address =

    2021

  42. [42]

    2016 , address =

    Tintle, Nathan and Chance, Beth and Cobb, George and Rossman, Allan and Roy, Soma and Swanson, Todd and VanderStoep, Jill , title =. 2016 , address =

    2016

  43. [43]

    Annual Review of Biomedical Data Science , year =

    Probabilistic Machine Learning for Healthcare , author =. Annual Review of Biomedical Data Science , year =

  44. [44]

    2024 , issn =

    International Journal of Forecasting , volume =. 2024 , issn =. doi:https://doi.org/10.1016/j.ijforecast.2023.05.002 , author =

    work page doi:10.1016/j.ijforecast.2023.05.002 2024

  45. [45]

    and Madigan, David and Hoeting, Jennifer A

    Raftery, Adrian E. and Madigan, David and Hoeting, Jennifer A. , title =. Journal of the American Statistical Association , year =

  46. [46]

    Handbook of Economic Forecasting , editor =

    Geweke, John and Whiteman, Charles , title =. Handbook of Economic Forecasting , editor =

  47. [47]

    A Comparative Study of

    Yu, Cheng. A Comparative Study of. Statistics and Data Science in Imaging , year =

  48. [48]

    Proceedings of the 31st International Conference on Neural Information Processing Systems , pages =

    Kendall, Alex and Gal, Yarin , title =. Proceedings of the 31st International Conference on Neural Information Processing Systems , pages =. 2017 , isbn =

    2017

  49. [49]

    British Journal of Mathematical and Statistical Psychology , year =

    Andrews, Mark and Baguley, Thom , title =. British Journal of Mathematical and Statistical Psychology , year =

  50. [50]

    , title =

    Lynch, Scott M. , title =. Annual Review of Sociology , year =

  51. [51]

    Journal of Statistics Education , year =

    Hoegh, Andrew , title =. Journal of Statistics Education , year =

  52. [52]

    Mueller, P. M. and Torres-Espín, Abel and Vonder Haar, Cole , title =. Neurotrauma Report , year =

  53. [53]

    Proceedings of the Third Annual Applied Statistics Education and Research Collaboration (ASEARC) Conference , year =

    Baglin, James and Da Costa, Christopher , title =. Proceedings of the Third Annual Applied Statistics Education and Research Collaboration (ASEARC) Conference , year =

  54. [54]

    The American Statistician , year =

    Dogucu, Mine and Hu, Jingchen , title =. The American Statistician , year =

  55. [55]

    What is POGIL? , year =

  56. [56]

    Biochemistry and Molecular Biology Education , year =

    Minderhout, Vicki and Loertscher, Jennifer , title =. Biochemistry and Molecular Biology Education , year =

  57. [57]

    Proceedings of the 43rd ACM Technical Symposium on Computer Science Education , pages =

    Kussmaul, Clifton , title =. Proceedings of the 43rd ACM Technical Symposium on Computer Science Education , pages =. 2012 , isbn =. doi:10.1145/2157136.2157246 , abstract =

    work page doi:10.1145/2157136.2157246 2012

  58. [58]

    Brown, Patricia J. P. , title =. Advances in Physiology Education , year =

  59. [59]

    , title =

    Rossman, Allan J. , title =

  60. [60]

    Education Finance and Policy , volume =

    Effects of Flipped Classroom Instruction: Evidence from a Randomized Trial , author =. Education Finance and Policy , volume =. 2021 , doi =

    2021

  61. [61]

    International Journal of Educational Technology in Higher Education , volume =

    Investigating Effects of Teachers in Flipped Classroom: A Randomized Controlled Trial Study of Classroom Level Heterogeneity , author =. International Journal of Educational Technology in Higher Education , volume =. 2023 , doi =

    2023

  62. [62]

    Journal of Statistics Education , volume =

    A Meta-Analytic Review of Studies of the Effectiveness of Small-Group Learning Methods on Statistics Achievement , author =. Journal of Statistics Education , volume =. 2014 , doi =

    2014

  63. [63]

    Academic Medicine , volume =

    Problem-based Learning: A Review of Literature on Its Outcomes and Implementation Issues , author =. Academic Medicine , volume =. 1993 , doi =

    1993

  64. [64]

    Proceedings of the National Academy of Sciences , volume =

    Active Learning Increases Student Performance in Science, Engineering, and Mathematics , author =. Proceedings of the National Academy of Sciences , volume =. 2014 , doi =

    2014

  65. [65]

    American Journal of Pharmaceutical Education , volume =

    A Meta-Analysis of Outcomes Comparing Flipped Classroom and Lecture , author =. American Journal of Pharmaceutical Education , volume =. 2018 , doi =

    2018

  66. [66]

    Journal of Statistics and Data Science Education , year =

    Cooperative Learning in Introductory Statistics: Assessing Student Perceptions, Performance, and Learning in Heterogeneous and Homogeneous Groups , author =. Journal of Statistics and Data Science Education , year =

  67. [67]

    Educational Psychologist , volume =

    Why Minimal Guidance During Instruction Does Not Work: An Analysis of the Failure of Constructivist, Discovery, Problem-Based, Experiential, and Inquiry-Based Teaching , author =. Educational Psychologist , volume =. 2006 , doi =

    2006

  68. [68]

    Learning and Instruction , volume =

    Effects of problem-based learning: A meta-analysis , author =. Learning and Instruction , volume =. 2003 , url =

    2003

  69. [69]

    Teaching Statistics and Data Science Collaboration via a Community of Practice,

    The Design and Implementation of a Bayesian Data Analysis Lesson for Pre-Service Mathematics and Science Teachers , author =. Journal of Statistics and Data Science Education , year =. doi:10.1080/26939169.2024.2362148 , note =

    work page doi:10.1080/26939169.2024.2362148 2024

  70. [70]

    Thinking, fast and slow , year =

    Kahneman, Daniel , biburl =. Thinking, fast and slow , year =

  71. [71]

    Journal of Statistics Education , year =

    Albert, Jim , title =. Journal of Statistics Education , year =

  72. [72]

    and Rossman, Allan J

    Albert, Jim H. and Rossman, Allan J. , title =. 2001 , url =

    2001

  73. [73]

    2019 , doi =

    Albert, Jim and Hu, Jingchen , title =. 2019 , doi =

    2019

  74. [74]

    Journal of Statistics Education , year =

    Albert, Jim and Hu, Jingchen , title =. Journal of Statistics Education , year =

  75. [75]

    and Ott, Miles Q

    Johnson, Alicia A. and Ott, Miles Q. and Dogucu, Mine , title =. 2022 , doi =

    2022

  76. [76]

    , title =

    Moore, David S. , title =. The American Statistician , year =

  77. [77]

    , title =

    Rossman, Allan J. , title =. 2019 , month = sep, day =

    2019