# Stat Modeling — complete archive in chronological order

Source: https://statmodeling.stat.columbia.edu/
Essays: 100 · Span: 2004-10-12 to 2005-03-10 · Total reading time: ~3 hours

Track your progress through this archive with a free account: https://www.evergreenessays.com/start/stat-modeling

## Essays

1. [A weblog for research in statistical modeling and applications, especially in social sciences](https://statmodeling.stat.columbia.edu/2004/10/12/a_weblog_for_re/) (2004-10-12) — 1 min
  Join Andrew Gelman and his team of experts as they delve into the fascinating world of statistical modeling and its applications, with a particular focus on the social sciences. From Bayesian statistics to multilevel modeling, causal inference, and political science, this weblog covers it all. With contributions from Gelman himself, his students, and esteemed colleagues, get ready for insightful and thought-provoking discussions. Engage in the conversation by sharing your comments, suggestions, and solutions to the open problems they present. Don't miss out on this valuable resource for researchers and enthusiasts alike.
2. [The Electoral College favors voters in small states](https://statmodeling.stat.columbia.edu/2004/10/13/the_electoral_c/) (2004-10-13) — 3 min
  Discover the surprising truth about the Electoral College and its impact on voters in small states in this eye-opening article by Andrew. Unveiling the flawed notion that larger states have more voting power, Andrew dives into the probabilities and statistics behind decisive voting and electoral influence. Uncover why swing states and small swing states play a crucial role in shaping election outcomes, and how your own voting power may be more substantial than you think. Prepare to question preconceived notions as this article challenges conventional wisdom with insightful analysis and thought-provoking findings.
3. [Why it’s rational to vote](https://statmodeling.stat.columbia.edu/2004/10/13/why_its_rationa/) (2004-10-13) — 4 min
  Exploring the rationality behind voting, Andrew presents an intriguing cost-benefit analysis that challenges conventional explanations. While the probability of one vote changing the election outcome is minuscule, Andrew argues that the impact of a preferred candidate winning can have far-reaching benefits for the country as a whole. Drawing on empirical evidence and social motivations, the article invites readers to reevaluate their perception of voting and consider the implications for collective well-being. Discover why voting may be more rational than previously believed in this thought-provoking essay.
4. [Bayes and Popper](https://statmodeling.stat.columbia.edu/2004/10/14/bayes_and_poppe/) (2004-10-14) — 4 min
  Andrew delves into the philosophical underpinnings of statistical inference in this thought-provoking essay. He explores the connection between the deductive reasoning of classical statistics and the inductive reasoning of Bayesian statistics. By challenging the assumptions and principles of both approaches, Andrew encourages us to critically evaluate our models and strive for deeper understanding. Whether you're a statistician or simply curious about the philosophy of science, this article will challenge your preconceptions and broaden your perspective on the art of inference.
5. [Overrepresentation of small states/provinces, and the USA Today effect](https://statmodeling.stat.columbia.edu/2004/10/15/overrepresentat/) (2004-10-15) — 4 min
  Andrew delves into the overrepresentation of small states/provinces in the political systems of various countries, with a particular focus on the United States. He highlights the disparities in electoral votes, Senate seats, and federal spending per capita between large and small states. Andrew investigates the reasons behind this imbalance and explores possible explanations such as the bargaining power of small states during constitution-writing processes and the influence of rural demographics. He also sheds light on the "USA Today effect" and how states are often presented as equals, despite significant population differences.
6. [Sensitivity Analysis of Joanna Shepherd’s DP paper](https://statmodeling.stat.columbia.edu/2004/10/18/sensitivity_ana/) (2004-10-18) — 2 min
  Delve into the intricate world of criminal justice economics with Andrew as he examines Joanna Shepherd's latest research on the deterrence effect of the death penalty. Discover how Andrew plans to test the sensitivity of Shepherd's findings by exploring changes in model specifications and analyzing county-by-year data. Explore the nuances and complexities of this thought-provoking topic as Andrew strategizes and collaborates with expert colleagues. Get ready for an eye-opening journey into the realm of academic research and data analysis in the realm of criminal justice economics.
