pith. sign in

arxiv: 2606.08986 · v1 · pith:3IGHBMQ5new · submitted 2026-06-08 · ⚛️ physics.ed-ph

Discovering Misconceptions and Misunderstandings From Administrations of Research-Designed Multiple Choice Instruments

Martin Segado , Aaron Adair , John Stewart , David Pritchard This is my paper

Pith reviewed 2026-06-27 14:25 UTC · model grok-4.3

classification ⚛️ physics.ed-ph
keywords misconceptionsForce Concept Inventoryitem response theoryNewtonian mechanicsphysics education researchmultidimensional modelingdistractor analysisformative assessment
0
0 comments X

The pith

A multidimensional model applied to 34,000 Force Concept Inventory responses extracts 22 coherent student misconceptions in Newtonian mechanics.

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

The paper uses a flexible multidimensional item-response model on a large set of Force Concept Inventory administrations to identify underlying dimensions in how students choose distractors. These dimensions group answers that share thematic content, allowing each to be labeled as a specific misconception or misunderstanding. The authors classify the 22 dimensions into ancient, medieval, and post-Newtonian categories and introduce misconception scores to measure how instruction affects each one. A sympathetic reader would care because the approach moves beyond total scores to diagnose which alternate ideas persist and for which students.

Core claim

Using a flexible multidimensional item-response model that lets different answer choices within each question point in different directions, the analysis of approximately 34,000 Force Concept Inventory administrations uncovers 22 robust, partly-overlapping dimensions. Each dimension is defined by distractors that share a coherent theme identifiable with a misconception or misunderstanding. These are sorted by historical era into Ancient, Medieval, and Post-Newtonian groups. Simple misconception scores are then computed for students and classes, revealing that some misconceptions remain largely unchanged by instruction while others are better remediated in below- or above-average students, wi

What carries the argument

The flexible multidimensional item-response model for multiple-choice data, which allows answer choices to occupy different directions in the knowledge space so that distinct misconceptions encoded in distractors can be separated.

If this is right

  • Misconception scores can be calculated for individual students or entire classes to track specific errors.
  • Instruction leaves some misconceptions largely unchanged while remediating others more effectively in students of higher or lower ability.
  • Many misconceptions remain poorly addressed for students of average or below-average ability.
  • Instructors gain a tool for class-level formative assessment focused on particular alternate ideas.

Where Pith is reading between the lines

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

  • The same modeling approach could be applied to other multiple-choice concept inventories to surface hidden misconceptions in different domains.
  • The historical classification implies that some persistent errors may benefit from teaching that explicitly contrasts pre-Newtonian ideas with modern ones.
  • If the dimensions prove stable across populations, they could serve as targets for controlled experiments comparing different remediation strategies.

Load-bearing premise

The dimensions extracted by the model correspond to genuine, stable student misconceptions rather than statistical artifacts of the chosen parameterization or the specific distractors in the test items.

What would settle it

Repeating the full analysis with an alternate multidimensional parameterization or on a different concept inventory and finding that the resulting dimensions no longer group into coherent, historically recognizable misconceptions would falsify the central claim.

Figures

Figures reproduced from arXiv: 2606.08986 by Aaron Adair, David Pritchard, John Stewart, Martin Segado.