7. [Unequal representation: comments from David Samuels](https://statmodeling.stat.columbia.edu/2004/10/19/unequal_represe/) (2004-10-19) — 2 min
  David Samuels, coauthor of a groundbreaking paper on unequal representation, shares insightful comments on the overrepresentation of small states/provinces. He recommends exploring Thies' research on over-representation in Japan and highlights Edward Gibson's work on the consequences of apportionment distortions in Latin America. Samuels also sheds light on the unique approach taken by the United States in implementing a 1-person 1-vote solution. His thought-provoking insights challenge democratic ideals and shed light on the political strategies that shape representation in different countries.
8. [Problems with Heterogeneous Choice Models](https://statmodeling.stat.columbia.edu/2004/10/20/problems_with_h/) (2004-10-20) — 2 min
  Andrew delves into the challenges of using heterogeneous choice models in statistical inference and their impact on estimating individual choices. Drawing on examples from political science, he highlights how variation in choices can be more intriguing than the determinants themselves. However, he discovers that current estimation techniques are flawed, leading to biased estimates and incorrect standard errors. Even under perfect specifications, bias and measurement errors persist. To overcome these limitations, Andrew suggests exploring alternative methods, specifically Bayesian estimation techniques, to offer a more effective approach to modeling heterogeneous choices. Discover the complexity and potential solutions in this eye-opening essay.
9. [Morris Fiorina on C-SPAN](https://statmodeling.stat.columbia.edu/2004/10/21/morris_fiorina/) (2004-10-21) — 1 min
  Join Morris Fiorina as he debunks the myth of a polarized America in his upcoming discussion on C-SPAN. In an era plagued by political division, Fiorina challenges popular belief and reveals that most Americans actually hold moderate views on major issues. As a senior fellow at the Hoover Institution and professor of political science at Stanford University, Fiorina brings insightful analysis to the table. Don't miss out on this eye-opening talk that challenges our assumptions about the culture war. Tune in this Sunday at 5:15pm and get ready to question everything you thought you knew.
10. [A fun demo for statistics class](https://statmodeling.stat.columbia.edu/2004/10/21/a_fun_demo_for/) (2004-10-21) — 4 min
  In this intriguing article, Andrew presents a unique and engaging demonstration for statistics classes that explores the challenges of random sampling. By using a bag of candies and a digital scale, students are tasked with estimating the weight of 100 candies. The results, which are surprisingly lower than expected, lead to a thought-provoking discussion on sampling methods and their implications. Andrew's demonstration serves as a fun and interactive way to introduce important statistical concepts, leaving students with a deeper understanding of the complexities involved in obtaining accurate data.
11. [Red State/Blue State Paradox](https://statmodeling.stat.columbia.edu/2004/10/22/red_stateblue_s/) (2004-10-22) — 3 min
12. [Statistical issues in modeling social space](https://statmodeling.stat.columbia.edu/2004/10/23/statistical_iss/) (2004-10-23) — 2 min
  In this intriguing article, Andrew delves into the statistical conundrums surrounding the modeling of social space, particularly within the venture capital industry. By mapping the geographic and industry dimensions of venture capital firms, Andrew aims to uncover the factors that influence their movements in this "social space." From understanding the dynamics of competition and social comparison to examining clustering patterns, this research offers valuable insights into the world of venture capitalism. Join Andrew as he explores the complexities of this statistical puzzle and sheds light on the sociological implications of these findings.
13. [2 Stage Least Squares Regression for Death Penalty Analysis](https://statmodeling.stat.columbia.edu/2004/10/26/2_stage_least_s/) (2004-10-26) — 1 min
  Author Andrew seeks advice on conducting a 2-stage least squares regression in Stata, specifically for replicating Shepherd's analysis on a dataset related to the death penalty. Facing challenges with system identification and excessive endogenous predictors, Andrew is seeking guidance on running the regression with a system of 4 equations, not in reduced form. If you're knowledgeable about this topic, your insights could greatly contribute to Andrew's research. Help Andrew tackle this complex analysis and be a part of advancing the field of death penalty analysis.
14. [Partial pooling of interactions](https://statmodeling.stat.columbia.edu/2004/10/26/partial_pooling/) (2004-10-26) — 3 min
  Andrew delves into the complex world of multi-way analysis of variance and the challenge of dealing with a vast number of predictors. By introducing the concept of partial pooling of interactions, he explores how factors with large main effects often have larger interactions. Using a Bayesian approach, Andrew proposes a model that allows for the variance of each interaction to depend on the coefficients of its component parts. He applies this model to real-world data on public spending and presents promising results. This thought-provoking article aims to pave the way for better models that can capture deep interactions.