Figure 1
Figure 1. Figure 1: BigeominQ-rotated distractor vectors from post-instruction High School Modeling matched with those from Large Public 3. The top 9 post-instruction vectors from Large Public 3 are on the left. Dimension 4 is doubled since it “matches” with two High School Modeling vectors—typical of comparisons with results where more dimensions are recovered. Correlation coefficients (uncentered Pearson) are shown in bold … view at source ↗
Figure 2
Figure 2. Figure 2: Uncentered Pearson correlations of post-instruction Large Public 3 vs High School Modeling distractor vectors. The 10 vectors from Large Public 3 (rows) are correlated with the 14 vectors from High School Modeling (columns); both used the BigeominQ rotation method. Dark shading indicates coefficients > 0.75 and highlights that 8 of the 10 Large Public 3 vectors correlate with one and only one “similar” vec… view at source ↗
Figure 3
Figure 3. Figure 3: Twelve of these distractor vectors correlate above 0.72. In contrast, the typical correlation [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 3
Figure 3. Figure 3: Uncentered Pearson correlations of pre- and post-instruction distractor vectors (columns and rows respectively) after BigeominQ rotation. Twelve of these vectors correlate at above 0.72, showing good overlap between the two sets of discovered sparse vectors. These strongly suggest that many of our sparse distractor vectors represent educationally-important clusters of distractors which are valid in multipl… view at source ↗
Figure 4
Figure 4. Figure 4: Eight candidate sparse solutions. These were obtained by applying our method to the combined FCI dataset including both pre- and post-instruction results from multiple schools. The eight solutions are largely similar, with the notable exception being the first dimension of the quartimin-rotated solutions. Further differences are discussed in the text. 13 [PITH_FULL_IMAGE:figures/full_fig_p013_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Hyperplane fractions for eight candidate solutions, here defined as the fraction of distractor slopes with magnitude less than 0.2. Solutions with more nearly-zero distractor slopes (i.e., greater hyperplane count) are typically easier to interpret and perhaps more likely to match real psychological processes. 3.5.1 A simple metric of overall solution quality One of the simplest quality metrics for a facto… view at source ↗
Figure 6
Figure 6. Figure 6: Histograms of extracted number of dimension in bootstrap evaluations. The three panes show the distributions, over 500 bootstrap samples, of the number of dimensions extracted from only pre-instruction data, only post-instruction data, and all data combined. using our MNCM-Bayes method, permitting up to 25 dimensions as when fitting the parent dataset and identifying the results to have orthogonal a vector… view at source ↗
Figure 7
Figure 7. Figure 7: Bootstrapped correlation coefficients for eight candidate solutions, computed per-dimension after applying a deadband of ±0.2 as described in the text. White points indicate median values across bootstrap samples, while thick and thin lines indicate interquartile ranges and central 95% percentile intervals, respectively. 17 [PITH_FULL_IMAGE:figures/full_fig_p017_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Four impetus misconception vectors. The first loads primarily on Q5 and Q18 which involve impetus force in circular motion, the second loads strongly on six distractors all involving an impetus force along the (straight-line) motion, and the third loads strongly on all distractors in the preceding two. The fourth describes a different but related concept: the continuation of a circular trajectory in the ab… view at source ↗
Figure 9
Figure 9. Figure 9: Binned m-scores and pre-post gains for predominantly ancient misconceptions, shown as functions of pre-test raw score. Dot areas are proportional to the number of students in each bin and error bars show linearized standard errors. Dashed lines represent “random guessing” baselines (see Section 5.2). 30 [PITH_FULL_IMAGE:figures/full_fig_p030_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Binned m-scores and pre-post gains for predominantly medieval miscon￾ceptions, shown as functions of pre-test raw score. Dot areas are proportional to the number of students in each bin and error bars show linearized standard errors. Dashed lines represent “random guessing” baselines (see Section 5.2). 31 [PITH_FULL_IMAGE:figures/full_fig_p031_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Binned m-scores and pre-post gains for predominantly post-Newtonian and novel misconceptions, shown as functions of pre-test raw score. Dot areas are proportional to the number of students in each bin and error bars show linearized standard errors. Dashed lines represent “random guessing” baselines (see Section 5.2). 32 [PITH_FULL_IMAGE:figures/full_fig_p032_11.png] view at source ↗
read the original abstract

Misconceptions are "alternate hypotheses" that are incorrect according to established theories of how the world works. Often held with confidence by students, they are relatively context-insensitive, can seem like common-sense views, and are noted for being resistant to remediation using traditional instruction. To find misconceptions in Newtonian mechanics, we analyze ~34,000 administrations of the pioneering Force Concept Inventory using a flexible multidimensional item-response model for multiple-choice data. In contrast to most earlier work, we allow answer choices within each question to have different directions in the multidimensional space of student knowledge, essential for concept inventories in which distractors often codify distinct misconceptions. We uncover 22 robust, partly-overlapping dimensions whose distractors share a coherent theme identifiable with a misconception or misunderstanding. Motivated by the realization that many mirror previously-accepted theories of mechanics, we broadly sort these by historical era: Ancient (learned by infants but codified by Greeks), Medieval (reactions and extensions of Aristotelian ideas), and Post-Newtonian (including known modern misconceptions as well as two which appear novel). We also present a simple approach for computing "misconception scores" for students and classes. Examining these scores before and after instruction reveals surprisingly varied patterns of remediation in our sample: some misconceptions persist largely unchanged by instruction, while others are better remediated in below- or above-average students. In general, we find that many misconceptions are poorly remediated for students of average or lower ability. We hope our work will serve as a guide for developing, evaluating, and improving interventions for these while providing physics instructors with a valuable tool for class-level formative assessment.