15. [Bayesian Methods for Variable Selection](https://statmodeling.stat.columbia.edu/2004/10/26/bayesian_method/) (2004-10-26) — 2 min
  Join Andrew as he delves into the fascinating world of Bayesian methods for variable selection in regression settings. Discover how these methods can help us determine the most influential predictors, improve prediction accuracy, and reduce future data collection costs. Delving into the works of George, McCulloch, Brown, Vannucci, and Fearn, Andrew explores the power of using latent binary vectors and imposing priors on regression parameters and possible models. Learn how Markov chain Monte Carlo (MCMC) techniques play a pivotal role in finding models with high posterior probability, and explore extensions to multinomial probit models for simultaneous classification and variable selection. Don't miss out on this enlightening journey into Bayesian variable selection.
16. [Reference for variable selection](https://statmodeling.stat.columbia.edu/2004/10/26/reference_for_v/) (2004-10-26) — 1 min
  Andrew provides a valuable reference on variable selection by Chipman, a renowned statistician. In this article, Chipman's work on Bayesian variable selection with related predictors is explored. This reference is a must-read for anyone looking to dive into the intricacies of variable selection and gain a deeper understanding of its applications in statistical modeling. Don't miss out on this insightful resource!
17. [The blessing of dimensionality](https://statmodeling.stat.columbia.edu/2004/10/27/the_blessing_of/) (2004-10-27) — 2 min
  Andrew challenges the notion of the "curse of dimensionality" in statistical inference, arguing that more predictors should be seen as a blessing rather than a curse. He explores how multilevel modeling can effectively handle high-dimensional data by grouping measurements and utilizing the structure within the predictors. Andrew provides real-world examples to support his perspective and suggests that, with careful consideration and proper methods, the abundance of data can lead to more accurate parameter estimation. Discover how this fresh perspective challenges traditional beliefs and offers new insights into the world of statistics.
18. [Why poll numbers keep hopping around by Philip Meyer](https://statmodeling.stat.columbia.edu/2004/10/28/why_poll_number/) (2004-10-28) — 5 min
  Discover why pre-election polls are so unpredictable and often misleading in this thought-provoking op-ed by Philip Meyer. Exploring the work of a team led by a Columbia University professor, Meyer uncovers the challenges faced by pollsters in capturing both voter choice and the composition of the actual electorate accurately. With the likelihood of voters shifting in and out of the "likely-voter" group depending on external factors, Meyer argues that instead of focusing solely on election predictions, polls should be used to understand coalition formations and the issues resonating with different voter groups. Dive into this insightful analysis of the polls to gain a fresh perspective on election dynamics.
19. [Matching, regression, interactions, and robustness](https://statmodeling.stat.columbia.edu/2004/10/29/matching_regres/) (2004-10-29) — 4 min
  Get ready for a deep dive into the world of causal inference! In this article by Daniel Ho, he engages in a thought-provoking conversation with Kosuke Imai and Don Rubin about the advantages of matching methods over regression in social science research. They explore the robustness and flexibility of matching, discussing how it offers more reliable inferences and avoids assumptions tied to regression models. If you're a social scientist curious about the strengths of matching and eager to explore alternative approaches to statistical analysis, this article is a must-read. Don't miss out on this stimulating discussion in the pursuit of better causal inference methods.
20. [Homer Simpson and mixture models](https://statmodeling.stat.columbia.edu/2004/11/01/homer_simpson_a/) (2004-11-01) — 2 min
  In this intriguing essay, Andrew delves into the perplexing phenomenon of why many Americans support the repeal of the estate tax, despite simultaneously believing that the wealthy should pay more in taxes. Drawing from Larry Bartels' study, Andrew explores the possible explanations for this seemingly contradictory stance, including confusion and ideological beliefs. Additionally, he proposes the use of a mixture model to better understand the diverse motivations behind individuals' preferences on tax cuts. This thought-provoking analysis sheds light on the complex interplay of economics, politics, and personal biases in shaping public opinion. Don't miss out on this fascinating exploration.