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

2 major / 2 minor

Summary. The paper applies a flexible multidimensional item-response model (allowing answer choices to point in different directions) to ~34,000 Force Concept Inventory administrations. It claims to extract 22 robust, partly-overlapping dimensions whose distractors share coherent themes identifiable as misconceptions, sorts them historically into Ancient, Medieval, and Post-Newtonian categories, and introduces misconception scores whose pre/post-instruction changes reveal varied remediation patterns, with many misconceptions poorly remediated for average or lower-ability students.

Significance. If the dimensions prove stable and generalizable beyond the specific model and item set, the work would supply physics education researchers with a data-driven taxonomy of misconceptions and a practical scoring method for formative assessment, potentially guiding more targeted interventions than unidimensional FCI scoring.

major comments (2)
  1. [Methods (IRT model and dimension extraction)] The central claim that the 22 dimensions are 'robust' and reflect genuine misconceptions (rather than artifacts of the chosen multidimensional IRT parameterization or distractor correlations) is load-bearing, yet the abstract supplies no description of the robustness procedure, cross-validation, hold-out testing, sensitivity to dimension count or link function, or comparison against unidimensional baselines. This information is required to evaluate the claim.
  2. [Results (dimension extraction and robustness)] No model-fit statistics, information criteria, or details on how the dimensionality was selected or validated are reported. Without these, it is impossible to determine whether the 22 dimensions are overparameterized or whether the observed thematic coherence arises from the model structure itself.
minor comments (2)
  1. [Discussion (historical classification)] The historical-era sorting of dimensions is presented as motivated by prior theories but would benefit from an explicit decision rule or inter-rater procedure to avoid appearing post-hoc.
  2. [Methods (misconception scores)] The misconception-score formula is described as 'simple' but its exact definition, normalization, and handling of overlapping dimensions should be stated explicitly with an equation.

Simulated Author's Rebuttal

2 responses · 0 unresolved

Thank you for the opportunity to respond to the referee's report. We address each major comment below and indicate the revisions we will make to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Methods (IRT model and dimension extraction)] The central claim that the 22 dimensions are 'robust' and reflect genuine misconceptions (rather than artifacts of the chosen multidimensional IRT parameterization or distractor correlations) is load-bearing, yet the abstract supplies no description of the robustness procedure, cross-validation, hold-out testing, sensitivity to dimension count or link function, or comparison against unidimensional baselines. This information is required to evaluate the claim.

    Authors: We agree that the abstract lacks a summary of the robustness checks. The full manuscript details the cross-validation and hold-out procedures used to establish the stability of the 22 dimensions, along with comparisons showing improved fit relative to unidimensional models. We will revise the abstract to include a concise description of these procedures and add explicit sensitivity analyses for dimension count in the methods section. We did not perform a full sensitivity analysis on the link function in the original work; this can be added as a supplementary check if requested. revision: yes

  2. Referee: [Results (dimension extraction and robustness)] No model-fit statistics, information criteria, or details on how the dimensionality was selected or validated are reported. Without these, it is impossible to determine whether the 22 dimensions are overparameterized or whether the observed thematic coherence arises from the model structure itself.

    Authors: We acknowledge that the manuscript would benefit from explicit reporting of model-fit statistics. In the revised version we will add AIC, BIC, and other information criteria, together with a step-by-step account of how dimensionality was selected through successive model comparisons and validation. These additions will allow readers to assess whether the 22 dimensions are overparameterized. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation is data-driven extraction from external administrations.

full rationale

The paper applies a multidimensional IRT model to an external dataset of ~34,000 FCI administrations and extracts dimensions whose themes are identified post-hoc from the data. No step defines the target dimensions in terms of the fitted parameters themselves, renames a fitted quantity as a prediction, or relies on a self-citation chain whose content is unverified outside the present work. The central result is therefore an empirical finding rather than a tautological re-expression of inputs or prior author claims.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

Only the abstract is available, so the ledger is necessarily incomplete and based on stated methodological choices rather than explicit equations.

free parameters (1)
  • dimensionality of the IRT model
    The number of dimensions (22) is extracted from data; the precise parameterization of the flexible model is not specified.
axioms (1)
  • domain assumption Distractors that load on the same dimension share a coherent misconception theme
    This interpretive step is required to label the statistical dimensions as misconceptions.