21. [Ideal point models](https://statmodeling.stat.columbia.edu/2004/11/02/ideal_point_mod/) (2004-11-02) — 2 min
22. [Analyzing Cross-Country Survey Data](https://statmodeling.stat.columbia.edu/2004/11/03/analyzing_cross/) (2004-11-03) — 3 min
23. [What does the 2004 Election Say About the Red/Blue Paradox?](https://statmodeling.stat.columbia.edu/2004/11/03/what_does_the_2/) (2004-11-03) — 1 min
24. [Living Poor, Voting Rich](https://statmodeling.stat.columbia.edu/2004/11/03/living_poor_vot/) (2004-11-03) — 4 min
25. [Bayesian Software Validation](https://statmodeling.stat.columbia.edu/2004/11/04/bayesian_softwa/) (2004-11-04) — 4 min
26. [Cross-validation for Bayesian multilevel modeling](https://statmodeling.stat.columbia.edu/2004/11/08/crossvalidation/) (2004-11-08) — 4 min
27. [Cost-benefit analysis, goal-based decision analysis, and Dave Krantz’s talk](https://statmodeling.stat.columbia.edu/2004/11/09/costbenefit_ana/) (2004-11-09) — 2 min
28. [How to save $10 billion](https://statmodeling.stat.columbia.edu/2004/11/10/how_to_save_10/) (2004-11-10) — 1 min
29. [Poetry and combinatorics](https://statmodeling.stat.columbia.edu/2004/11/11/poetry_and_comb/) (2004-11-11) — 4 min
30. [Expedient Methods in Environmental Indexing](https://statmodeling.stat.columbia.edu/2004/11/12/expedient_metho/) (2004-11-12) — 1 min
31. [The Stubborn American Voter](https://statmodeling.stat.columbia.edu/2004/11/15/the_stubborn_am/) (2004-11-15) — 1 min
32. [Institutional decision analysis](https://statmodeling.stat.columbia.edu/2004/11/16/institutional_d/) (2004-11-16) — 3 min
33. [The usual model for before-after data is wrong](https://statmodeling.stat.columbia.edu/2004/11/17/the_usual_model/) (2004-11-17) — 2 min
34. [Estimating spatial interactions in forest clearing](https://statmodeling.stat.columbia.edu/2004/11/18/estimating_spat/) (2004-11-18) — 2 min
35. [Vote swings in Florida in counties with and without e-voting](https://statmodeling.stat.columbia.edu/2004/11/19/vote_swings_in/) (2004-11-19) — 5 min
36. [A linear regression example, and a question](https://statmodeling.stat.columbia.edu/2004/11/22/a_linear_regres/) (2004-11-22) — 2 min
37. [Arsenic in Bangladesh; sharing wells](https://statmodeling.stat.columbia.edu/2004/11/23/arsenic_in_bang/) (2004-11-23) — 3 min
38. [More on voting patterns in Florida (from Jasjeet Sekhon)](https://statmodeling.stat.columbia.edu/2004/11/24/more_on_voting/) (2004-11-24) — 2 min
39. [Knowing what you don’t know](https://statmodeling.stat.columbia.edu/2004/11/29/knowing_what_yo/) (2004-11-29) — 1 min
40. [Ranking colleges](https://statmodeling.stat.columbia.edu/2004/11/30/ranking_college/) (2004-11-30) — 4 min
41. [Ideal-point models with absentions](https://statmodeling.stat.columbia.edu/2004/12/09/idealpoint_mode/) (2004-12-09) — 2 min
42. [Against parsimony](https://statmodeling.stat.columbia.edu/2004/12/10/against_parsimo/) (2004-12-10) — 1 min
43. [Too many polls](https://statmodeling.stat.columbia.edu/2004/12/13/too_many_polls/) (2004-12-13) — 2 min
44. [Multiple imputation for model checking:  completed-data plots with missing and latent data](https://statmodeling.stat.columbia.edu/2004/12/14/multiple_imputa/) (2004-12-14) — 4 min
45. [Statistical teaching, application, and research (STAR) conference at Columbia](https://statmodeling.stat.columbia.edu/2004/12/15/statistical_tea/) (2004-12-15) — 3 min
46. [Wacky computer scientists](https://statmodeling.stat.columbia.edu/2004/12/16/wacky_computer_1/) (2004-12-16) — 1 min
47. [Radon webage is back up and running](https://statmodeling.stat.columbia.edu/2004/12/17/radon_webage_is/) (2004-12-17) — 1 min