pith-pipeline@v0.9.1-grok · 5833 in / 1148 out tokens · 17221 ms · 2026-06-27T14:25:06.077847+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

61 extracted references · 22 canonical work pages · 2 internal anchors

  1. [1]

    Segado, Martin and Adair, Aaron and Stewart, John and Ma, Yunfei and Drury, Byron and Pritchard, David , year =. A. Frontiers in Psychology , volume =. doi:10.3389/fpsyg.2025.1506320 , langid =

    work page doi:10.3389/fpsyg.2025.1506320 2025

  2. [2]

    Student Misconceptions about Newtonian Mechanics: Origins and Solutions through Changes to Instruction

  3. [3]

    1957 , Address =

    Concepts of Force: A Study in the Foundations of Dynamics , Author =. 1957 , Address =

    1957

  4. [4]

    Chi, Michelene T. H. and Feltovich, Paul J. and Glaser, Robert , year =. Categorization and. Cognitive Science , volume =. doi:10.1207/s15516709cog0502_2 , langid =

    work page doi:10.1207/s15516709cog0502_2

  5. [5]

    Educational and Psychological Measurement , volume = 65, number = 5, pages =

    Gradient Projection Algorithms and Software for Arbitrary Rotation Criteria in Factor Analysis , author =. Educational and Psychological Measurement , volume = 65, number = 5, pages =

  6. [6]

    Journal of the American Statistical Association , publisher =

    Variational Inference: A Review for Statisticians , author =. Journal of the American Statistical Association , publisher =

  7. [7]

    Psychometrika , volume = 37, number = 1, pages =

    Estimating Item parameters and Latent Ability when Responses Are Scored in Two or More Nominal Categories , author =. Psychometrika , volume = 37, number = 1, pages =

  8. [8]

    Applied Psychological Measurement , volume = 12, number = 3, pages =

    Full-Information Item Factor Analysis , author =. Applied Psychological Measurement , volume = 12, number = 3, pages =

  9. [9]

    Characterizing the mathematical problem-solving strategies of transitioning novice physics students , author =. Phys. Rev. Phys. Educ. Res. , volume =. 2020 , publisher =

    2020

  10. [10]

    Philip , year = 2012, journal =

    Chalmers, R. Philip , year = 2012, journal =. Mirt: A Multidimensional Item Response Theory Package for the

    2012

  11. [11]

    The Physics Teacher , volume = 30, number = 3, pages =

    Force concept inventory , author =. The Physics Teacher , volume = 30, number = 3, pages =

  12. [12]

    Force Concept Inventory, revised version (v95) , author =

  13. [13]

    Hestenes, David and Jackson, Jane , year = 2010, howpublished =. Table

    2010

  14. [14]

    Psychometrika , volume = 76, number = 4, pages =

    Exploratory Bi-Factor Analysis , author =. Psychometrika , volume = 76, number = 4, pages =

  15. [15]

    Adam: A Method for Stochastic Optimization

    Adam: A Method for Stochastic Optimization , author =. doi:10.48550/arXiv.1412.6980 , howpublished =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1412.6980

  16. [16]

    Handbook of Item Response Theory, Volume 2: Statistical Tools , year = 2016, publisher =

    2016

  17. [17]

    , year = 2016, journal =

    Natesan, Prathiba and Nandakumar, Ratna and Minka, Tom and Rubright, Jonathan D. , year = 2016, journal =. Bayesian Prior Choice in. doi:10.3389/fpsyg.2016.01422 , pmcid =

    work page doi:10.3389/fpsyg.2016.01422 2016

  18. [18]

    Proceedings of the Sixth ACM Conference on Learning @ Scale , location =

    Mining Students Pre-instruction Beliefs for Improved Learning , author =. Proceedings of the Sixth ACM Conference on Learning @ Scale , location =

  19. [19]

    , year =

    Brown, David E. , year =. Students'. Science & Education , volume =. doi:10.1007/s11191-013-9655-9 , langid =

    work page doi:10.1007/s11191-013-9655-9

  20. [20]

    and Kryjevskaia, Mila and Stetzer, MacKenzie R

    Gette, Cody R. and Kryjevskaia, Mila and Stetzer, MacKenzie R. and Heron, Paula R. L. , year =. Probing Student Reasoning Approaches through the Lens of Dual-Process Theories:. Physical Review Physics Education Research , volume =

  21. [21]

    and Slotta, James D

    Chi, Michelene T.H. and Slotta, James D. , year =. The. Cognition and Instruction , volume =

  22. [22]

    1993 , journal =

    Toward an. 1993 , journal =

    1993

  23. [23]

    Composable Effects for Flexible and Accelerated Probabilistic Programming in NumPyro

    Phan, Du and Pradhan, Neeraj and Jankowiak, Martin , year = 2019, publisher =. Composable Effects for Flexible and Accelerated Probabilistic Programming in. doi:10.48550/arXiv.1912.11554 , howpublished =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1912.11554 2019

  24. [24]

    Frontiers in Psychology , volume = 8, doi =

    Bayesian Dimensionality Assessment for the Multidimensional Nominal Response Model , author =. Frontiers in Psychology , volume = 8, doi =

  25. [25]

    Psychometrika , volume = 31, number = 1, pages =

    A Generalized Solution of the Orthogonal Procrustes Problem , author =. Psychometrika , volume = 31, number = 1, pages =

  26. [26]

    Physical Review Physics Education Research , volume = 17, number = 1, pages =

    Examining the Relation of Correct Knowledge and Misconceptions Using the Nominal Response Model , author =. Physical Review Physics Education Research , volume = 17, number = 1, pages =

  27. [27]

    Psychometrika , volume = 52, number = 3, pages =

    On the Relationship between Item Response Theory and Factor Analysis of Discretized Variables , author =. Psychometrika , volume = 52, number = 3, pages =

  28. [28]

    Handbook of Polytomous Item Response Theory Models , publisher =

    The Nominal Categories Item Response Model , author =. Handbook of Polytomous Item Response Theory Models , publisher =. doi:10.4324/9780203861264.ch3 , isbn =

    work page doi:10.4324/9780203861264.ch3

  29. [29]

    Handbook of Item Response Theory , publisher =

    Nominal Categories Models , author =. Handbook of Item Response Theory , publisher =

  30. [30]

    Exploring the structure of misconceptions in the Force Concept Inventory with modified module analysis , author =. Phys. Rev. Phys. Educ. Res. , volume =. 2019 , publisher =

    2019

  31. [31]

    Comparing conceptual understanding across institutions with module analysis , author =. Phys. Rev. Phys. Educ. Res. , volume =. 2022 , publisher =

    2022

  32. [32]

    Psychometrika , volume = 85, number = 2, pages =

    A Note on Exploratory Item Factor Analysis by Singular Value Decomposition , author =. Psychometrika , volume = 85, number = 2, pages =

  33. [33]

    2017 , journal =

    Using the Method of Dominant Incorrect Answers with the. 2017 , journal =. doi:10.1088/1361-6552/52/1/015006 , copyright =

    work page doi:10.1088/1361-6552/52/1/015006 2017

  34. [34]

    , author =

    Local Minima and Factor Rotations in Exploratory Factor Analysis. , author =. 2023 , journal =. doi:10.1037/met0000467 , langid =

    work page doi:10.1037/met0000467 2023

  35. [35]

    , year =

    Scherr, Rachel E. , year =. Modeling Student Thinking:. American Journal of Physics , volume =. doi:10.1119/1.2410013 , abstract =

    work page doi:10.1119/1.2410013

  36. [36]

    Using module analysis for multiple choice responses: A new method applied to Force Concept Inventory data , author =. Phys. Rev. Phys. Educ. Res. , volume =. 2016 , publisher =

    2016

  37. [37]

    Exploring the structure of misconceptions in the. Phys. Rev. Phys. Educ. Res. , author =. 2020 , doi =

    2020

  38. [38]

    Quantitatively ranking incorrect responses to multiple-choice questions using item response theory , author=. Phys. Rev. Phys. Educ. Res. , volume=. 2020 , publisher=

    2020

  39. [39]

    1963 , journal =

    The Uniqueness and Significance of Simple Structure Demonstrated by Contrasting Organic ``Natural Structure'' and ``Random Structure'' Data , author =. 1963 , journal =. doi:10.1007/BF02289548 , abstract =

    work page doi:10.1007/bf02289548 1963

  40. [40]

    Chi, Michelene T. H. , editor =. Three. Handbook of. 2008 , pages =

    2008

  41. [41]

    Dijksterhuis, E. J. , editor =. The. Critical. 1969 , pages =

    1969

  42. [42]

    2010 , booktitle =

    Docktor, Jennifer and Mestre, Jose , pages =. 2010 , booktitle =

    2010

  43. [43]