48. [Fully Bayesian Computing](https://statmodeling.stat.columbia.edu/2004/12/22/fully_bayesian/) (2004-12-22) — 2 min
49. [Family demography and public policy seminar](https://statmodeling.stat.columbia.edu/2004/12/23/family_demograp/) (2004-12-23) — 3 min
50. [Anything worth doing is worth doing repeatedly, or The fundamental connection between frequentist statistics and hierarchical models](https://statmodeling.stat.columbia.edu/2004/12/24/anything_worth/) (2004-12-24) — 2 min
51. [The “law of parsimony”?](https://statmodeling.stat.columbia.edu/2004/12/27/the_law_of_pars/) (2004-12-27) — 2 min
52. [Adjusting polls for party identification](https://statmodeling.stat.columbia.edu/2004/12/28/adjusting_polls/) (2004-12-28) — 2 min
53. [Type 1, type 2, type S, and type M errors](https://statmodeling.stat.columbia.edu/2004/12/29/type_1_type_2_t/) (2004-12-29) — 3 min
54. [What is the value of a life?](https://statmodeling.stat.columbia.edu/2004/12/30/what_is_the_val/) (2004-12-30) — 2 min
55. [A day late and a dollar short:  summarizing prediction errors of time series and spatial models](https://statmodeling.stat.columbia.edu/2004/12/31/a_day_late_and/) (2004-12-31) — 2 min
56. [What is the probability that Elvis Presley was an identical twin?](https://statmodeling.stat.columbia.edu/2005/01/03/what_is_the_pro/) (2005-01-03) — 1 min
57. [Cool research abstracts from the Monthly Labor Review](https://statmodeling.stat.columbia.edu/2005/01/04/cool_research_a/) (2005-01-04) — 1 min
58. [Twins](https://statmodeling.stat.columbia.edu/2005/01/05/twins/) (2005-01-05) — 3 min
59. [CrashStat](https://statmodeling.stat.columbia.edu/2005/01/06/crashstat/) (2005-01-06) — 1 min
60. [Could propensity score analysis fix the Harvard Nurses study?](https://statmodeling.stat.columbia.edu/2005/01/07/could_propensit/) (2005-01-07) — 3 min
61. [Data-driven Vague Prior Distributions](https://statmodeling.stat.columbia.edu/2005/01/12/datadriven_vagu/) (2005-01-12) — 4 min
62. [Neurobiology and decision making](https://statmodeling.stat.columbia.edu/2005/01/18/neurobiology_an/) (2005-01-18) — 1 min
63. [Blog about statistics teaching](https://statmodeling.stat.columbia.edu/2005/01/19/blog_about_stat/) (2005-01-19) — 1 min
64. [Brouhaha about multilevel models](https://statmodeling.stat.columbia.edu/2005/01/20/brouhaha_about/) (2005-01-20) — 1 min
65. [The (John) Smiths](https://statmodeling.stat.columbia.edu/2005/01/21/the_john_smiths/) (2005-01-21) — 1 min
66. [A question about measurement and educational policy](https://statmodeling.stat.columbia.edu/2005/01/21/a_question_abou/) (2005-01-21) — 2 min
67. [Estimating the probability of events that have never occurred](https://statmodeling.stat.columbia.edu/2005/01/24/estimating_the/) (2005-01-24) — 3 min
68. [Why I don’t use the term “fixed and random effects”](https://statmodeling.stat.columbia.edu/2005/01/25/why_i_dont_use/) (2005-01-25) — 4 min
69. [The pinch-hitter syndrome:  a general principle?](https://statmodeling.stat.columbia.edu/2005/01/26/the_pinchhitter/) (2005-01-26) — 2 min
70. [Thoughts on Eric Johnson’s talk](https://statmodeling.stat.columbia.edu/2005/01/27/thoughts_on_eri/) (2005-01-27) — 6 min
71. [The death penalty . . . for forgery?](https://statmodeling.stat.columbia.edu/2005/01/31/the_death_penal/) (2005-01-31) — 3 min
72. [Spatial statistics and voting](https://statmodeling.stat.columbia.edu/2005/02/01/spatial_statist/) (2005-02-01) — 1 min
73. [Using base rate information?](https://statmodeling.stat.columbia.edu/2005/02/02/using_base_rate/) (2005-02-02) — 5 min
74. [Social networks and voter turnout](https://statmodeling.stat.columbia.edu/2005/02/03/social_networks/) (2005-02-03) — 1 min
75. [This is what a blog entry is supposed to look like](https://statmodeling.stat.columbia.edu/2005/02/03/this_is_what_a/) (2005-02-03) — 1 min
76. [Physicists modeling social phenomena; social scientists invoking physics](https://statmodeling.stat.columbia.edu/2005/02/04/physicists_mode/) (2005-02-04) — 2 min
77. [Agressive treatment, agressive teaching](https://statmodeling.stat.columbia.edu/2005/02/07/agressive_treat/) (2005-02-07) — 10 min
78. [A Very Delayed Lightbulb Over my Head](https://statmodeling.stat.columbia.edu/2005/02/07/a_very_delayed/) (2005-02-07) — 4 min
79. [FAQ on DIC and pD in bugs and bugs.R](https://statmodeling.stat.columbia.edu/2005/02/08/faq_on_dic_and/) (2005-02-08) — 2 min
80. [Reserving political offices for women in India](https://statmodeling.stat.columbia.edu/2005/02/09/reserving_polit/) (2005-02-09) — 3 min
81. [Parent to children asset transfers](https://statmodeling.stat.columbia.edu/2005/02/10/parent_to_child/) (2005-02-10) — 1 min
82. [Mythinformation](https://statmodeling.stat.columbia.edu/2005/02/11/mythinformation/) (2005-02-11) — 1 min
83. [Reducing arsenic exposure in Bangladesh](https://statmodeling.stat.columbia.edu/2005/02/15/reducing_arseni/) (2005-02-15) — 1 min
84. [Rationality and ideology:   Carrie McLaren’s reactions to Malcolm Gladwell](https://statmodeling.stat.columbia.edu/2005/02/15/rationality_and/) (2005-02-15) — 7 min
85. [Power calculations](https://statmodeling.stat.columbia.edu/2005/02/16/power_calculati/) (2005-02-16) — 2 min
86. [Bayesian modeling for kidney filtering](https://statmodeling.stat.columbia.edu/2005/02/17/bayesian_modeli/) (2005-02-17) — 2 min
87. [More on social networks and voting](https://statmodeling.stat.columbia.edu/2005/02/18/more_on_social/) (2005-02-18) — 5 min
88. [Bayes for medical diagnosis](https://statmodeling.stat.columbia.edu/2005/02/21/bayes_for_medic/) (2005-02-21) — 1 min
89. [Jasjeet’s R package for matching in observational studies](https://statmodeling.stat.columbia.edu/2005/02/22/jasjeets_r_pack/) (2005-02-22) — 1 min
90. [Causal inference and decision trees](https://statmodeling.stat.columbia.edu/2005/02/23/causal_inferenc/) (2005-02-23) — 2 min
91. [Contingency and alternative history](https://statmodeling.stat.columbia.edu/2005/02/24/contingency_and/) (2005-02-24) — 4 min
92. [International data](https://statmodeling.stat.columbia.edu/2005/02/25/international_d/) (2005-02-25) — 1 min
93. [Is voting contagious?](https://statmodeling.stat.columbia.edu/2005/02/28/is_voting_conta/) (2005-02-28) — 3 min
94. [Matching and matching](https://statmodeling.stat.columbia.edu/2005/03/01/matching_and_ma/) (2005-03-01) — 2 min
95. [EDA for HLM](https://statmodeling.stat.columbia.edu/2005/03/02/eda_for_hlm/) (2005-03-02) — 3 min
96. [Meritocracy won’t happen:  the problem’s with the “ocracy”](https://statmodeling.stat.columbia.edu/2005/03/03/meritocracy_the/) (2005-03-03) — 2 min
97. [The secret weapon](https://statmodeling.stat.columbia.edu/2005/03/07/the_secret_weap/) (2005-03-07) — 2 min
98. [Still more on R software for matching for causal inference](https://statmodeling.stat.columbia.edu/2005/03/08/still_more_on_r/) (2005-03-08) — 2 min
99. [p (A|B) != p (B|A)](https://statmodeling.stat.columbia.edu/2005/03/09/p_ab_p_ba/) (2005-03-09) — 1 min
100. [Contingency and ideology](https://statmodeling.stat.columbia.edu/2005/03/10/contingency_and_1/) (2005-03-10) — 3 min