    Journal of the History of Ideas , author =

    Impetus. Journal of the History of Ideas , author =. 1975 , pages =. doi:10.2307/2709009 , number =

    work page doi:10.2307/2709009 1975

  44. [44]

    Generating a growth-oriented partial credit grading model for the. Phys. Rev. Phys. Educ. Res. , author =. 2019 , pages =. doi:10.1103/PhysRevPhysEducRes.15.020151 , number =

    work page doi:10.1103/physrevphyseducres.15.020151 2019

  45. [45]

    The cognitive revolution in educational psychology , pages=

    The impact of the cognitive revolution on science learning and teaching , author=. The cognitive revolution in educational psychology , pages=. 2005 , publisher=

    2005

  46. [46]

    Lee, Sunbok and Chen, Zhongzhou and Pritchard, David and Kimn, Alex and Paul, Andrew , year =. Factor. Proceedings of the. doi:10.1145/3051457.3053984 , copyright =

    work page doi:10.1145/3051457.3053984

  47. [47]

    2006 , journal =

    Tucker's Congruence Coefficient as a Meaningful Index of Factor Similarity , author =. 2006 , journal =. doi:10.1027/1614-2241.2.2.57 , abstract =

    work page doi:10.1027/1614-2241.2.2.57 2006

  48. [48]

    1976 , journal =

    Procrustes Matching by Congruence Coefficients , author =. 1976 , journal =. doi:10.1007/BF02296973 , abstract =

    work page doi:10.1007/bf02296973 1976

  49. [49]

    1975 , journal =

    Generalized Procrustes Analysis , author =. 1975 , journal =. doi:10.1007/BF02291478 , abstract =

    work page doi:10.1007/bf02291478 1975

  50. [50]

    Multivariate Behavioral Research , author =

    The. Multivariate Behavioral Research , author =. 1992 , note =. doi:10.1207/s15327906mbr2704_5 , abstract =

    work page doi:10.1207/s15327906mbr2704_5 1992

  51. [51]

    , year =

    Hattori, Minami and Zhang, Guangjian and Preacher, Kristopher J. , year =. Multiple. Multivariate Behavioral Research , volume =. doi:10.1080/00273171.2017.1361312 , abstract =

    work page doi:10.1080/00273171.2017.1361312 2017

  52. [52]

    , year =

    Hake, Richard R. , year =. Interactive-Engagement versus Traditional Methods:. American Journal of Physics , volume =

  53. [53]

    Revuelta, Javier and. Factor. 2020 , journal =. doi:10.1080/10705511.2019.1668276 , abstract =

    work page doi:10.1080/10705511.2019.1668276 2020

  54. [54]

    Strategies for

    Lee, Eun and Forthofer, Ronald , year =. Strategies for. Analyzing. doi:10.4135/9781412983341.n4 , isbn =

    work page doi:10.4135/9781412983341.n4

  55. [55]

    and Corbett, Albert T

    Koedinger, Kenneth R. and Corbett, Albert T. and Perfetti, Charles , year =. The. Cognitive Science , volume =. doi:10.1111/j.1551-6709.2012.01245.x , langid =

    work page doi:10.1111/j.1551-6709.2012.01245.x 2012

  56. [56]

    Students' proficiency scores within multitrait item response theory , author=. Phys. Rev. Phys. Educ. Res. , volume=. 2015 , publisher=

    2015

  57. [57]

    and Schumayer, D

    Scott, T.F. and Schumayer, D. and Gray, A.R. , Journal =. Exploratory factor analysis of a. 2012 , Number =

    2012

  58. [58]

    and Dietz, R.D

    Semak, M.R. and Dietz, R.D. and Pearson, R.H. and Willis, C.W. , journal =. Examining evolving performance on the. 2017 , publisher =

    2017

  59. [59]

    and Zabriskie, C

    Stewart, J. and Zabriskie, C. and DeVore, S. and Stewart, G. , journal =. Multidimensional item response theory and the. 2018 , publisher =

    2018

  60. [60]

    and Wells, J

    Yang, J. and Wells, J. and Henderson, R. and Christman, E. and Stewart, G. and Stewart, J. , journal=. Extending modified module analysis to include correct responses:. 2020 , publisher=

    2020

  61. [61]

    What Babies Know: Core Knowledge and Composition:

    Spelke, Elizabeth , year =. What Babies Know: Core Knowledge and Composition: